JP7448205B2 - Invention document analysis system, invention document analysis method, and invention document analysis processing program - Google Patents

Description

The present disclosure relates to an invention document analysis system and related technology.

Conventionally, machine translation devices are known (for example, see Patent Document 1).

Japanese Patent Application Publication No. 2018-190287

Conventional machine translation devices can output appropriate translation results only when conditions such as the Japanese document to be translated is written in a certain format are met, but it is difficult to express technical ideas in Japanese. This is not necessarily useful when a translator or the like translates an invention document that has been created. For example, invention documents are not necessarily translated properly by conventional machine translation devices, and translators need to check the results translated by the machine translation device and correct them as appropriate.

The present disclosure provides an invention document system that can analyze an invention document expressing a technical idea in Japanese and output useful information to a person using the invention document. For example, the present disclosure allows a translator to efficiently translate a translation result document based on an invention document that expresses a technical idea in Japanese (that is, a document that is the result of translating an invention document into a foreign language such as English or Chinese). To provide an invention document analysis system capable of outputting information useful for purposeful or appropriate creation. Further, the present disclosure provides a method (invention document analysis method) and an invention document analysis processing program related to the invention document analysis system. Further, the present disclosure provides a display document that is generated by the invention document analysis method , is data that may include a script as a computer program, and includes a document written in a markup language.

The invention document analysis system in the present disclosure is an invention document analysis system that analyzes an invention document, which is a document that describes technical ideas in Japanese and describes claims and specifications for patent applications. a settings information acquisition unit that acquires configuration information including condition information indicating extraction conditions; a document acquisition unit that acquires the invention document; and an analysis of the invention document acquired by the document acquisition unit based on the configuration information. an analysis unit; and an output unit that outputs information generated based on an analysis result of the analysis unit regarding the invention document acquired by the document acquisition unit, and the analysis unit includes an invention document acquired by the document acquisition unit. the output unit includes a plurality of important words and phrases extracted by the important words and phrases from the invention document acquired by the document acquisition unit; Output information generated based on each phrase.

In addition, the invention document analysis method in the present disclosure uses a device equipped with a microprocessor to analyze an invention document, which is a document that describes claims and specifications for a patent application expressing a technical idea in Japanese, and to display a document for display. An invention document analysis method that performs an invention document analysis process that generates an invention document, wherein the invention document analysis process includes a document acquisition step for acquiring an invention document, and an invention document acquired by the document acquisition step that is determined in advance. A patent generated based on the analysis result of the analysis step of the invention document acquired by the document acquisition step by extracting important words based on the conditions for extraction of important words an output step of generating and outputting the display document for displaying application-related information, the display document including a document written in a markup language, and being interpreted by a browser in a device equipped with a microprocessor. Patent application related information including claim information representing the content of claims in an invention document for a patent application that expresses the technical idea in Japanese, and specification information representing the content of the specification in the invention document. is included as content to be displayed, and the claim information representing the content of the claim is extracted based on the conditions for extracting predetermined important words from among the words and phrases of the claim. For each important phrase, display control information for controlling the display mode of the display element of the important phrase is included, and among the important phrases expressed in the claim information, a plurality of character strings constituting the important phrases are different from each other. Each of the display control information relating to each of the important terms includes information for controlling the display elements of each of the plurality of important terms to be displayed in a display mode with mutually different display colors, The display control information related to an important phrase that is composed of the same character string as any one of the plurality of important phrases whose character strings are different from each other among the important phrases that are displayed. contains information for controlling the display element of the important term to be displayed in a display mode with the same display color as the display color added as the display mode of the display element of the one important term, The specification information representing the contents of the specification includes, for each important term extracted based on the conditions for extracting important terms among the terms in the specification, the display mode of the display element of the important term. Any one of the plurality of important words and phrases that include display control information for controlling the information, and among the important words and phrases represented in the specification information, the character strings forming the important words in the claim information are different from each other. The display control information relating to the same important word or phrase as the one important word or phrase may cause the display element of the important word to be displayed in the same display color as the display color added as the display mode of the display element of the one important word or phrase in the claim information. Contains information for controlling display in a display mode with added.
In addition, the invention document analysis processing program in the present disclosure uses a device equipped with a microprocessor to analyze and display an invention document, which is a document containing claims and specifications for a patent application expressing technical ideas in Japanese. This is a program for realizing the above-described invention document analysis method by performing an invention document analysis process that generates a document .

According to the invention document analysis system , invention document analysis method , and invention document analysis processing program in the present disclosure, useful information can be output. According to the display document in the present disclosure, useful information can be displayed. Further, the invention document analysis system, invention document analysis method, display document, and invention document analysis processing program in the present disclosure can be used, for example, to enable a translator to efficiently or appropriately create a translation result document based on an invention document. may be useful.

FIG. 1 is a diagram showing an example of a work flow related to the translation of an invention document. FIG. 2 is a diagram showing an example of the flow of translation work by a translator. FIG. 3 is a diagram listing examples of factors that impede efficiency or appropriateness of translation for acquiring useful rights. FIG. 4 is a diagram showing an example of the configuration of the invention document analysis system according to the first embodiment. FIG. 5 is a diagram showing an example of the contents of setting information according to the first embodiment. FIG. 6 is a diagram illustrating an example of the content of patent application related information according to the first embodiment. FIG. 7 is a flowchart showing an example of the invention document analysis process according to the first embodiment. FIG. 8 is a diagram showing an example of the hardware configuration of the invention document analysis system according to the first and second embodiments. FIG. 9 is a diagram showing an example of the hardware configuration and network configuration of the invention document analysis system according to the first and second embodiments. FIG. 10 is a diagram showing another example of the hardware configuration and network configuration of the invention document analysis system according to the first and second embodiments. FIG. 11 is a diagram showing an example of the configuration of an invention document analysis system according to the second embodiment. FIG. 12 is a diagram showing an example of main data types handled by the invention document analysis system according to the second embodiment. FIG. 13 is a diagram showing an example of the contents of setting information according to the second embodiment. FIG. 14 is a diagram showing an example of standard provisions for invention documents for translation. FIG. 15 is a diagram showing an example of the content of patent application related information according to the second embodiment. FIG. 16 is a diagram showing a configuration example of an HTML document, which is an example of a display document related to patent application related information. FIG. 17 is a diagram showing an example of a display screen displayed by a browser or the like based on a display document related to patent application related information. FIG. 18 is a diagram showing an example of important phrase management information. FIG. 19 is a diagram showing an example of items of attribute information of important words and phrases in the important words and phrases management information. FIG. 20 is a diagram illustrating a display example of the inclusion relationship information portion of the important word/phrase relationship information of patent application related information as a display document. FIG. 21 is a diagram showing an example of warning information items and description items of patent application related information. FIG. 22 is a diagram showing an example of claim information in patent application related information as a display document displayed on a browser. FIG. 23 is a diagram showing an example of information displayed in response to an operation on an important phrase in the complaint information displayed on the browser. FIG. 24 is a diagram showing an example of a screen displayed in response to an operation on an important phrase in the complaint information displayed on the browser. FIG. 25 is a diagram showing an example of claim-based specification information in patent application related information as a display document displayed on a browser. FIG. 26 is a diagram showing an example of specification information in patent application related information as a display document displayed on a browser. FIG. 27 is a diagram showing an example of a screen displayed in response to an operation on a code added to an important phrase in the specification information displayed on the browser. FIG. 28 is a diagram showing an example of a specification heading list in patent application related information as a display document displayed on a browser. FIG. 29 is a diagram showing another example of a specification heading list in patent application related information as a display document displayed on a browser. FIG. 30 is a diagram showing an example of a code list and a drawing code list in patent application related information as a display document displayed on a browser. FIG. 31 is a diagram showing an example of a screen displayed in response to an operation on a code in the code list displayed on the browser. FIG. 32 is a diagram showing an example of a constituent claim tree in patent application related information as a display document displayed on a browser. FIG. 33 is a diagram showing an example in which a citation relationship is severed in response to an operation on a symbol for expressing a citation relationship in a constituent claim tree displayed on a browser. FIG. 34 is a diagram showing an example of claim translation information in patent application related information as a display document displayed on a browser. FIG. 35 is a diagram showing an example of a menu for selecting elements of a set of translation word candidates displayed in response to an operation on a translation word from an important word in the complaint translation information displayed on the browser. FIG. 36 is a diagram showing an example of an important word list in patent application related information as a display document displayed on a browser. FIG. 37 is a diagram illustrating an example in which a set of translation word candidates in the important word list displayed on the browser is changed to a corresponding translation word in response to an operation of selecting an element thereof. FIG. 38 is a diagram showing another example of an important word list in patent application related information as a display document displayed on a browser. FIG. 39 is a diagram showing an example of a display screen displayed by the display unit of the invention document analysis system according to the second embodiment. FIG. 40 is a flowchart illustrating an example of invention document analysis processing according to the second embodiment. FIG. 41 is a flowchart showing an example of the citation relationship analysis process. FIG. 42 is a flowchart illustrating an example of complaint important phrase extraction processing. FIG. 43 is a flowchart showing an example of specification important word extraction processing. FIG. 44 is a flowchart showing an example of inter-element relationship determination processing. FIG. 45 is a flowchart illustrating an example of a process for determining the importance of each important term. FIG. 46 is a flowchart showing an example of warning information generation processing. FIG. 47 is a flowchart showing an example of complaint information generation processing. FIG. 48 is a flowchart showing an example of complaint information generation processing. FIG. 49 is a flowchart showing an example of code determination processing. FIG. 50 is a flowchart showing an example of claim-based drawing information generation processing. FIG. 51 is a flowchart showing an example of claim-based specification information generation processing. FIG. 52 is a flowchart showing an example of claim-based summary information generation processing. FIG. 53 is a flowchart showing an example of the modified complaint information generation process. FIG. 54 is a flowchart showing an example of claim tree generation processing. FIG. 55 is a flowchart illustrating an example of statement information generation processing. FIG. 56 is a flowchart illustrating an example of generation processing of various translation information. FIG. 57 is a flowchart illustrating an example of important word list generation processing. FIG. 58 is a flowchart showing an example of a learning model generation process for important word/phrase translation. FIG. 59 is a flowchart showing an example of a translation learning model generation process.

The invention document analysis system in the present disclosure is an invention document analysis system that analyzes an invention document, which is a document that describes technical ideas in Japanese and describes claims and specifications for patent applications. a settings information acquisition unit that acquires configuration information including condition information indicating extraction conditions; a document acquisition unit that acquires the invention document; and an analysis of the invention document acquired by the document acquisition unit based on the configuration information. an analysis unit; and an output unit that outputs information generated based on an analysis result of the analysis unit regarding the invention document acquired by the document acquisition unit, and the analysis unit includes an invention document acquired by the document acquisition unit. the output unit includes a plurality of important words and phrases extracted by the important words and phrases from the invention document acquired by the document acquisition unit; This is an invention document analysis system that outputs information generated based on each word.

As a result, the invention document analysis system outputs information based on the results of analyzing an invention document containing claims and specifications prepared by, for example, an inventor (for example, the inventor or a person related to the inventor) or a patent attorney. For example, for a translator who translates the invention document (for example, a translator or other person using the invention document), the understanding of the invention document may be facilitated by referring to the output information. . Facilitating the understanding of invention documents will be useful for translators to appropriately or efficiently translate invention documents. Information generated based on key words in an invention document can be useful in cases where a translator determines corresponding translated words corresponding to key words in an invention document before translating the entire invention document.

Here, the condition information includes important term candidate character string pattern information indicating an important term candidate character string pattern representing a pattern of character strings that are candidate important terms, and a reference used to refer to the previously described character string. Reference string pattern information that indicates a reference string pattern that represents a string pattern, and a code character that indicates a code string pattern that represents a code string pattern that is appended to a string that represents an element name. column pattern information and specific particle string pattern information indicating a specific particle string pattern representing a pattern of strings that are predetermined particles; A predetermined character string whose first character is katakana is determined to correspond to the relevant important word candidate character string pattern, and the reference character string pattern information is at least '' is determined to correspond to the reference character string pattern, and the code string pattern information is such that at least a predetermined character string whose first character is a number corresponds to the code string pattern. The specific particle string pattern information includes at least one of the following characters: "wo", "ni", "ga", "to", "kara", "he" and "wa". The characters are determined to correspond to the specific particle character string pattern, and the important word/phrase extraction unit extracts a first character string that corresponds to the reference character string pattern into the invention document acquired by the document acquisition unit. a first extraction unit that extracts a character string that is the same as the second character string as an important word when a second character string that corresponds to the important word candidate character string pattern exists subsequent to the key word candidate character string pattern, and the document acquisition unit If there is a period character following the third character string that corresponds to the important word candidate character string pattern in the claim of the invention document obtained by A second extraction unit extracts a fifth character string that corresponds to the code character string pattern subsequent to a fourth character string that corresponds to the important word candidate character string pattern in the invention document acquired by the document acquisition unit. exists, a third extraction unit that extracts a character string that is the same as the fourth character string as an important term, and an invention document acquired by the document acquisition unit that corresponds to the important term candidate character string pattern. a fourth extraction unit that extracts a character string that is the same as the sixth character string as an important phrase when there is a seventh character string that follows the sixth character string and that corresponds to the specific particle character string pattern; In the extraction of important words, when a plurality of character strings correspond to the important word candidate character string pattern, one of the plurality of character strings is replaced by another character string. When the string is included in the description of the invention document, the remaining character strings excluding the one character string among the plurality of character strings are selected as the selection target, and the important words are selected and extracted, and the output unit: Outputting patent application related information generated based on the analysis result by the analysis unit regarding the invention document acquired by the document acquisition unit, the patent application related information is generated from the invention document acquired by the document acquisition unit. Each of the plurality of important words and phrases extracted by the important word and phrase extracting unit is output as a display document that is included as content to be displayed, and the patent application related information is to be displayed in the display document. The claim information representing the content of the claim of the invention document acquired by the document acquisition unit includes claim information representing the content of the claim of the invention document acquired by the document acquisition unit, and the claim information representing the content of the claim of the invention document acquired by the document acquisition unit is Contains display control information for controlling the display mode of each of the plurality of important words and phrases in the claim extracted by the extraction unit, and the patent application related information should be displayed in the display document. , the specification information representing the content of the specification of the invention document acquired by the document acquisition unit, the specification information representing the content of the specification of the invention document acquired by the document acquisition unit, the invention For each of the plurality of important words and phrases in the specification extracted from the document by the important word and phrase extraction unit, display control information for controlling the display mode of the important words and phrases is included, and in the claim information and the specification information, The display control information for controlling the display mode of each of the plurality of key words extracted by the key word/phrase extracting unit controls the display mode of the key word/phrase corresponding to the display control information. Includes highlighting information for differentiating the display mode of words other than the plurality of important words in the claim information and the specification information, and the character strings constituting the important words among the important words expressed in the claim information. Each of the display control information for controlling the display mode corresponding to each of the plurality of important terms and phrases that is different from each other includes information that is different from each other for controlling each of the plurality of important terms and phrases to be displayed in a different display mode. of the important words and phrases represented in the specification information, the character strings constituting the important words in the claim information are the same as any one of the plurality of important words and phrases that are different from each other. The display control information for controlling the display mode of the important word/phrase may cause the display element of the key word/phrase to be displayed in a display mode that is at least partially similar to the display mode of the display element of the one key word/phrase in the claim information. Includes information for display control, and among the important words and phrases represented in the specification information, the character strings constituting the important words in the claim information are different from any of the plurality of important words and phrases that are different from each other. The display control information for controlling the display mode of the key word/phrase controls the display element of the key word/phrase to be displayed in a display mode different from any of the display elements of the plurality of key words/phrases in the claim information. It may also include information for

As a result, for example, translators, etc. (translators, inventors, etc., patent attorneys, or other persons) can use the claim information and specification information of patent application-related information related to display documents displayed by browsers, etc. It may be possible to identify important words and phrases whose display manner is controlled. In addition, translators and others who view patent application-related information will be able to identify whether each important word or phrase described in the specification information in the patent application-related information is an important word or phrase described in the claim. , it may be possible to quickly understand whether a plurality of important words and phrases are the same or different depending on the display mode of each important word and phrase. In addition, if the invention document analysis system controls the display mode of important words and phrases, for example, by highlighting them (for example, by displaying them in bold), translators, etc. can quickly grasp the important words and phrases in claim information. It can be. Note that the display document generated by this invention document analysis system (that is, the display document that includes patent application related information to be displayed) is subject to interpretation by a browser installed on a device equipped with a microprocessor. As a result, patent application related information will be displayed. Translators and the like can efficiently understand the invention document by viewing the patent application related information in the display document generated based on the invention document.

Also, for example, for each of a plurality of important words in the claim extracted by the important word extraction unit, the claim information representing the contents of the claims of the invention document acquired by the document acquisition unit, The display control information includes information indicating a display color to be added to the important word/phrase, and includes the important word/phrase included in the specification information representing the content of the specification of the invention document acquired by the document acquisition unit. The display control information related to each of the plurality of important words and phrases in the specification extracted by the extraction unit includes information indicating a display color to be added to the important word and phrase, and is displayed in the claim information. The display control information relating to each of a plurality of important words and phrases whose character strings are different from each other among the important words and phrases that are displayed includes changing the display element of each of the plurality of important words to the identification information for each important word. Correspondingly, a character string that includes information for controlling display in a display mode with a different display color added, and that constitutes an important phrase in the claim information among the important phrases represented in the specification information. The display control information relating to an important phrase that is the same as any one of the plurality of important phrases whose values are different from each other is configured to change the display element of the important phrase to a display element of the one important phrase in the claim information. The information may include information for controlling the element to be displayed in a display mode with the same display color added as the display color added as the display mode of the element.

This allows a translator or the like to view patent application related information as a display document using, for example, a PC, smartphone, or other computer equipped with a browser. Translators, etc. who view patent application related information as a display document displayed by a browser etc. can identify the display color added to important words in the patent application related information and identify whether a certain important word is a claim. It becomes possible to identify whether or not a plurality of important words and phrases are listed, and it is possible to quickly understand whether a plurality of important words and phrases are the same or different.

Further, for example, the display document is an HTML document that includes the patent application related information as content to be displayed and includes an HTML file written in HTML (Hypertext Markup Language), and important words and phrases in the claim information Display control information related to an important phrase in the claim information, which is composed of the same character string as any one of the plurality of important phrases whose character strings are different from each other, , is for controlling the display in a display mode with the same display color as the display color added as the display mode of the one important word/phrase, and the character string forming the important word/phrase in the complaint information. The display control information relating to each of a plurality of important words and phrases that are different from each other displays each of the plurality of important words in a display manner in which a left border line having a predetermined width is added in a display color that is different for each important word. For each of the plurality of important words and phrases in the specification information, the key word and phrase are the same as any one of the plurality of important words and phrases in the claim information. For example, the display control information regarding the important word/phrase in the specification information may display the important word/phrase in the same display color as the display color of the left border line related to the display mode of the one important word/phrase in the claim information. This is to control the display in a display mode with a left border line having a predetermined width, and to control whether the relevant important word in the specification information is different from any of the plurality of important words in the claim information. For example, the display control information regarding the important word/phrase in the specification information is for controlling the important word/phrase to be displayed in a display mode in which the left border line having the predetermined width is not added. It may also be a thing.

As a result, a translator or the like can view patent application related information as a display document, which is an HTML document, using a PC, smartphone, or other computer equipped with a browser compatible with HTML. Translators who view patent application related information as a display document displayed by a browser, etc., can distinguish between different important words included in claim information in the patent application related information by comparing the left and right words added to each important word. By identifying the display color of the border line, it can be quickly distinguished. In addition, there is a method of expressing different important words by using different display colors of the left border added to the important words. This has the advantage that the legibility of the important words themselves is not reduced regardless of whether the background color of the area is dark (for example, black) or light (for example, white). In addition, the left border that can be added to some display elements in HTML documents is not a character, so it will be ignored when copying and pasting the text of the display content that includes important words, so the display content It does not get in the way when copying and pasting into a text editor etc. to edit descriptions that include important words.

Further, for example, the identification information of a plurality of important words and phrases in which the character strings forming the important words are different from each other in the claim information representing the contents of claims of the invention document acquired by the document acquisition unit. includes information indicating a serial number determined in ascending order of the earliest position of an important phrase composed of the same character string as the important phrase in the claim, and includes information on the plurality of important phrases in the claim information. The display control information concerning each of the plurality of important words and phrases includes two important words and phrases among the plurality of important words of the first group in which the serial number related to the identification information for each important word is equal to or less than a predetermined number. The minimum value of the difference in the display color added to each important term in each set for all of the sets is the number of important terms in the second group whose serial number is greater than the predetermined number and the plurality of important terms in the first group. Among a plurality of important words and phrases in a mixed group of words and phrases, the difference in the display color added to each important word in each set of two important words is greater than the minimum value for all of the sets, It may be for controlling the display color added to each important word/phrase.

As a result, when each important term in the claim information is displayed in the patent application related information as a display document using a browser, etc., the translator etc. who viewed the claim information can quickly identify each important term in the claim information. Identification may be possible. Adjustment of the display color based on the serial number related to the identification information of important words (for example, the first line number of the important word in the claim of an invention document) using the display control information of important words can be performed in the section where the claim information has a relatively small item number. Since the difference in display color between a plurality of important words arranged in a relatively important portion (that is, a relatively important portion) can be made relatively large, the distinguishability between important words in that portion can be improved.

Further, for example, the predetermined width is the width of one character, and for the plurality of important words and phrases in the complaint information whose character strings forming the important words are different from each other, the display control information related to the important words and phrases is , is for controlling the important word/phrase to be displayed in a display mode with no right border line, top border line, and bottom border line added, and the difference in display color is due to the hue component of the display color. It may be a difference or a difference in each component of RGB values representing display colors.

As a result, the boundary line added to the important word/phrase becomes a simple one with only the left border line, so the visibility of the important word/phrase is improved. Also, if the left border line has a width of one character (for example, a width represented by 1em), a rectangle filled with the display color of one character is displayed immediately before each of the plurality of important words in the complaint information. Therefore, even if a rectangle for one character is mixed in a character string, it does not easily disturb the arrangement of each character in the complaint information, so the effect is that the visibility of each character is not easily obstructed. For example, if a monospaced font is used as the font for displaying each character in claim information in a display document, if a rectangle for one character is mixed in the character string, the arrangement of each character may be Even if a proportional font that is not monospaced is used as the font for displaying each character in the claim information, the width of a rectangle for one character is Because there is not much difference in width (especially the width of relatively many Japanese full-width characters such as kanji, hiragana, katakana, etc.), when viewing complaint information visually, the arrangement of each character is disturbed and visibility is greatly hindered. There is no such thing.

Further, for example, the patent application related information includes each line of the claim information, and for each line or item of the included claim information, immediately after the content of the line or item, the line or The patent application related information includes claim translation information in which a translation line containing the result of replacing the important word/phrase included in the one paragraph with a corresponding translated word corresponding to the important word/phrase is inserted, and the patent application related information includes each line of the specification information. Then, for each line of the included specification information, immediately after the contents of the one line, a translation line containing the result of replacing the important word/phrase included in the one line with the corresponding translated word corresponding to the important word/phrase is inserted. The claim translation information and the specification translation information include corresponding translations corresponding to the same important phrase as each of the plurality of important phrases in which the character strings constituting the important phrases in the claim information are different from each other. includes display control information for controlling the display mode of the corresponding translated word corresponding to the important word/phrase, and includes display control information for controlling the display mode of the corresponding translated word corresponding to the important word/phrase in the claim translation information and the specification translation information. The display control information is for controlling the corresponding translated word to be displayed in a display mode in which a left border line having the predetermined width and the same color as the display color added to the important word/phrase is added. You can also use it as

As a result, for example, a translator etc. who viewed claim translation information and specification translation information of patent application related information related to a display document displayed on a browser etc. can add important words and phrases and corresponding translated words. By identifying the display color of the left border line, it becomes possible to quickly grasp which important word/phrase described in a claim a certain corresponding translated word corresponds to.

Further, for example, the patent application related information includes claim translation information representing the result of replacing important words included in the content of the claim information with corresponding translated words corresponding to the important words, and the patent application related information includes: The claim translation information and the specification translation information include specification translation information representing the result of replacing important words and phrases included in the content of the specification information with corresponding translated words corresponding to the important words and phrases, and the claim translation information and the specification translation information are Display control information for controlling the display mode of the corresponding translated word corresponding to the important word/phrase with respect to the corresponding translated word corresponding to the same key word/phrase as each of the plurality of important word/phrases whose character strings forming the key word/phrase are different from each other. and the display control information related to the corresponding translated word corresponding to the important word in the claim translation information and the specification translation information displays the corresponding translated word in the same color as the display color added to the important word. The display may be controlled to display in a display manner in which the left border line having the predetermined width is added.

As a result, for example, translators who view claim information, claim translation information, and specification translation information of patent application-related information related to display documents displayed on a browser, etc., can check the important words and phrases of the claim information and the corresponding translations. By identifying the display color of the left border line added to a word, it is possible to quickly understand which important phrase in the claim information a certain corresponding translated word in claim translation information and specification translation information corresponds to. It becomes possible to do so.

Further, for example, the setting information includes information for specifying an important word/phrase setting dictionary that lists important words to be extracted, and in the important word/phrase setting dictionary, at least some of the listed important words are Corresponding translation words are associated with each other, and the invention document analysis system includes an important word/phrase setting dictionary acquisition unit that acquires the important word/phrase setting dictionary based on the setting information, and the important word/phrase extraction unit is configured to match the important word/phrase setting dictionary. If the important words listed in the setting dictionary are included in the invention document acquired by the document acquisition unit, the important words are extracted from the invention document and the important words in the claim information or the specification information are extracted. If a word is an important word that is associated with a corresponding translation in the important word/phrase setting dictionary, the important word/phrase setting is used as a corresponding translation of the important word in the claim translation information or the specification translation information. A corresponding translation word associated with the important word/phrase in a dictionary may be used.

This allows, for example, an inventor or patent attorney who has created an invention document to specify certain words as important words, or to specify corresponding translations for at least some of the important words. By generating an important word/phrase setting dictionary, the invention document analysis system can output appropriate patent application related information. Therefore, the translator may be able to efficiently or appropriately translate the invention document by referring to the patent application related information. In addition, when translators have defined corresponding translations for important words, by creating an important word setting dictionary, invention documents provided by inventors or patent attorneys can be used in the invention document analysis system. When analyzed, appropriate patent application related information can be output.

Further, for example, in the display document related to the patent application related information generated for the invention document acquired by the document acquisition unit, the display related to the important words extracted from the invention document by the important word extraction unit The control information changes the display element of the key term to a web page that includes the key term, a definition of the key term, a synonym of the key term, or a synonym of the key term on a search site on the network, depending on the operation. The present invention may function as a graphical user interface (GUI) that executes the process of searching for related images or the process of translating the important phrases at a translation site on the network.

As a result, translators, etc. who viewed patent application related information related to display documents displayed on a browser etc. can perform operations on important words in the displayed patent application related information (in other words, operations on display elements of important words). , it may be possible to obtain information such as search results for the important words (for example, web pages containing the important words, definitions of the important words, synonyms, related images), translation results of the important words, etc. It may become possible to understand efficiently or appropriately. Therefore, the translator may be able to efficiently or appropriately determine translation terms for important terms.

Further, for example, in the display document related to the patent application related information generated for the invention document acquired by the document acquisition unit, the display related to the important words extracted from the invention document by the important word extraction unit The control information is configured to display at least part of the attribute information of the important term or the importance of the important term determined based on the attribute information of the important term, depending on the operation of the display element of the important term. The attribute information of the important words extracted from the invention document is used as a GUI for displaying the important words and phrases when there is a claim section defining the important words in the claims of the invention document. Information indicating the section number of the relevant section, information indicating the section number of the section if there is a claim section that refers to the important term in the claims of the invention document, and information indicating the section number of the section in the claim of the invention document that references the important word In the case where there is a description in which a code is added to a word, information indicating the code and information identifying whether or not there is a description in which the important word or phrase is used as a subject in the invention document are included. Good too.

As a result, translators who view patent application related information related to display documents displayed on a browser, etc., can obtain information on the importance level or attribute information of the displayed important words by operating on the displayed important words. Therefore, it may be possible to recognize to some extent which important words and phrases in the invention document are important, and it may be possible to efficiently or properly understand the important words and phrases. Therefore, the translator may be able to translate efficiently by, for example, preferentially determining translation terms for important key terms.

Further, for example, in the display document related to the patent application related information generated for the invention document acquired by the document acquisition unit, the display related to the important words extracted from the invention document by the important word extraction unit The control information is a GUI for executing an important word inclusion description enumeration process that enumerates and displays the display element of the important word from the claim information and the specification information, in accordance with the operation. The important word inclusion description enumeration process enumerates and displays the lines containing the important words to be operated from the claim information and the specification information, and The relevant important words and phrases are highlighted, and the displayed line is displayed as a GUI for displaying links to multiple lines including the relevant line in the claim information or specification information, depending on the operation on the relevant line. It may also be a process of adding display control information for functioning.

As a result, a translator, etc. who viewed patent application related information related to a display document displayed on a browser, etc., can perform operations on important words in the displayed patent application related information, and as a result of the important word inclusion description enumeration process. You can check the lines containing the important words in the claim information and specification information that are listed and displayed. Therefore, the translator or the like may be able to efficiently or appropriately understand important words and phrases in the invention document. Moreover, for this reason, the translator may be able to efficiently or appropriately determine translation words for important words. In addition, the translator or the like can accurately identify the important word/phrase because the important word/phrase is highlighted in the line enumerated by the important word/phrase inclusion description enumeration process by including the manipulated important word/phrase. In addition, translators, etc. can perform further operations on one of the lines enumerated by the important word inclusion description enumeration process by including the manipulated important word, so that it will be further displayed. Multiple lines in claim information, specification information, or another invention document that precede and follow one line can be easily confirmed.

Further, for example, the patent application related information includes an important word list in which important words are listed on a new line for each important word extracted from the invention document by the important word extraction unit, and each line of the important word list is arranged in a predetermined manner. The order is the order of importance determined for the important words and phrases for each line, the order according to the result of sorting the important words and phrases for each line in character code order, and the invention for the important words and phrases for each line. It may be arranged in the order of the number of uses in the document, or the order of the number of uses in the claims of the invention document for each line of important words.

As a result, translators can quickly understand the important words included in an invention document by checking the list of important words in patent application related information, and for example, decide which words to translate for the important words listed in the list By starting the translation of the entire invention document after doing so, efficient translation may be possible. In addition, for translators who view a list of important words in which important words are arranged in order of importance, in order of frequency of use in invention documents, or in order of frequency of use in claims, the important words are arranged in order of importance, frequency of use in invention documents, or frequency of use in claims. Efficient translation may be possible by sequentially determining translation terms for each important term. In addition, for translators who view the important word list in which important words are arranged in the order according to the sorting result in the order of character codes, it is important to note that multiple important words that are arranged consecutively in the important word list and differ only in their endings ( For example, by comparing "temporary recording folder" and "temporary recording folder" or "external storage device" and "external storage sending"), it is possible to quickly detect errors in the endings of important words. obtain. In addition, inventors, patent attorneys, or translators can edit the list of important words in the patent application related information that is output as a result of having the invention document analysis system analyze the invention document. It may be possible to efficiently generate and update an important word/phrase setting dictionary that can be used to appropriately extract important words.

Further, for example, the display control information relating to each of a plurality of important words and phrases in which the character strings constituting the important words are different from each other in the claim information may be such that each of the plurality of important words is displayed in a different display color for each important word. This is to control the display in a display mode in which a left border line having a predetermined width is added, and for each important word in the important word list, the important word is one of the plurality of important words in the claim information. If the important word is the same as any one important word in the important word list, the important word is displayed on the left border line according to the display mode of the one important word in the claim information. Display control information may be added that controls the display to be displayed in a display mode in which a left border line having the predetermined width is added in the same display color as the color.

As a result, translators and others who view patent application related information as a display document will be able to distinguish between important words and phrases listed in the claim information from among the important words and phrases in the list of important words. It may be possible to quickly identify key words in the list of key words that correspond to each key word in the key word list.

Further, for example, the analysis unit may generate a set of corresponding translation word candidates for one or more important words extracted by the important word extraction unit from the invention document acquired by the document acquisition unit. It includes a translation word candidate set specifying unit for identifying a certain translation word candidate set, and each line of the important word list includes one important word and a corresponding translation word or translation word candidate set corresponding to the important word phrase. It may also be a thing.

As a result, by referring to the translation word candidate set or the corresponding translated words that correspond to the important words included in the important words list of patent application related information output by the invention document analysis system, the translator can, for example, respond to the important words. By selecting one translated word/phrase from a set of translated word candidates, it may be possible to efficiently or appropriately determine the translated word/phrase corresponding to the important word/phrase.

Further, for example, in a line containing an important word and a translation word candidate set corresponding to the important word in the important word list, if the number of elements of the translation word candidate set is plural, the translation word The multiple elements of the candidate set are listed in order of validity, which indicates the evaluation value of the degree of validity as a translation word of the relevant important word, and are included in the translation word candidate set that includes multiple elements and is included in the important word list. In the corresponding display element, each element of the translation word candidate set is displayed in a selectable manner, the selected element is determined as a corresponding translation word, and the content of the display element of the translation word candidate set is displayed as the corresponding translation word. In order to have a function as a GUI that executes the process of replacing and displaying translated words, and the process of generating and outputting translation auxiliary information, which is information that associates the important words and phrases with the corresponding translated words. Display control information is added, and the analysis unit calculates the translated word using predetermined calculation processing based on data generated by collecting and integrating a plurality of pieces of translation auxiliary information output in the past by the invention document analysis system. It is also possible to have a function of calculating the validity of the elements of the candidate set.

As a result, by referring to a set of translation word candidates that includes elements in order of validity that correspond to important words included in the important word list, translators can efficiently or appropriately select translation words corresponding to important words. It may be possible to determine. In addition, translators who view the list of important words or people related to the translator (for example, other translators who belong to the same department or group as the translator) may have previously selected translation words from a set of candidate words. It is relatively easy to operate an invention document analysis system in such a way that the validity of the elements of the translation word candidate set corresponding to the important words can be calculated based on the results obtained, so it is important to keep the validity high. This can be achieved relatively easily.

Further, for example, the analysis unit includes a warning information generation unit that generates warning information based on the invention document acquired by the document acquisition unit, and the patent application related information includes the warning information, and the patent application related information includes the warning information. When the generation unit detects that the invention document acquired by the document acquisition unit includes a description that does not satisfy the condition that the same element name is written immediately before the same code, includes indication information indicating the same code and a plurality of element names with different expressions to which the code is added, in the specification of the invention document acquired by the document acquisition unit. If there is a parenthetical statement between opening parentheses and closing parentheses, the condition that the relationship between the parenthetical statement and the description of the character string immediately preceding the parenthetical statement is clear is met. When it is detected that a statement in parentheses that is not specified is included, it is added to the above-mentioned warning information to indicate that the meaning of the statement in parentheses is not clear and the position of the statement in parentheses in the specification of the invention document. If a word or phrase is written following the reference character string in any section of a claim of the invention document acquired by the document acquisition unit, the referenced phrase that is the same as the word or phrase is included in the claim of the invention document acquired by the document acquisition unit. For reference strings that do not meet the conditions that the reference string must be listed in the part that precedes the reference string, or that the section must be listed in the cited section based on the citation chain, If it is detected that such a description is included, a message indicating that there is no previous description of the referenced phrase, the number of the section that includes the description of the detected reference character string, and the reference character. By including indication information indicating the content of the description related to the column in the warning information and using a predetermined list of subordinate concept words, other clauses of the claims of the invention document acquired by the document acquisition unit can be included in the warning information. If it is detected that an unquoted independent term contains a statement that does not meet the condition of not using words that indicate subordinate concepts, the item number of the independent claim and the word or phrase are The warning information includes indication information indicating that words and phrases indicating subordinate concepts are used, and the important words and phrases extracted by the important phrase extraction unit from the claims of the invention document acquired by the document acquisition unit are included in the warning information. If it is detected that the description of the invention document does not satisfy the condition that the word/phrase is stated in the specification, the important word/phrase and the fact that the important word/phrase is not stated in the specification are determined. The warning information includes indication information indicating that the two important words extracted by the important word extraction unit from the invention document acquired by the document acquisition unit are one of the important words. If a predetermined number of characters are deleted from one of the important words and phrases that constitute the same, the important words and phrases of one of the deleted words and phrases are the same as the other important word and phrase of the other important words and phrases. In addition, the warning information includes indication information indicating that there is a possibility of a typo in the two important words and one of the two important words, and the invention document acquired by the document acquisition unit is When two important phrases extracted by the important phrase extraction section are converted from half-width characters constituting one of the important phrases to full-width characters, the one important phrase after the conversion, In cases where there is a relationship in which the other important words and phrases are the same, an indication that the two important words and phrases and the notation of the two important words and phrases are formally inconsistent. information may be included in the warning information.

As a result, by referring to the warning information in the patent application related information output by the invention document analysis system (for example, the pointed information in the warning information, etc.), the inventor, etc. or patent attorney who created the invention document can Amendments to the invention document may be made, if any. In addition, by receiving the invention documents to be translated and the patent application related information from the inventor or patent attorney, the translator may be able to receive information from the patent application related information even if the warning information in the patent application related information remains. It can be assumed that the inventor or patent attorney has determined that there is no need to take particular action on the indication, so it is possible to obtain an appropriate invention document to be translated, and as a result, the translation of the invention document It may be possible to perform the following efficiently or appropriately.

For example, the warning information generation unit may display the claim corresponding to the indication of the indication information in at least one display element of the indication information in the warning information included in the patent application related information as the display document. Display control information may be added to allow the display element to function as a GUI for displaying information or a link to a description location in the specification information.

As a result, an inventor, etc. or patent attorney who viewed the warning information of patent application related information related to a display document displayed by a browser etc. can point out the pointed information by operating the pointing information in the displayed warning information. Since it may be possible to confirm the description location in the claim information or specification information that corresponds to the above, the invention document can be appropriately revised as necessary.

Further, for example, the patent application related information includes a specification heading list in which lines including headings in the specification of the invention document acquired by the document acquisition unit are listed in the order of description in the specification, and the invention The document analysis system inserts one or more types of predetermined parentheses (for example, corner brackets) at the beginning of each line in the specification of the invention document or at a position following one or more blank characters at the beginning of the line. (brackets, angle brackets, round brackets, etc.) that are sandwiched between opening and closing parentheses, either unconditionally or on the condition that there is no period character "." at the end of the line. , it may be handled as a line that includes the description of the heading.
As a result, a translator or the like who has viewed the specification heading list in the patent application related information can obtain an overview of what is stated in the specification of the invention document. In addition, by viewing the specification heading list, the inventor, etc. or patent attorney can quickly confirm whether the description of each heading in the specification of the invention document is appropriate. For example, in the specification heading list in the patent application related information output to the invention document analysis system, the descriptions "(First Embodiment)" and "(Third Embodiment)" are included as two consecutive lines. Accordingly, it can be understood that the heading "(Second Embodiment)" is missing in the specification of the invention document that is the subject of analysis in the invention document analysis system.

Further, for example, the invention document analysis system treats a line including a description sandwiched between opening and closing corner brackets as a line including a description of the heading, and includes the invention in the specification heading list. For a line containing a heading representing a paragraph number in the specification of a document, important words extracted from the claims of the invention document by the important phrase extraction unit are included within the paragraph range related to the paragraph number in the specification. If there is an important phrase that is the same as the phrase, the important phrase may be added to the corresponding line of the specification heading list.
As a result, a translator or the like who views the specification heading list in the patent application related information may be able to quickly infer or confirm the outline of what is stated in the specification of the invention document. If the translator can quickly guess the outline of the specification, the possibility that the translator will be able to efficiently or appropriately translate the sentences written in the specification increases. In addition, for each important word/phrase (that is, the important word/phrase extracted from the claim) added to the line containing the description of the heading representing the paragraph number in this specification heading list, important words with the same character string are displayed in the same color. (For example, the display color of the left boundary line of the important word/phrase) may be added, and different important words/phrases may be given different display colors. Adding a display color to an important word or phrase is realized, for example, by display control information for a display element regarding the important word or phrase. This can increase the possibility that a translator or the like who has viewed the list of specification headings related to a display document displayed by a browser or the like can quickly recognize whether each important phrase is the same or not. In addition, in addition to adding important words extracted from the claims to the line containing the description of the heading representing the paragraph number in the specification heading list, important words and phrases extracted from the specification that are not included in the claims of the invention document are added. If an important word/phrase exists in the paragraph with that paragraph number, the important word/phrase may be added so that it is displayed in a different display mode than the important word/phrase extracted from the claim.

Further, for example, the analysis unit determines the relationship between elements expressed by each of the important words extracted by the important word extraction unit from the invention document acquired by the document acquisition unit, and the important words and phrases that indicate the result of the determination. The inter-element relationship determination unit includes an inter-element relationship determination unit that generates relationship information, the patent application related information includes the important word-phrase relationship information, and the inter-element relationship determination unit extracts the important word-phrase from the invention document into the important word-phrase relationship information. Inclusion relationship information expressing the structural inclusion relationship between the elements represented by each of the plurality of important words extracted by the important word/phrase section, and the relationship between the elements represented by each of the plurality of important words/phrases extracted from the invention document by the important word/phrase extraction section. If there is a relationship between a recording medium and the contents of the recording medium, record-related information expressing the relationship, and elements expressed by each of the plurality of important words extracted from the invention document by the important word extraction section are related to the program and the relevant information. Program realization related information that expresses the relationship when there is a relationship with the content realized by the program, and the equivalent relationship between the elements expressed by each of the plurality of important words extracted from the invention document by the important word extraction unit. Equivalent relationship information expressing the relationship, and hierarchical conceptual relationship information expressing the relationship when there is a conceptual hierarchical relationship between the elements expressed by each of the plurality of important phrases extracted from the invention document by the important phrase extraction unit. It is also possible to include.
As a result, translators who view the important word-related information in the patent application-related information output by the invention document analysis system can efficiently understand the relationships between multiple important words in the invention document that was analyzed by the invention document analysis system. Therefore, it may be possible to efficiently or appropriately determine translation words corresponding to each of the plurality of important words.

Further, for example, the setting information acquired by the setting information acquisition unit includes comment specifying information that is information for specifying a comment line or comment, and the analysis unit analyzes the claim information in the patent application related information. a claim tree generation unit that generates a claim tree expressing the citation relationship of each term of the claim expressed in a tree structure format, the patent application related information includes the claim tree, and the claim tree generation unit For each claim item in the claim information, write information on a new line in descending order of item number, and information indicating the item number of the item that the item cites, using a symbol for expressing the citation relationship and the item number. generate a claim tree, and for each item related to each line of the claim tree, if the line immediately before the item in the claim information is a comment line in a predetermined format specified based on the comment identification information, , the contents of the comment line are added to the line that describes the item number of the item in the claim tree, and if the line immediately before the item in the claim information is not the comment line, the same line as the heading of the item is added. If a comment is written in a predetermined format specified based on the comment identification information, the content of the comment is added to the line in which the item number of the section in the claim tree is written, and If the line immediately preceding the relevant section in the information is not the comment line, and the comment is not stated in the same line as the heading of the relevant section, the important word extraction that is stated in the relevant section of the claim of the invention document When there are one or more important words extracted by the section, at least one of the one or more important words is added to the line in which the item number of the section in the claim tree is written. Good too.
As a result, translators can check comments that inventors or patent attorneys have included in claims of invention documents using the claim tree in the patent application related information output by the invention document analysis system based on the invention document. , it may be possible to efficiently or accurately understand the significance of each clause by commenting on the citation relationship between each clause of the claim. In addition, for sections in which no comments are written regarding the claim section of an invention document, the important words and phrases of that section are shown in the claim tree in the patent application related information output by the invention document analysis system based on the invention document. Therefore, it is possible that a translator, an inventor, etc., or a patent attorney who has forgotten some of the content of the invention document due to the passage of time since its creation may be able to understand the meaning of that section of the claim. If a translator fully understands the meaning of each section of a claim, he or she will be able to translate each section of the claim efficiently or appropriately.

Further, for example, the claim tree generation unit generates a constituent claim tree that expresses the citation relationship of each term of the claim expressed in the claim information in the patent application related information in a tree structure format, and The application-related information includes the constituent claim tree, and the claim tree generation unit rearranges the rows of each claim section in the claim information in descending order of item number, and generates the item number of the section that the section cites. A constituent claim tree is generated in which information indicating information is expressed using symbols and item numbers for expressing citation relations, and for each term related to each line of the constituent claim tree, the important words and phrases described in the relevant clause in the claim information are extracted. When there are one or more important words and phrases extracted by the section, each of the important words and phrases is added to the line in which the section number of the relevant section is written, and if there are multiple important words and phrases that should be added to the relevant section, those multiple important words and phrases are added. If there is a structural inclusion relationship between the elements represented by each important word or phrase, a description expressing the relationship between the important words may be added to the line.
As a result, by viewing the constituent claim tree in the patent application related information output by the invention document analysis system based on the invention document, the translator or the translator can easily avoid forgetting the content of the invention document as time has passed since its creation. The inventor, etc. or patent attorney who created the invention can understand the citation relationship between each claim section in the invention document, as well as the relationship between multiple important words and phrases stated in the claim section. There is a possibility that the meaning of the terms can be understood efficiently.

Further, for example, the invention document analysis system may analyze characters that correspond to the inappropriate character string pattern indicated by the inappropriate character string pattern information, based on correction information associated with predetermined inappropriate character string pattern information. column exists in a claim in an invention document acquired by the document acquisition unit, the content of the revised claim is added to the patent application related information by modifying the claim based on the modification information. The claim tree generation unit includes modified claim information generated as shown in FIG. In response to the operation of the display element, display control information is added to provide a GUI function for displaying the section of the claim information in the display document with the corresponding section number, and the claim tree or In the display element for the symbol for expressing the citation relationship of the constituent claim tree, in response to the operation of the display element, a citation relationship cutting process for disconnecting the citation relationship represented by the symbol. Display control information is added to provide a function as a GUI, and the citation relationship disconnection process is a process of updating the content of the revised claim information in the display document to reflect the disconnection of the citation relationship. , a process of changing the symbol for expressing the citation relationship as the display element to a predetermined symbol to express that the citation relationship has been severed, and redisplaying the display element. It may also be a thing.
As a result, inventors or patent attorneys can, for example, prepare invention documents for application in Japan, and then decide whether to apply for foreign applications based on application strategies in each country, differences in legal systems related to patents, etc. If you want to use the invention document as a Japanese text for translation by changing the claim organization, you can use the claim tree or claim tree of patent application related information as a display document output by the invention document analysis system based on the invention document. By operating a part of the GUI of the constituent claim tree, it is possible to efficiently remove the quotation from one claim clause to another clause, and generate or generate revised claim information that reflects the cancellation of the quotation. Can be updated. By subsequently acquiring and viewing the patent application related information as a display document, the translator can perform translation for foreign applications, etc. based on the revised claim information. Further, the inventor or the patent attorney may, for example, copy and paste the displayed content of the revised claim information after operating the GUI, hand over a document generated to a translator, and request translation of the document.

Further, for example, the invention document analysis system may represent important words corresponding to codes among the important words extracted from the invention document by the important words extracting unit, by associating the code with the important words. The invention document analysis system generates a code list that is a list and includes it in the patent application related information, and the invention document analysis system enumerates codes written subsequent to the important words extracted from the invention document by the important word extraction unit. and expresses the symbol in association with the corresponding important phrase, and if there is no description in the invention document in which the symbol is added to the important phrase extracted by the important phrase extracting section, the important phrase is When an element expressed by a word or phrase structurally includes an element expressed by another key word or phrase, or is structurally included in an element expressed by another key word or phrase, the structural inclusion of the key word or phrase and the other key word or phrase; The code list may be generated by determining a code based on the relationship and expressing the determined code in association with the important word/phrase.
As a result, the invention document analysis system determines whether the invention document includes a description with a code added for each of the plurality of important words and phrases in which there is a structural inclusion relationship between the elements represented by each of the plurality of important words and phrases in the invention document. If there is no code for the important term, the code for the important term is automatically determined based on the inclusion relationship, and a code list that associates the code with the important term is generated. etc. or a patent attorney may be able to efficiently or properly complete an invention document. For example, an inventor or a patent attorney can efficiently invent an invention by creating an invention document without writing codes and inserting the necessary codes into the invention document based on the code list generated by the invention document analysis system. It may be possible to complete the document.

Further, for example, the document acquisition unit further acquires a drawing document corresponding to the invention document acquired by the document acquisition unit, and generates a drawing document based on an analysis result of the invention document acquired by the document acquisition unit. Among the plurality of important phrases extracted from the invention document by the important phrase extraction unit in the claim information or the specification information of the patent application related information, the element name described with a code added in the invention document. With respect to an important word or phrase that is the same as a word that represents the important word or phrase, the display control information related to the important word or phrase changes the key word or phrase according to the operation so that the code related to the important word or phrase in the drawing document corresponding to the invention document corresponds to the key word or phrase. It may also include information for displaying the described portion as a GUI element.
As a result, the translator can use the GUI related to the same important words and phrases in the claim information or specification information of the patent application related information as the display document that include the description with the code added in the invention document. By operating on the key word, you can quickly check the part of the drawing where the code corresponding to the key word is written, making it possible to understand the key word efficiently. The possibility that a person can efficiently or appropriately translate an invention document may be increased.

Further, for example, the document acquisition unit acquires a plurality of invention documents and a plurality of translation result documents that are translation results of each of the plurality of invention documents, and the invention document analysis system analyzes the inventions to be translated. The invention document standard provision acquisition unit for translation acquires the invention document standard provision for translation that defines a plurality of conditions that the document should satisfy, and the invention document A determination is made as to whether at least some of the plurality of conditions specified in the standard provisions for invention documents for translation are satisfied, and the conditions for which the determination has been made are made among the plurality of invention documents. a document evaluation unit that selects an invention document whose satisfaction exceeds a predetermined degree; and a set of the invention document selected by the document evaluation unit and a translation result document corresponding to the invention document as training data. a machine learning unit that generates a learning model for translation, which is a learning model for machine translation, by machine learning that is based on the machine learning, and the output unit includes a learning unit that outputs the learning model for translation generated by the machine learning unit. It may also include a model output section.
As a result, invention documents with relatively high conformity to the standard provisions for invention documents for translation are selected, and the pair with the corresponding translation result document is used as training data, so invention documents containing inappropriate descriptions are Due to translation errors generated based on the inappropriate invention document and the inappropriate invention document, the combination of the inappropriate invention document and the translated document with inconsistent content becomes difficult to use as training data, and the invention document analysis system In this way, an appropriate translation learning model can be generated. Furthermore, according to a system (for example, invention document analysis system 20, etc.) that machine-translates an invention document through arithmetic processing using the generated translation learning model, a certain degree of appropriate translation results can be obtained, so that translators can By referring to the translation results and making appropriate corrections, it may be possible to translate the invention document efficiently or appropriately.

The display document according to the present disclosure, which is output by the above-mentioned invention document analysis system (for example, a system that can be realized by causing a device equipped with a microprocessor to perform invention document analysis processing by executing an invention document analysis processing program) (In other words, a display document that includes patent application related information to be displayed) is data that may include a script as a computer program, and is data that includes a document written in a markup language, and is data that can be used by an invention document analysis system. The generated patent application related information is included as content to be displayed. The display document according to the present disclosure can be recorded on a computer-readable non-transitory recording medium such as a ROM, an optical disk, or a hard disk, and then distributed or circulated. When the display document according to the present disclosure is distributed, the display document is subject to interpretation by a browser etc. installed on a device equipped with a microprocessor, and as a result, the display document is not generated by the invention document analysis system described above. Information related to patent applications will be displayed. Translators and the like can efficiently understand the invention document by viewing the patent application related information in the display document generated based on the invention document. In addition, by viewing the displayed patent application related information, the translator can, for example, efficiently or appropriately interpret the claims and specifications of the invention document, which are respectively represented by the claim information and specification information included in the patent application related information. It may be possible to translate it into In addition, by viewing and referring to the displayed patent application related information, inventors, etc. or patent attorneys can, for example, view the claims and specifications of invention documents represented by the claim information and specification information included in the patent application related information, respectively. The specification may be updated with appropriate content, for example to make it easier to translate.

The display document in this disclosure includes, for example, a document written in a markup language, is subject to interpretation by a browser, etc. in a device equipped with a microprocessor, and is an invention for patent application that expresses a technical idea in Japanese. A display document that includes patent application-related information as content to be displayed, including claim information representing the contents of the claims in the document and specification information representing the contents of the specification in the invention document, the display document including the claims. The claim information representing the contents of the claim controls the display mode of the display element of each important word extracted based on predetermined important word extraction conditions among the words of the claim. The display control information relating to each of a plurality of important words and phrases in which the character strings constituting the important words are different from each other among the important words and phrases represented in the claim information is This is to control each display element to be displayed in a display mode with different display colors added, and among the important words and phrases expressed in the claim information, the character strings forming the important words are different from each other. The display control information related to an important phrase that is composed of the same character string as any one of the plurality of different important phrases is configured to change the display element of the important phrase to the display of the one important phrase. This is to control the display in a display mode with the same display color as the display color added as the display mode of the element, and the specification information representing the contents of the specification is Among the words and phrases, for each important phrase extracted based on the above-mentioned conditions for extracting important words, display control information for controlling the display mode of the display element of the important phrase is included, and the information is displayed in the statement information. The display control information relating to an important phrase that is the same as any one of the plurality of important phrases in which the character strings constituting the important phrases in the claim information are different from each other, The display element for the important term is controlled to be displayed in a display mode with the same display color as the display color added as the display mode for the display element for the one important term in the complaint information. , among the important words and phrases represented in the specification information, the display control information relating to the important words and phrases that are different from any of the plurality of important words and phrases in which the character strings forming the important words in the claim information are different from each other, This is a display document for controlling a display element of an important term to be displayed in a display mode different from any of the display elements of a plurality of important terms in the complaint information.
As a result, when patent application related information is displayed on a browser etc., multiple important words and phrases consisting of mutually different character strings extracted from the claims of an invention document are displayed in the claim information or specifications in the patent application related information. An important term in claim information or specification information that is displayed in a display mode with different display colors and is composed of the same character string as any one of the plurality of important terms. is displayed in a display manner in which the same display color as that one important word is added. Therefore, translators who view patent application related information displayed on a browser can quickly identify the same important words and different important words from each of the multiple important words using the added display color. , it may be possible to efficiently understand the content of claim information or specification information. For translators, efficient understanding of the content of claim information or specification information is useful for efficient translation of claims and specification. Therefore, a translator who has obtained a display document containing patent application-related information from the inventor or patent attorney who created the invention document will have the browser interpret the display document and display the patent application-related information. There is a possibility that invention documents can be translated efficiently by viewing the . In addition, a translator who has obtained an invention document from an inventor or patent attorney who created the invention document can patent a patent based on the invention document using the invention document analysis system (for example, invention document analysis system 20, etc.) in this disclosure. It is possible to efficiently translate invention documents by generating a display document containing application-related information, having the browser interpret the display document, displaying the patent application-related information, and viewing its contents. .

Further, for example, the display document is an HTML document including an HTML file in which the patent application related information is described in HTML (Hypertext Markup Language), and the conditions for extracting important words are as follows: In the claim or the specification, a character that is a candidate for an important word follows a first character string that corresponds to a reference character string pattern that represents a pattern of a reference character string that is used to refer to a character string previously described. A first extraction condition for extracting a character string that is the same as the second character string as an important word when there is a second character string that corresponds to the important word candidate character string pattern representing the pattern of the column; , a second extraction condition for extracting a character string that is the same as the third character string as an important word when there is a period character following the third character string that corresponds to the important word candidate character string pattern, and the claim Or, in the specification, a code string representing a pattern of code character strings that is added to the fourth character string corresponding to the important word candidate character string pattern and subsequent to the character string representing the element name. a third extraction condition for extracting a character string that is the same as the fourth character string as an important word when a fifth character string that corresponds to the pattern exists, and the important word candidate character string in the claim or the specification; If the sixth character string that corresponds to the pattern is followed by a seventh character string that corresponds to a specific particle string pattern that represents a pattern of character strings that are predetermined particles, then it is the same as the sixth character string. A fourth extraction condition for extracting a character string of as an important word, the first extraction condition, the second extraction condition, the third extraction condition, and the fourth extraction condition, the important word candidate character In the case where a plurality of character strings correspond to a column pattern, if one of the plurality of character strings is included in another character string in the statement of the claim or the specification, the plurality of character strings and a selection extraction condition for selecting and extracting important words from the remaining character strings after excluding the one character string, and the important word candidate character string pattern includes at least a first character. A predetermined character string that is a kanji character and a predetermined character string whose first character is katakana are applicable, at least the character string "said" is applicable to the reference character string pattern, and the code string pattern is a predetermined character string. corresponds to at least a predetermined character string whose first character is a number, and the specific particle string pattern includes at least "wo", "ni", "ga", "to", "kara", " The display control information relating to each of the plurality of important words and phrases to which any of the characters ``to'' and ``ha'' correspond, and in which the character strings constituting the important words in the claim information are different from each other, This is for controlling each display element to be displayed in a display mode in which a left border line having a predetermined width is added in a display color that differs for each important term, and for each important term in the specification information. , if the important phrase is the same as any one of the plurality of important phrases in the claim information, the display control information related to the important phrase in the specification information Displaying the display element in a display mode in which a left border line having the predetermined width is added in the same display color as the display color of the left border line according to the display mode of the display element of the one important phrase in the claim information. If the important word/phrase in the specification information is different from any of the plurality of important words/phrases in the claim information, the display control information regarding the important word/phrase in the specification information The display element may be controlled to be displayed in a display mode in which the left border line having the predetermined width is not added.
As a result, translators who view patent application related information as a display document displayed on a browser etc. can add different important words included in claim information to each important word in the patent application related information. By identifying the display color of the left border line, you can quickly distinguish between the two. In addition, there is a method of expressing different important words by using different display colors of the left border added to the important words. Regardless of whether the background color of the area is dark (e.g. black) or light (e.g. white), it has the advantage of not reducing the readability of the important words themselves.For example, if the basic background color is set to white or This is useful in cases where the color can be changed to black or the like, or in cases where the display document is equipped with a GUI that allows the background color to be selected. In addition, the left border that can be added to some display elements in HTML documents is not a character, so it will be ignored when copying and pasting the display content text that includes important words. It does not get in the way when copying and pasting into a text editor etc. to edit descriptions that include important words.

Further, for example, the predetermined width is the width of one character, and the patent application related information includes a plurality of important words and phrases in which the character strings forming the important words in the claim information are different from each other. A new line is added to include the enumerated important word list, and for each important word in the important word list, display control information for controlling the display mode of the display element of the important word is added. The display control information related to the important word/phrase may be the same as any one of the plurality of important words/phrases in the claim information in which the character strings forming the important word/phrase are different from each other. For example, a left border line having the predetermined width is added to the display element of the important word/phrase in the same display color as the left border line according to the display mode of the display element of the one important word/phrase in the claim information. It may also be for controlling the display in a different display mode.
As a result, translators and others who view patent application related information as a display document will be able to distinguish between important words and phrases listed in the claim information from among the important words and phrases in the list of important words. It may be possible to quickly identify key words in the key word list that correspond to each key word in the key word list. Also, if the left border line has a width of one character (for example, a width represented by 1em), a rectangle filled with the display color of one character is displayed immediately before each of the plurality of important words in the complaint information. Therefore, even if a rectangle for one character is mixed in a character string, it does not easily disturb the arrangement of each character in the complaint information, so the effect is that the visibility of each character is not easily obstructed. For example, if a monospaced font is used as the display font for each character in the claim information in the display document, the placement of each character may change if a rectangle for one character is mixed in the character string. Even if a font that is not monospaced is used to display each character in the claim information, the width of a rectangle for one character is Since there is not much difference in width, when viewing the claim information, the arrangement of each character will not be disturbed and visibility will not be significantly hindered. It should be noted that the visibility of characters in patent application-related information is good, or the ability to quickly identify important words with different character strings makes it possible to efficiently understand the contents of claim information and specification information. obtain. For translators, efficient understanding of claim information and specification information is useful for efficient translation of claims and specifications of invention documents.

Furthermore, for example, each row of the important word/phrase list includes one important word and a corresponding translation word corresponding to the important word/phrase, or a translation word candidate set that is a set of candidates for corresponding translation words for the important word/phrase. The display control information related to the important word in the important word list changes the display element of the translation word candidate set for the important word to any element of the translation word candidate set according to the operation. This function is used to control functions as a graphical user interface (GUI) that enables selection of the selected translation word candidate set according to operations on display elements of the set of translated word candidates while functioning as the GUI. The elements of the translated word candidate set that have been selected are determined as the corresponding translated words corresponding to the important word/phrase, and the contents of the display elements of the translated word candidate set are replaced and displayed with the determined corresponding translated words. It may also include a script for executing the process.
As a result, translators who have viewed patent application related information as a display document can use the GUI to select a set of translation word candidates that correspond to important words in the list of important words in the patent application related information displayed on a browser, etc. By selecting any element of the translation word candidate set, the corresponding translation word corresponding to the important word in the important word list can be easily determined and displayed. In addition, the translator etc., for example, after determining the corresponding translation word corresponding to the important word by selecting any element of the translation word candidate set, copy and paste the important word list displayed on the browser etc. By pasting it into a text editor etc. and editing it as necessary, the invention document analysis system (for example, invention document analysis system 20, etc.) in this disclosure can extract important information from the invention document that is the basis for generating patent application related information. It is possible to relatively easily generate an important word/phrase setting dictionary that can be used to extract words (for example, an important word/phrase setting dictionary that can be acquired and used by the important word/phrase setting dictionary acquisition unit 203).

Further, for example, the display control information related to the important word/phrase in the patent application related information may be configured to display the display element of the important word/phrase on a search site on the network, such as the web page containing the important word/phrase, the It functions as a GUI that executes the process of searching for definitions of important words, synonyms of the important words, or images related to the important words, or the process of having the important words translated by a translation site on the network. It may also be a thing.
As a result, translators, etc. who viewed patent application related information related to display documents displayed on a browser etc. can perform operations on important words in the displayed patent application related information (in other words, operations on display elements of important words). , it may be possible to obtain the search results for the key words (e.g., web pages containing the key words, definitions of the key words, synonyms, related images), or translation results of the key words, so that information related to the patent application can be obtained. It may be possible to efficiently or properly understand important words and phrases in the invention document that served as the basis for the generation of information. Therefore, the translator may be able to efficiently or appropriately determine translation terms for important terms.

Further, for example, the display control information related to the important term in the patent application related information may change the display element of the important term to display at least part of the attribute information of the important term, or It functions as a GUI for displaying the importance of the important word/phrase determined based on the attribute information of the important word/phrase, and the attribute information of the important word/phrase indicates the importance of the important word/phrase in the claim represented by the claim information. If there is a claim section that defines the term, information indicating the section number of that section, and if there is a claim section that refers to the important phrase in the claim represented by the claim information, the information of that section. The information may include information indicating an item number and, if there is a description in which a code is added to the important term in the specification represented by the specification information, information indicating the code.
As a result, translators who view patent application related information related to display documents displayed on a browser, etc., can obtain information on the importance level or attribute information of the displayed important words by operating on the displayed important words. As a result, it becomes possible to recognize to some extent which important words and phrases in the invention document that were the basis for the generation of the patent application related information, and it is possible to efficiently or properly understand the important words and phrases. It could be possible. Therefore, the translator may be able to translate efficiently by, for example, preferentially determining translation terms for important key terms.

For example, the display control information related to the important word/phrase in the patent application related information may change the display element of the important word/phrase to a line containing the important word/phrase from the claim information and the specification information according to the operation. It functions as a GUI for executing the enumeration process of inclusion of important words to be enumerated, and the enumeration process of enumeration of descriptions that include important words is performed by selecting the line containing the important term to be operated from the claim information and the specification information. The important words and phrases in each line listed and displayed are highlighted, and the line in the claim information or specification information is displayed in the displayed line, depending on the operation on the line. It may also be a process of adding display control information for making the line function as a GUI for displaying links to the multiple lines included.
As a result, a translator, etc. who viewed patent application related information related to a display document displayed on a browser, etc., can perform operations on important words in the displayed patent application related information, and as a result of the important word inclusion description enumeration process. You can check the lines containing the important words in the claim information and specification information that are listed and displayed. Therefore, the translator or the like may be able to efficiently or appropriately understand important words and phrases in the invention document that served as the basis for generating the patent application related information. Moreover, for this reason, the translator may be able to efficiently or appropriately determine translation words for important words. In addition, the translator or the like can accurately identify the important word/phrase because the important word/phrase is highlighted in the line enumerated by the important word/phrase inclusion description enumeration process by including the manipulated important word/phrase. In addition, translators, etc. can perform further operations on one of the lines enumerated by the important word inclusion description enumeration process by including the manipulated important word, so that it will be further displayed. Multiple lines in claim information, specification information, or another invention document that precede and follow one line can be easily confirmed.

Further, for example, the patent application related information includes claim translation information representing the result of machine translation for the claim represented by the claim information, and machine translation result for the specification represented by the specification information. A line or a paragraph in the claim translation information that represents the result of machine translation that targets a line or paragraph of the claim information is a line or paragraph that represents the result of machine translation of a line or paragraph of the claim information. Replace the complete translation line expressed in the target language, or the important words included in the line or section of the claim information, with the corresponding translated word corresponding to the important word or phrase, or a set of translation word candidates for the important word. This is a partial translation line that is the result of machine translation, and a line that represents the result of machine translation of one line of the specification information in the specification translation information is a partial translation line that is the result of machine translation of one line of the specification information. This is the result of replacing the complete translation line represented by , or the important word/phrase included in the relevant specification information line, with the corresponding translated word corresponding to the important word/phrase, or a set of translation word candidates for the important word/phrase. The claim translation information and the specification translation information are partial translation lines, and the claim translation information and the specification translation information are such that the character strings constituting the important words in the claim information are the same as any one of the plurality of important words and phrases that are different from each other. Regarding important phrases, it includes display control information for controlling the display mode of display elements of machine translation results for which the important phrases are subject to machine translation, and the display control information regarding the display elements of the machine translation results is: Display of the display element so that the same display color as that added to the display element of the important word in the claim information and the specification information is added to at least a part of the display mode of the display element. It may be for controlling the aspect.
As a result, for example, a translator who viewed patent application related information related to a display document displayed on a browser etc. can add important words and phrases in the patent application related information and the results of machine translation of important words. By identifying the displayed display color, it is possible to quickly understand which important words in claim information or specification information the result of machine translation of important words in claim translation information or specification translation information corresponds to. It becomes possible.

For example, the claim translation information includes each line of the claim information, and for each line or item of the included claim information, the content of the line or item is immediately followed by the content of the line or item. The complete translation line or the partial translation line representing the result of machine translation corresponding to the machine translation target is inserted, and the specification translation information includes each line of the specification information. For each line of specification information, the complete translation line or partial translation line representing the result of machine translation corresponding to the machine translation target is inserted immediately after the line, and the claim translation information And the specification translation information includes a display mode of display elements of at least a part of the machine translation target in the claim translation information and the specification translation information, and a display of at least a part of the result of machine translation corresponding to the machine translation target. Display control is applied to each display element in order to control the display mode of the element to be different in at least one of the following: text background color, text color, text decoration, font size, font style, and font weight. Information may also be added.
As a result, translators who view patent application related information as a display document displayed on a browser etc. can check the lines to be machine translated and the machine translation target line in the claim translation information and specification translation information in the patent application related information. Translation result lines can be quickly identified by differences in at least one of character background color, character color, character decoration, font size, font style, and font weight. In addition, while visually confirming the line to be machine translated in the claim translation information and specification translation information, the translator should check the contents of the machine translation result line placed after the line to be machine translated as appropriate. It may be possible to modify and efficiently create a more appropriate translated sentence.

Further, for example, the patent application related information includes warning information, and the warning information indicates that the claim or the specification of the invention document meets a plurality of conditions specified in predefined invention document standard regulations for translation. Indication information that indicates whether the conditions are met and points out if the conditions are not satisfied, and the claims or the specification of the invention document meet the multiple conditions specified in the standard provisions for invention documents for translation. Inventions that indicate the evaluation value of the conformity of the invention document to the invention document standard provisions for translation, which is calculated by a predetermined calculation based on the results of testing whether at least some of the conditions are satisfied. The document evaluation information may include document evaluation information, regulation access information for accessing the standard provisions for invention documents for translation, and version information indicating a version of the standard provisions for invention documents for translation.
As a result, by referring to the warning information (for example, pointed information in the warning information, etc.) in the patent application-related information as a display document displayed by a browser, etc., it became the basis for generating the patent application-related information. The inventor, etc. or patent attorney who prepared the invention document may amend the invention document if necessary. In addition, the translator can, for example, receive a display document related to patent application-related information from an inventor or a patent attorney, so that even if the warning information in the patent application-related information remains, the translator can point out the problem. It can be assumed that this is the result of the inventor or patent attorney's judgment that there is no need to take any particular action, so it is possible to understand the appropriate content of the translation target, and as a result, it is possible to efficiently translate the invention document. Or it may be possible to do it properly. Furthermore, by referring to the regulation access information in the warning information, the translator etc. can access the standard regulations for invention documents for translation and view their contents. Furthermore, by referring to the version information of the standard provisions for invention documents for translation in the warning information, the translator etc. can confirm which version of the standard provisions for invention documents for translation has been used.

In addition, the invention document analysis processing program according to the present disclosure allows a device equipped with a microprocessor to analyze an invention document that is a document that describes claims and specifications for a patent application expressing technical ideas in Japanese. An invention document analysis processing program for generating a display document as described above by performing an analysis process, wherein the invention document analysis process includes a document acquisition step for acquiring an invention document, and a document acquisition step. an analysis step of analyzing the obtained invention document by extracting important words based on predetermined conditions for extracting important words; and an analysis step of analyzing the invention document obtained by the document obtaining step. The invention document analysis processing program includes an output step of generating and outputting the display document for displaying the patent application related information generated based on the analysis result.
By installing this invention document analysis processing program into a device or the like having a microprocessor, the device or the like can realize the invention document analysis processing. The invention document analysis processing realized by this invention document analysis processing program generates a display document as described above, and the display document is interpreted by a browser etc. to obtain patent application related information related to the display document. Browsing may facilitate understanding of invention documents for translators and the like. Facilitating the understanding of invention documents will be useful for translators to appropriately or efficiently translate invention documents.

Here, for example, the invention document analysis process may display a GUI screen that enables a selection operation of any one of the one or more display documents that have been output in the output step, and select the selected display document. The method may include processing for making the document subject to interpretation by a browser in the device.
As a result, by having a device equipped with a microprocessor execute an invention document analysis processing program and selecting one of the display documents generated in the past as a result of the execution, patent application related information related to the display document can be obtained. will be displayed by the browser. Therefore, when there are multiple display documents generated as a result of execution of the invention document analysis processing program, translators, etc. can compare and view the patent application related information related to each of the multiple display documents. etc., there is a possibility that patent application related information related to a display document generated based on a specific invention document among the plurality of display documents can be efficiently understood.

The above-mentioned invention document analysis system and related technology may be applied to various devices, devices, integrated circuits, systems, methods, programs (i.e., computer programs), computer-readable recording media, etc., in whole or in part. It can be embodied in various ways.

Hereinafter, in order to facilitate understanding of the technology in the present disclosure, embodiments will be illustrated and described in detail with reference to the drawings as appropriate. However, descriptions of well-known techniques will be omitted as appropriate. Note that the drawings are schematic diagrams and are not strictly illustrated. Further, the number, arrangement, shape, attribute, state, numerical value of various elements in the description of this embodiment and the drawings, or the connection manner, execution order, inclusion relationship, etc. between elements are merely examples, and the scope of the claims This does not limit the scope of claims described in . Among the various elements in the description of this embodiment and the drawings, elements not stated in the independent claims can be added arbitrarily.

(Embodiment 1)
An example of an embodiment of an invention document analysis system will be described below with reference to the drawings.

[1. Information related to the embodiment of the invention document analysis system]
The invention document analysis system described in the embodiments of this disclosure is a system that executes invention document analysis processing for analyzing invention documents.

In the present disclosure, an invention document is a document that expresses a technical idea in Japanese and describes claims or specifications for a patent application. By the way, in order to obtain patent rights (for example, patent rights) in regions other than Japan (for example, sending documents written in a foreign language other than Japanese to a foreign country other than Japan) , for patent applications, etc.), a translator may translate an invention document created by an inventor or the like to create a translation result document. The translation result document is a document that mainly includes at least one of claims and specifications written in a foreign language. One example of the foreign language used to describe the translation result document is English, and another example is Chinese. Hereinafter, for convenience, English will be used as an example of a foreign language used to describe the translation result document, but this is not intended to limit the language used to describe the translation result document to English.

An example of the purpose of creating the invention document analysis system shown in the embodiments of this disclosure is to enable translators to efficiently or efficiently translate translated documents based on invention documents in order to acquire useful rights in regions other than Japan. The purpose is to provide a useful system for creating properly.

[1-1. Invention document]
An invention document is a document that expresses a technical idea in Japanese, and also contains claims or specifications for a patent application. Here, a document means document data, and the contents of a document may include descriptions of one or more documents. Note that the invention document may be a document that includes both claims and specifications. For example, an invention document is composed of multiple lines. Each line is a part of the document data separated by line feed codes.

In terms of language, an invention document, which is a document expressing a technical idea in Japanese, only needs to be written mainly in Japanese, and words or phrases in a foreign language may be used partially. A claim is, for example, the content described in the document "Claims" in a patent application in Japan, or the content described in one or more "claims" in a patent application in Japan. Note that even if a document is configured in a specific format that is different from the description format of a patent application in Japan, it is possible to write a statement that substantially corresponds to the contents of the "claims" or "specification" in a patent application in Japan. By converting the specific format, the containing document can be treated as a document containing the "claims" or "specification" of a patent application in Japan. I think so. For example, a claim may be the content described in the document "Claims" in an international application (that is, an international application based on the Patent Cooperation Treaty).

An example of an invention document is a document of a patent application in Japan that expresses the technical idea in Japanese, including claims and specifications, regardless of whether it is filed before or after the application. Further, an example of an invention document is an unfinished document in the preparation stage of a patent application in Japan that expresses a technical idea in Japanese, including at least one of claims and specifications. Also, as an example of an invention document, a foreign patent application (that is, a patent application filed in a region other than Japan) that expresses the technical idea in Japanese, including a description corresponding to at least one of the claims and the specification. documents or documents in the stage before being translated into a foreign language as a foreign patent application (for example, a Japanese manuscript for translation). Furthermore, an example of an invention document is a document such as an invention memo that expresses a technical idea in Japanese, including a description corresponding to at least one of a claim and a specification. Furthermore, an example of an invention document is data that substantially corresponds to the contents of a patent publication (for example, a patent publication publication, a patent publication publication, etc.) published by a Japanese or foreign government agency or international bureau.

[1-2. Work related to translation of invention documents]
An example of the flow of work related to the translation of invention documents will be described below.

FIG. 1 shows an example of a work flow related to the translation of an invention document. As shown in the figure, the inventor first creates an invention memo that describes the content of the technical idea he created, organizes the content of the invention memo, and essentially creates a claim or specification for a patent application. Create an invention document that describes the The inventor, etc. (e.g., the inventor, or a person related to the inventor, such as an applicant who has been assigned the right to obtain a patent from the inventor, or a collaborator in the organization to which the inventor belongs) must consult a patent attorney to prepare the invention document. Sometimes I leave it to them. Note that the inventor, etc. or patent attorney may also create a drawing document containing drawings for patent application.

When an inventor, etc. or a patent attorney requests a translator to translate an invention document, the inventor, etc. or patent attorney may, as necessary, specify corresponding translation words for important words (i.e., so-called keywords) in the invention document, or specify the rights to the translation. Special translation instructions may be given depending on the region in which the application is to be obtained (for example, in response to a strategy that takes into account the laws and precedents of the foreign country to which the application is being applied). Important words and phrases in invention documents written primarily in Japanese (i.e., documents expressing technical ideas in Japanese) are one or more words (i.e., character strings consisting of one or more characters), and are not written in Japanese. It may be one or more words. The translator creates a translation result document by translating the invention document, specifying corresponding translation words for important words, special translation instructions, etc.

FIG. 2 shows an example of the flow of translation work by a translator. As shown in the figure, for example, the translator first understands the outline of the invention document, and if corresponding translation words have been specified, uses this as a reference to determine the corresponding translation words for important words and phrases in the invention document. . Then, the translator proofreads, changes, etc. the invention document as necessary to make it easier to translate the invention document. Then, the translator translates the invention document after proofreading, changes, etc., and creates a translation result document. When translating, a translator can, for example, have a machine translation system or the like perform machine translation, and create a translation result document by utilizing the results of the machine translation. Finally, the translator reviews the translation result document as a translation result and makes corrections to the translation result document as necessary.

[1-3. Consideration of factors that impede the efficiency or appropriateness of translation for acquiring useful rights]
FIG. 3 is a diagram listing examples of factors that impede efficiency or appropriateness of translation for acquiring useful rights.

If the invention document is inadequate, for example, it may be difficult for translators who are not technical experts to understand the invention document or to decide how to modify and translate the invention document. Become. As a result, it takes time to conduct research to understand the invention document, to communicate with the creator of the invention document for confirmation, etc., and the translator may not be able to translate efficiently. Furthermore, if a translator misunderstands an inappropriate invention document, there is a possibility that the translator will not be able to perform a proper translation.

As shown in Figure 3, invention documents that are inappropriate for acquiring useful rights include, for example, those that contain typographical errors, inconsistent descriptions, ambiguous expressions, relative expressions, etc., and words that do not have corresponding appropriate translated words. These include sentences that are unclear in their exact meaning, statements that do not formally comply with foreign laws and precedents, and statements that may pose problems in enforcing patent rights. An example of a relative expression is the description "heavy first member." For example, if a relative expression is included, the translator may be confused as to how to translate the relative expression in order to obtain useful rights. An example of a sentence whose exact meaning is not clear is a sentence that lacks a subject, making it difficult for anyone but a technical expert to determine what the subject is. An example of a description that does not formally conform to foreign laws and precedents is that under the conditions that a translated document is required to obtain a patent right in a specified foreign country, the document is subject to patent protection in that specified foreign country. An example of this is a statement of a claim that is directed to something that is not covered by the United States (for example, a program in the United States). Note that descriptions that may pose a problem in the enforcement of patent rights do not necessarily pose a problem, but they do have the potential to become a problem. Examples of statements that may pose problems in enforcing patent rights include "Preferably..." and "Features of the present invention" using the word "invention" with the meaning of the claimed subject matter in the United States. is...'', etc. Translators may be at a loss as to whether to modify such descriptions in order to obtain useful rights.

Special translation instructions that are inappropriate for obtaining useful rights include special translation instructions whose specific work content is unclear. An example of a special translation instruction in which the specific content of the work is unclear is an instruction to resolve multiple dependent claims for which the revised content cannot necessarily be determined uniformly. Special translation instructions with unclear specific work content can hinder efficient translation.

Additionally, when a translator uses machine translation results from a machine translation system, etc., if the translation accuracy of the machine translation is not sufficient, the translator may need a large amount of time to review the machine translation results. .

[1-4. Example of use of embodiment of invention document analysis system]
In an embodiment of the invention document analysis system according to the present disclosure, for example, information indicating deficiencies in invention documents, evaluation, etc. can be output. Further, in one embodiment of the invention document analysis system, for example, information that supports determination of corresponding translation words of important words and phrases of an invention document can be outputted. Further, in one embodiment of the invention document analysis system, for example, information useful for improving the accuracy of machine translation of invention documents can be output.

Therefore, for example, an inventor or a patent attorney who has created an invention document can utilize an embodiment of the invention document analysis system in the present disclosure and utilize output based on the analysis of the invention document by the embodiment. This may provide an opportunity to review inappropriate invention documents and revise them appropriately.

For example, an inventor, etc. or a patent attorney who has created an invention document may use an embodiment of the invention document analysis system in the present disclosure, and information regarding output results based on the analysis of the invention document by the embodiment (for example, output When providing the translator with the result itself or information that allows the translator to obtain a similar output result, the translator can perform the translation work (see Figure 2) efficiently or appropriately. there is a possibility. This is because the translator can easily resolve any confusion that may arise when translating in order to obtain useful rights by using the provided information regarding the output results based on the analysis results of the invention document.

For example, a translator who receives an invention document etc. from an inventor or patent attorney uses an embodiment of the invention document analysis system in this disclosure, and utilizes the output based on the analysis of the invention document by the embodiment. By doing so, there is a possibility that the translation work (see FIG. 2) for the invention document can be performed efficiently or appropriately. Furthermore, when using a machine translation system or the like, translators can improve the accuracy of machine translation by utilizing information output by an embodiment of the invention document analysis system of the present disclosure. 2) may be performed efficiently or appropriately.

[2-1. Configuration example of invention document analysis system according to Embodiment 1]
FIG. 4 shows an example of the configuration of the invention document analysis system 10.

The invention document analysis system 10 is a system that executes an invention document analysis process for analyzing invention documents.

The invention document analysis system 10 is a system for analyzing invention documents and outputting information such as patent application related information based on the analysis results. be done. Here, the explanation will be continued assuming that the invention document analysis system 10 is realized by a computer equipped with an operating system (OS) including a file system, etc., and other programs (for example, application programs, etc.). . Here, the scale and mode of the computer including the microprocessor are not limited, and may be, for example, a supercomputer, a general-purpose computer, a minicomputer, a PC (Personal Computer), a tablet, a smartphone, or the like. Further, the microprocessor may be, for example, a quantum processor, and the computer including the microprocessor may be, for example, a quantum computer. A computer may include, for example, an input device and a display device. The computer input device may be a pointing device such as a mouse, a keyboard, a touch pad, a voice input device, or the like. The display device of the computer may be, for example, a liquid crystal display, an organic electro-luminescence (OEL) display, or the like. Note that the input device and display device of the computer may be integrated; for example, the computer may include a touch panel in which a touch pad is superimposed on the display surface of the display device.

Users of the invention document analysis system 10 include, for example, creators of invention documents (e.g., inventors or patent attorneys), translators who use invention documents (e.g., people who translate invention documents into foreign languages, This is assumed to be a translator who creates a translation result document for a foreign application, or a user of the invention document other than the translator. Note that the user of the invention document analysis system 10 may be a person who is not an inventor, a patent attorney, a translator, or the like.

As shown in FIG. 4, the invention document analysis system 10 includes a setting information acquisition section 101, a document acquisition section 102, an important phrase setting dictionary acquisition section 103, an invention document standard provision acquisition section for translation 104, an analysis section 105, and a document evaluation section. 106, a machine learning section 107, and an output section 108. The functions of the setting information acquisition unit 101, document acquisition unit 102, important word/phrase setting dictionary acquisition unit 103, invention document standard provision acquisition unit for translation 104, analysis unit 105, document evaluation unit 106, machine learning unit 107, and output unit 108 are as follows. , for example, by one or more computers executing a program.

[2-1-1. Setting information acquisition section]
The setting information acquisition unit 101 acquires setting information that the invention document analysis system 10 uses as a basis for analyzing an invention document. The setting information includes information to be referred to when analyzing the invention document. Analysis of invention documents includes extraction of key words from the invention documents. FIG. 5 shows an example of the contents of the setting information.

The setting information is information that the invention document analysis system 10 refers to. The setting information may exist inside or outside the invention document analysis system 10, for example. The configuration information may be in any format. Configuration information may exist, for example, as the content of a file (e.g., a text file) or as the content of a database, and may be one or more environment variables managed by the computer's OS and defined by the user. It may exist as content or as data forming part of the invention document analysis system 10 (for example, data written in a program for implementing the invention document analysis system 10). When the setting information exists as data that constitutes a part of the invention document analysis system 10, the invention document analysis system 10 stores all or all of the setting information via a graphical user interface (GUI) or the like. Part of the content may be displayed so that the user can change part of the setting information.

For example, if the setting information exists outside the invention document analysis system 10 (for example, as the contents of an external file), the setting information acquisition unit 101 reads the setting information into the invention document analysis system 10. , the setting information may be acquired by converting it into a data format that is easy to use within the invention document analysis system 10. Further, the setting information acquisition unit 101 reads externally existing setting information (for example, setting information as the contents of a file), and uses the read content as a default value in advance as data constituting a part of the invention document analysis system 10. The setting information may be acquired by overwriting all or part of the setting information in which the setting information is defined.

The setting information includes condition information indicating conditions for extracting important words from an invention document to be analyzed by the invention document analysis system 10. A character string that satisfies the extraction conditions indicated by the condition information is an important phrase. It is assumed that condition information is set and operated so that this important word is a word that is estimated to be important to some extent, but the accuracy of the estimation of its importance is 100%. There's no need. Important words and phrases are, for example, keywords, key phrases, and the like. A character string is composed of one or more characters such as kanji, kana characters, alphabetic characters, numbers, symbols, and other characters that can be represented by a character code. One character code representing a certain character is determined depending on an encoding format such as UTF-8, Shift_JIS, or Unicode.

The condition information includes, for example, important word/phrase candidate character string pattern information, reference character string pattern information, code character string pattern information, specific particle character string pattern information, and the like. Important term candidate character string pattern information, reference character string pattern information, code character string pattern information, and specific particle character string pattern information are, for example, information that represents a character string pattern indicating a set of character strings to be extracted using a regular expression. It is. Note that the condition information may include information related to extraction conditions indicating other character string patterns for extracting important words and phrases, information related to extraction exception conditions indicating exceptions to extraction, and the like.

The important word/phrase candidate character string pattern information indicates an important word/phrase candidate character string pattern representing a pattern of character strings that are candidates for important words. Important word candidate character string pattern information includes, for example, a predetermined character string whose first character is a kanji character (for example, a character string such as "control device" or "control system"), and another predetermined character string whose first character is a katakana character. A character string (for example, a character string such as "sensor" or "sensing method") is determined to correspond to the important word candidate character string pattern. Note that the important word/phrase candidate character string pattern information does not necessarily have to be determined so that all character strings whose first characters are kanji or katakana correspond to the important word/phrase candidate character string pattern. For example, "receive", which is part of the character string of the verb "receive" written in the invention document, may not be included in a predetermined character string whose first character is a kanji character. Certain conditions can be attached to a string in order for it to be a predetermined string, so for example, a string that is part of the beginning of a verb may not be included in the predetermined string even if the first character is a kanji or katakana. Alternatively, a character string immediately followed by "su" may not be included in the predetermined character string even if the first character is a kanji or katakana character.

The reference character string pattern information indicates a reference character string pattern representing a pattern of a reference character string used for referencing a previously described character string. The reference character string pattern information is determined such that, for example, character strings such as "said", "above", "that", "mentioned above", "that", etc. correspond to the reference character string pattern.

The code character string pattern information indicates a code character string pattern representing a pattern of a character string of a code appended subsequent to a character string representing an element name. Each reference numeral is added to a portion of the drawing for the patent application. Also, in order to refer to a part of the figure in the specification for a patent application, etc., the same code as that added to a part of the figure is used following the element name indicating the part of the figure. , may be described. The code may also be referred to as a reference code.

The code string pattern information is, for example, a predetermined character string whose first character is a number (for example, a character string such as "100", "101a", "122-1", "123'", etc.) It is determined that it corresponds to a pattern. Note that the code string pattern information does not necessarily have to be determined so that all character strings whose first character is a number correspond to the code string pattern. Further, the code string pattern information may be determined such that, for example, a string whose first character is other than a number (for example, an alphabetic letter) also corresponds to the code string pattern.

The specific particle character string pattern information indicates a specific particle character string pattern that represents a pattern of character strings that are predetermined particles. The specific particle character string pattern information indicates that, for example, if any of the characters "wo", "ni", "ga", "to", "kara", "he", and "ha" is the specific particle It is defined to correspond to a string pattern.

The setting information further includes, for example, information for specifying the important word/phrase setting dictionary.

The information for specifying the important word/phrase setting dictionary is information necessary for the important word/phrase setting dictionary acquisition unit 103 of the invention document analysis system 10 to acquire the important word/phrase setting dictionary, and for example, a file containing the important word/phrase setting dictionary. This information is necessary for accessing the file, such as file path information indicating the location of the file, etc. Note that, for example, the identification information of the important word/phrase setting dictionary may be the important word/phrase setting dictionary itself.

The setting information may include, for example, information for specifying the storage location of the output file. An example of the information for specifying the storage location of the output file is information that specifies a predetermined folder (that is, a predetermined directory) in the file system.

[2-1-2. Document acquisition department]
The document acquisition unit 102 acquires an invention document. The document acquisition unit 102 can acquire each of a plurality of invention documents sequentially, but may be able to acquire a plurality of invention documents simultaneously, for example. The invention document acquired by the document acquisition unit 102 can be used by the analysis unit 105 and the machine learning unit 107.

For example, when an invention document exists as a file stored in a recording medium or the like, the document acquisition unit 102 has a function of acquiring the invention document (i.e., the document data of the invention document) that is the file content by reading the file. have Note that the document acquisition unit 102 may, for example, retrieve an invention document from data in a database, data stored in the memory of a computer constituting the invention document analysis system 10 (for example, when a clipboard is available in the OS, an application program (data that has been copied and stored in a memory area used as a clipboard), the invention document can be read out from a database, memory, etc. It may also have a function to obtain .

For example, the document acquisition unit 102 can acquire, in addition to a plurality of invention documents, a plurality of translation result documents that are translation results of each of the plurality of invention documents. The translation result document is used, for example, by the machine learning unit 107.

For example, in addition to the invention document, the document acquisition unit 102 may acquire a drawing document corresponding to the invention document. A drawing document corresponding to an invention document is a document representing a drawing referred to in the specification, etc. of the invention document. Generally, the same reference numerals as those added to a part in the drawings for a patent application are written following words representing elements corresponding to the part in the specification for the patent application. A drawing document exists as a file such as a PDF (Portable Document Format) file, for example. If the drawing document exists as a searchable PDF file with text information added (in other words, a PDF file in which character strings are searchable), the symbols and other information in each drawing expressed in the drawing document It is possible to easily search for character strings. Furthermore, even if the drawing document exists as a PDF file that is not a searchable PDF file, conventional optical character recognition (OCR) technology or character recognition technology using machine learning can be used. Thus, for example, it is possible to search for codes and other character strings in each drawing represented in a drawing document, and for example, it is also possible to convert the PDF file into a searchable PDF file using existing technology.

[2-1-3. Important word setting dictionary acquisition part]
The important word/phrase setting dictionary acquisition unit 103 obtains an important word/phrase setting dictionary based on the identification information of the important word/phrase setting dictionary in the setting information acquired by the setting information acquiring unit 101 . An important word/phrase setting dictionary is information that lists multiple important words to be extracted (in other words, information that includes multiple important words), and for each of all or some of the multiple important words, This information is configured such that the important word/phrase may be associated with a corresponding translated word translated into a predetermined language other than Japanese (for example, English). In the important word/phrase setting dictionary, for example, some important words are associated with corresponding translation words.

The important word/phrase setting dictionary may exist, for example, as a file, as the contents of a database, or as data within the invention document analysis system 10.

The important word/phrase setting dictionary can be used by the important word/phrase extraction unit 1051 in the analysis unit 105 to extract important words.

[2-1-4. Invention document standard provision acquisition department for translation]
The invention document standard provision acquisition unit 104 obtains standard provisions for translation invention documents that define a plurality of conditions that an invention document to be translated must satisfy.

The standard regulations for invention documents for translation include, for example, information that defines the conditions that an invention document to be translated should satisfy, and information that indicates patterns of character strings for potentially inappropriate descriptions that do not meet the conditions. Contains inappropriate character string pattern information indicating the character string pattern. In addition, the information that defines the conditions that the invention document to be translated should satisfy may include, for example, inappropriate character string pattern information that represents the character string pattern of the potentially inappropriate written portion that does not satisfy the conditions. Contains correction information on how to correct the potentially inappropriate written portion in association with all or part of the information.

The invention document standard provisions for translation are used, for example, by the analysis unit 105 and the machine learning unit 107. Further, the modification information, which is an example of the content of the invention document standard provisions for translation, is used to generate modified claim information and modified specification information that can be included in the patent application related information that can be output by the output unit 108. obtain.

[2-1-5. Analysis Department]
The analysis unit 105 analyzes the invention document acquired by the document acquisition unit 102 based on the setting information acquired by the setting information acquisition unit 101. The analysis unit 105 includes an important phrase extraction unit 1051, a translation word candidate set identification unit 1052, a classification unit 1053, and a translation unit 1054. For example, the analysis unit 105 determines whether the invention document acquired by the document acquisition unit 102 satisfies at least some of the conditions defined in the standard regulations for invention documents for translation. , based on the result of the determination, generates warning information to point out potentially inappropriate parts of the invention document. Among the plurality of conditions defined in the standard regulations for invention documents for translation, the conditions for which the analysis unit 105 determines whether or not an invention document satisfies the conditions are, for example, selection information defined in the setting information. The invention document analysis system 10 can be selected (that is, narrowed down) based on the above criteria or by a selection operation by the user of the invention document analysis system 10.

This warning information may be included in the patent application related information that may be output by the output unit 108. Note that the determination by the analysis unit 105 as to whether or not an invention document satisfies the conditions defined in the standard provisions for invention documents for translation is based on, for example, inappropriate character string pattern information that may be included in the standard provisions for invention documents for translation. This is done depending on whether a character string corresponding to the indicated inappropriate character string pattern exists in the invention document. Further, the warning information generated as a result of the analysis of the invention document by the analysis unit 105 may include, for example, invention document evaluation information indicating the conformity of the invention document to the invention document standard provisions for translation. The invention document evaluation information indicates, for example, an evaluation value as a result of evaluating the conformity of the invention document to the invention document standard provisions for translation. For example, the evaluation value as a result of evaluating the conformity of an invention document to the standard provisions for invention documents for translation is based on the number of conditions selected as a judgment target from among the plurality of conditions specified in the standard provisions for invention documents for translation. It may be a conformance rate that is a ratio of the number of conditions determined to be satisfied by the invention document. In addition, for example, the height of the precision rate can be divided into multiple ranges, and a value representing the category to which the precision rate of the judgment result falls (for example, a serial number specified for each category in order from the lowest precision rate to the highest). ) may be used as the evaluation value.

[2-1-5-1. Important word extraction section]
The important term extraction unit 1051 extracts important terms from the invention document acquired by the document acquisition unit 102 based on the condition information included in the setting information. By appropriately determining the condition information, the important term extraction unit 1051 can, for example, extract information from individual patent publications among several patent publications published by the Japan Patent Office, as an example of an invention document. A plurality of important words can be extracted from the claims or specifications.

The important words and phrases extracted by the important word and phrase extraction unit 1051 are used, for example, to generate information such as patent application related information output by the invention document analysis system 10. Here, using an example in which an invention document includes claims and a specification, the important word extraction unit 1051 that extracts important words from the invention document will be described in detail. The important phrase extraction unit 1051 includes a first extraction unit 10511 and a second extraction unit, each of which performs extraction according to a first extraction condition, a second extraction condition, a third extraction condition, and a fourth extraction condition for extracting important words. 10512, a third extraction section 10513, and a fourth extraction section 10514.

The first extraction unit 10511 adds condition information to the claim and specification of the invention document acquired by the document acquisition unit 102, following the character string corresponding to the reference character string pattern indicated by the reference character string pattern information in the condition information. If there is a character string that corresponds to the important word candidate character string pattern indicated by the important word candidate character string pattern information in , extract the same character string as the important word candidate character string pattern as an important word. Has a function.

The second extraction unit 10512 adds a period character to the claim of the invention document acquired by the document acquisition unit 102 after a character string corresponding to the important term candidate character string pattern indicated by the important term candidate character string pattern information in the condition information. exists, the same character string as the character string corresponding to the important word/phrase candidate character string pattern is extracted as an important word/phrase. The period character is, for example, ".".

The third extraction unit 10513 extracts a character string corresponding to the important term candidate character string pattern indicated by the important term candidate character string pattern information in the condition information into the invention document acquired by the document acquisition unit 102, If there is a character string that corresponds to the code string pattern indicated by the code string pattern information, it has a function of extracting the same character string as the important word/phrase candidate character string pattern as an important word/phrase.

The fourth extraction unit 10514 extracts an important term candidate character string that is a character string corresponding to the important term candidate character string pattern indicated by the important term candidate character string pattern information in the condition information into the invention document acquired by the document acquisition unit 102. Subsequently, when there is a specific particle string that is a character string that corresponds to the specific particle string pattern indicated by the specific particle string pattern information in the condition information, a character string that corresponds to the important word candidate string pattern and It has a function to extract identical character strings as important words.

Note that the important term extraction unit 1051 may extract important terms from the invention document, for example, based on information related to extraction conditions included in the condition information. In addition, the important phrase extraction unit 1051 uses a first extraction unit 10511, a second extraction unit 10512, a third extraction unit 10513, or a fourth extraction unit 10514, for example, based on information related to extraction exception conditions included in the condition information. Some of the important words and phrases extracted from the invention document may be treated as non-important words. Further, the important phrase extraction section 1051 may be modified so as not to include any one of the first extraction section 10511, the second extraction section 10512, the third extraction section 10513, and the fourth extraction section 10514. .

Further, the important word/phrase extraction unit 1051 may detect, for example, that the same important words as those included in the important word/phrase setting dictionary acquired by the important word/phrase setting dictionary acquisition unit 103 are included in the invention document acquired by the document acquisition unit 102. If the key words and phrases are included, the same key words and phrases as those included in the key word and phrase setting dictionary may be extracted from the invention document.
Further, the important word/phrase extraction unit 1051 may use one or a small number of relatively complex important word/phrase candidate character string patterns to extract the important word/phrase, or may use a relatively concise important word/phrase candidate character string pattern. It is also possible to use more than one. For example, in order to extract important words, if a plurality of character strings match the important word candidate character string pattern, the important word/phrase extraction unit 1051 extracts important words. When a character string is included in the description of an invention document, important words may be selected and extracted using the remaining character strings after excluding one of the plurality of character strings. That is, when extracting important words using the first extraction condition, second extraction condition, third extraction condition, and fourth extraction condition, the important word extraction unit 1051 determines whether a plurality of character strings correspond to the important word candidate character string pattern. In the case where one of the plurality of character strings is included in another character string in the claim or description of the invention document, that one of the plurality of character strings is It is possible to use a selection extraction condition that selects and extracts important words and phrases from the remaining character strings that have been excluded.

[2-1-5-2. Translation word candidate set identification unit]
For each important word extracted by the important word extraction section 1051 from the invention document acquired by the document acquisition section 102, the translation word candidate set specifying section 1052 identifies a translation word that is a set of corresponding translation word candidates for the important word. Identify candidate sets. The translation word candidate information representing the translation word candidate set specified by the translation word candidate set specifying unit 1052 for each important term is included in, for example, the patent application related information output by the invention document analysis system 10.

The translation word candidate set includes one or more elements (that is, corresponding translation word candidates). The corresponding translated word is a word that is the result of translating an important word or phrase into a foreign language. Here, as an example, the corresponding translated words will be explained as English words, but the corresponding translated words may be words in a language other than English (for example, Chinese). For example, when the important phrase is "amount of change," examples of elements of the translation word candidate set include "variation" and "amount of change."

As a specific example, the translation word candidate set identification unit 1052 uses bilingual dictionary information (for example, conventionally known Japanese-English dictionary data) that associates Japanese words and phrases with foreign language words, and uses the important word extraction unit By searching the bilingual dictionary information using the important word/phrase extracted by 1051 as a key, one or more bilingual words corresponding to the important word/phrase are detected, and a set of the detected words/phrases is searched for the important word/phrase. is identified as a set of translation word candidates. A specific example of bilingual dictionary information is, for example, the Japanese-English dictionary "edict2" created by the Japanese dictionary project founded by Mr. James William Breen of Australia, or the processing and editing of that Japanese-English dictionary. For example, the data obtained by The translation word candidate set specifying unit 1052 can hold bilingual dictionary information, for example.

By the invention document analysis system 10, the output unit 108 can output patent application related information including each important term and translation word candidate information representing a corresponding translation word candidate set. In addition, the invention document analysis system 10 calculates, for each important term in the invention document, a corresponding element selected by the user of the invention document analysis system 10 from among the set of translation word candidates identified corresponding to the important term. Translation auxiliary information, which is information that associates a translated word with the important phrase, may be output. In this case, one or more users select one or more corresponding translation words corresponding to important words when each of the plurality of invention documents is sequentially analyzed by the invention document analysis system 10. The above translation auxiliary information may be stored in a recording medium or the like inside or outside the invention document analysis system 10. The recording medium may be any medium capable of recording data, such as a memory, hard disk, or optical disk. One or more pieces of translation auxiliary information stored in a recording medium or the like may be integrated to generate translation auxiliary information in a format in which each important phrase is associated with a set of corresponding translation words. The translation auxiliary information may be output in the form of file contents, database contents, etc., or may be output in the form of display. In a case where already generated translation auxiliary information can be used, the translation word candidate set specifying unit 1052 uses the already generated translation auxiliary information to identify the important words extracted by the important phrase extraction unit 1051. By searching the already generated translation auxiliary information using the important word as a key, we can detect words that have been selected by any user in the past as one or more corresponding translation words corresponding to the important word. The set of words/phrases may be specified as a set of translation word candidates for the important word/phrase.

As another specific example, the translation word candidate set identification unit 1052 uses the classification unit 1053 or the technical field identification unit 10531 to identify a group to which the invention document acquired by the document acquisition unit 102 belongs, or a group corresponding to the invention document. When a technical field is identified, bilingual dictionary information corresponding to that group or technical field is selected from a plurality of bilingual dictionary information, and the important words extracted by the important word extraction unit 1051 are used as keys. By searching the selected bilingual dictionary information, one or more bilingual words corresponding to the important word are detected, and the set of detected words is identified as a set of translation word candidates for the important word. It may also be a thing. One specific example of this is a group-by-group or technical-field-by-technical field in which, for each of a plurality of groups or technical fields, Japanese words and phrases in invention documents belonging to that group or technical field are matched with words and phrases in the target language. It can be used when multiple bilingual dictionary information exists. In each of this bilingual dictionary information, for example, the order in which the words are estimated to be most likely to be selected as important words in invention documents in the group or technical field corresponding to the bilingual dictionary information (in other words, the corresponding translated words) The translation word candidate set specifying unit 1052 orders the words and phrases of one or more bilingual translations corresponding to the important word (in descending order of validity), and the translation word candidate set specifying unit 1052 arranges the words and phrases in order to match the order of the words in the selected bilingual dictionary information. When information related to the translation word candidate set is output from the invention document analysis system 10, the arrangement order of each element of the translation word candidate set may be determined.

As another specific example, when the classification unit 1053 specifies a group to which one invention document acquired by the document acquisition unit 102 belongs, the translation word candidate set identification unit 1052 identifies other words belonging to the group. A plurality of pairs of important words and phrases extracted by the important word and phrase extracting unit 1051 and corresponding translated words corresponding to the important words are identified from each of the plurality of invention documents and each of the corresponding translation result documents, and the plurality of pairs are extracted. One of the documents acquired by the document acquisition unit 102 through machine translation using a learning model for important phrase translation, which is a learning model generated by machine learning, using the following as training data for machine learning for important phrase translation. Identify the elements of the translation word candidate set for each important word extracted by the important word extraction unit 1051 from the invention document of (specific).

[2-1-5-3. Classification section]
The classification unit 1053 uses the important word group, which is a plurality of important words extracted by the important word extraction unit 1051, for each of the plurality of invention documents acquired by the document acquisition unit 102, and extracts each extracted important word group. Based on the correlation between the plurality of invention documents, group classification is performed to specify which of the plurality of groups each of the plurality of invention documents belongs to. In order to evaluate the correlation between the important word groups, the classification unit 1053 utilizes an algorithm that counts the number of the same important words that are common to both important word groups and a synonym dictionary. An algorithm that counts the number of important words and phrases for each group of important words, using a predetermined number of predetermined strings (e.g., thousands, tens of thousands, etc.) as elements, for example, the character string of that element is important. Define vector data with an element value of 1 if it exists as a word or phrase and 0 if it does not exist as an important word, and use any algorithm such as various algorithms to calculate the correlation coefficient between the vector data. Good too. Further, the classification unit 1053 may have a function of specifying a group to which each invention document belongs, for example, based on a specification by a user of the invention document analysis system 10. The group classification unit, method, etc. is determined by the operation unit of the invention document analysis system 10 (for example, invention documents, bilingual dictionary information, or information output by the invention document analysis system 10, which is used by each user who communicates with each other). It may be useful for each invention document analysis system 10) to be unified.

The classification unit 1053 includes a technical field identification unit 10531.

For each of a plurality of technical fields, the technical field identification unit 10531 extracts important words and phrases from the important word extraction unit 1051 after each of one or more invention documents known to be related to the technical field has been acquired in the past by the document acquisition unit 102. By using the results of past extraction of groups, important phrases extracted by the important phrase extraction unit 1051 from invention documents in which the technical field is not specified and which are acquired by the document acquisition unit 102 among a plurality of technical fields. Any technical field whose correlation with the word group is evaluated to be at least a predetermined degree is specified as a technical field corresponding to an invention document whose technical field is not specified. The technical field can be considered an example of a group. A technical field is expressed, for example, by International Patent Classification (IPC). The technical field specifying unit 10531 may use classification up to any level of the IPC as the technical field, and may treat only sections, or only sections and classes, etc., as one technical field. Further, the technical field specifying unit 10531 may express the technical field using all of the IPC sections, classes, subclasses, main groups, and subgroups. For example, if a technical field is to be distinguished only by IPC sections, the technical field identification unit 10531 uses patent publications etc. to acquire a plurality of invention documents belonging to each section of the IPC, and It is possible to obtain each invention document by the invention document analysis system 10 and obtain a group of important words extracted by the important word extraction unit 1051 that corresponds to the section. For example, by using this group of important words for each section obtained in the past, the technical field identification unit 10531 can extract important words from the invention document for which no section as a technical field has been specified, which has been acquired by the document acquisition unit 102. A section whose correlation with the important word group extracted by the word/phrase extraction unit 1051 is evaluated to be at least a predetermined degree can be specified as the technical field of the invention document.

The technical field information, which is information indicating the technical field (for example, IPC) specified by the technical field specifying section 10531 corresponding to the invention document acquired by the document acquisition section 102, is, for example, a patent application outputted by the invention document analysis system 10. Included in related information.

Further, the group specified by the classification unit 1053 to which the invention document belongs, or the technical field specified corresponding to the invention document by the technical field specification unit 10531, for example, the translation word candidate set specification unit 1052 It can be used to select bilingual dictionary information for specifying a set of translation word candidates corresponding to each important word/phrase extracted from. Furthermore, the group classification by the classification unit 1053 can be used by the machine learning unit 107 to generate a translation learning model for each group.

[2-1-5-4. Translation Department]
The translation unit 1054 receives the invention document acquired by the document acquisition unit 102 as input and performs machine translation of the invention document using a translation learning model and performs arithmetic processing predetermined for translation. A learning model for translation is a learning model for machine translation that is generated by machine learning using multiple sets of an invention document and a translation result document as a result of translation by a translator based on the invention document as training data. It is a learning model.

Here, as an example, the translation learning model will be explained as a learning model used to machine translate an invention document, which is a document mainly written in Japanese, into an English document. It may also be a learning model used for machine translation into documents in Chinese or other languages instead of English. An example of a method for generating a translation learning model for translating Japanese into English is to use the international publication bulletin of an international application written in Japanese and the international application in English such as the United States or the United Kingdom. The invention document and the translation result document should be paired with a gazette published by the government office of the designated country that contains content equivalent to the translation result document translated into English in order to enter the national phase of the designated country. One method is to perform supervised machine learning by treating the data as supervised data.

In addition, when an invention document can be classified into multiple groups, it is better to generate a translation learning model for each group to improve the accuracy of machine translation when an appropriate translation learning model is selectively used. There is a good possibility that an improvement in the appropriateness of translation can be expected. An example of group classification is an example in which invention documents are classified into a group of invention documents related to electrical technology, a group of invention documents related to mechanical technology, and a group of invention documents related to chemical technology. Another example of group classification is an example in which the technical field to which an invention document belongs is classified by IPC section, or those with different sections and classes are classified as different groups. Further, as another example of group classification, in order to distinguish the department to which the inventor who created the invention indicated in one invention document belongs from other departments, it is possible to group the department according to the department to which the inventor of the invention An example of categorizing is given.

When a translation learning model exists for each of a plurality of groups or technical fields, the translation unit 1054 uses the classification unit 1053 to specify, for example, the group or technical field to which the invention document acquired by the document acquisition unit 102 belongs. , all or part of the invention document may be machine translated by selectively using a translation learning model corresponding to the group or technical field to which the invention document belongs.

The results of machine translation of the invention document by the translation unit 1054 may be included in the patent application related information output by the invention document analysis system 10, for example.

[2-1-6. Document Evaluation Department]
The document evaluation unit 106 determines, for each of the plurality of invention documents acquired by the document acquisition unit 102, whether the invention document satisfies at least some of the conditions defined in the standard regulations for invention documents for translation. A determination is made as to whether or not the invention is true, and from among the plurality of invention documents, one or more invention documents that satisfy the condition to which the determination was made exceed a predetermined degree are selected. Among the plurality of conditions defined in the standard regulations for invention documents for translation, the conditions for which the document evaluation unit 106 determines whether or not an invention document satisfies the conditions are, for example, the selection criteria defined in the setting information. The selection may be made based on information or the like or by a selection operation by the user of the invention document analysis system 10.

The invention document selected by the document evaluation unit 106 can be used by the machine learning unit 107, for example.

[2-1-7. Machine Learning Department]
The machine learning unit 107 generates a learning model for machine translation of an invention document by machine learning.

For example, the machine learning unit 107 calculates, for each invention document selected by the document evaluation unit 106, a set of the invention document and the translation result document acquired by the document acquisition unit 102 corresponding to the invention document. A learning model for translation is generated by executing machine learning using multiple sets of training data as training data.

As another example, the machine learning unit 107 calculates, for each of a plurality of groups, the following information about each invention document classified into that group by the classification unit 1053 among the plurality of invention documents acquired by the document acquisition unit 102. Translation, which is a learning model for machine translation, is performed by machine learning using each set of the invention document and the translation result document obtained by the document acquisition unit 102 as the translation result of the invention document as training data. Generate a learning model for This makes it possible to generate a learning model for translation corresponding to each of a plurality of groups related to the invention document. In this example as well, the invention documents used as teacher data may be limited to, for example, invention documents selected by the document evaluation unit 106.

As another example, the machine learning unit 107 detects, in the invention document acquired by the document acquisition unit 102, inappropriate characters indicated by inappropriate character string pattern information related to the conditions specified in the invention document standard regulations for translation. If there is a character string that corresponds to the column pattern, the invention for translation is created based on the content of the invention document and the correction information associated with the inappropriate character string pattern information in the standard provisions for invention documents for translation. A revised invention document (that is, a document with content equivalent to the revised invention document) that is generated by modifying the content of the invention document to meet the conditions specified in the document standard provisions, and the invention. A learning model for translation is generated by machine learning using a set of a translation result document acquired by the document acquisition unit 102 corresponding to the document as training data. For example, a modified version of an invention document is used when a character string in an invention document corresponds to an inappropriate character string pattern indicated by inappropriate character string pattern information to which modification information is associated in the standard provisions for invention documents for translation. It can be generated by deleting a string or replacing it with another character string, etc. according to its modification information. The modified version of the invention document includes, for example, modified version claim information that represents the content in which the character string of the claim of the invention document has been modified based on the modification information corresponding to the inappropriate character string pattern, and the characters of the specification of the invention document. The text includes at least one of the modified specification information representing the content modified based on the modification information corresponding to the relevant inappropriate character string pattern.

The translation learning model generated by the machine learning unit 107 can be used by the translation unit 1054, for example. The translation learning model generated by the machine learning unit 107 can be used, for example, in a machine translation system or a machine translation program for translating invention documents, or in a general machine translation system.

[2-1-8. Output section]
The output unit 108 outputs information generated based on the analysis result of the analysis unit 105 regarding the invention document acquired by the document acquisition unit 102.

The output unit 108 outputs, for example, information generated based on each of a plurality of key terms extracted by the key term extraction unit 1051 from the invention document acquired by the document acquisition unit 102.

As a specific example, the output unit 108 outputs patent application related information generated based on the analysis result by the analysis unit 105 of the invention document acquired by the document acquisition unit 102. FIG. 6 shows an example of the content of patent application related information output by the output unit 108.

The patent application related information output by the output unit 108 may include warning information generated by the analysis unit 105.

For example, the output unit 108 outputs patent application related information including, as contents to be displayed, each of a plurality of important terms extracted by the important term extraction unit 1051 from the invention document acquired by the document acquisition unit 102. It has a function to output as a display document. The patent application related information as a display document includes, for example, an important word list listing a plurality of important words.

Here, the display document is, for example, a document written in a markup language or the like that can be displayed by an application program such as a browser (Web Browser). The display document is, for example, an HTML document written in HTML (Hypertext Markup Language). The HTML document includes at least an HTML file, and the HTML file is written in accordance with, for example, HTML5. In the display document, the content to be displayed, such as important words, should be written as the contents of the <body> tag of the HTML file (that is, written between the <body> and </body> tags). is displayed by a browser, etc. For example, the characters to be displayed in the content of the <body> tag or the content of tags in the hierarchy below it may be set to be displayed in a monospaced font, for example. Monospaced fonts allow multiple characters to be displayed aligned. Further, the characters to be displayed in the content of the <body> tag or the content of tags in the hierarchy below it may be set to be displayed in a proportional font instead of a monospaced font. For example, the HTML file may include a description for reading an external file as necessary, or instead, the contents corresponding to the external file may be written directly in the HTML file. The external file is, for example, a CSS file written in CSS (Cascading Style Sheets) regarding the display mode of display elements, a script file containing a script written in a scripting language such as JavaScript (registered trademark), or the like. Therefore, an HTML document as an example of a display document may be composed of a group of files such as one or more HTML files, one or more CSS files, and one or more script files. The HTML file includes information that specifies, for example, a CSS file and a script file. The HTML document may be composed of only HTML files, HTML files and CSS files, or HTML files and script files. Note that CSS allows, for example, control of the display mode of characters in patent application related information, and, for example, the font-family property of CSS makes it possible to specify the font of characters to be displayed.

For example, the patent application related information includes, for each of a plurality of important terms extracted by the important term extraction unit 1051 from the invention document acquired by the document acquisition unit 102, additional information related to the important term is added to the important term. Contains information about Note that the additional information in the patent application related information does not necessarily need to be added to each of the important words extracted from the invention document by the important word extraction unit 1051, but may be added to some of all the extracted important words. It may be added only to each of a plurality of important words.

The additional information related to the important word/phrase in the patent application related information may be, for example, translated word information as information regarding a corresponding translated word that is a translated word/phrase expressing the meaning of the important word/phrase. Here, the translated phrase is a phrase that includes a word expressed in one type of language other than Japanese (eg, English, Chinese, etc.). For example, when the important word is "object," an example of the translated word is "object." The translated word information as information regarding the corresponding translated words is, for example, information representing the corresponding translated words themselves (for example, machine translation results for important words, etc.), a set of translated word candidates that are a set of candidates for the corresponding translated words, and the like. When the translation word information is translation word candidate information that is information representing a translation word candidate set for important words, the translation word candidate information further includes information related to the translation word candidate set (for example, translation word candidate information). For each element of the candidate set, a value indicating the level of validity as a corresponding translation word, information indicating a rank, etc.) may be included.

Further, additional information related to an important term in the patent application related information may be obtained by, for example, selecting any element of the translation word candidate set specified by the translation word candidate set specifying unit 1052 for the important term. It includes display control information for displaying the set of translation word candidates as a graphical user interface (GUI) element so that they can be selected. Here, the operation may be performed by a user of the computer, for example, on an input device of the computer, when the patent application related information is displayed on the display device of the computer, such as in a browser. An example of a GUI element is a button (for example, a push button, a radio button, etc.) that can be operated in a displayed state, a menu, etc., but a display element such as a character string can also be a GUI element. In an HTML document, which is an example of a display document, a script (for example, a function written in JavaScript (registered trademark) A character string or other element described as the content of an HTML tag (for example, a <span> tag) to which a HTML tag (e.g., <span> tag) can be added can be an operation-compatible GUI element that is an operable GUI element.

The display control information included in the additional information related to an important word/phrase includes, as one specific example, translation word candidate information representing a set of translation word candidates, and translated word candidates for the important word/phrase displayed as a GUI element. A script for executing a determination process that determines an element selected from a set according to an operation as a corresponding translation word corresponding to the important word, and each script whose corresponding translation word is determined as a result of the execution of the determination process. It includes a script for executing processing for generating and outputting translation auxiliary information, which is information that associates important words and phrases with corresponding translated words. By executing the script in response to the operation, some processing (such as determining a corresponding translation word for an important term) is performed, and for example, the patent application related information displayed on the computer display device is The displayed content may change depending on the result of the processing. In the determination process of determining an element selected from a translation word candidate set as a corresponding translation word, for example, along with the determination, the displayed translation word candidate set may be replaced with a corresponding translation word and displayed.

In addition, all or part of the display content displayed by the browser in the patent application related information as a display document can be selected by the user's operation and can be the target of copy and paste or cut and paste. By pasting (that is, pasting), it can be included as the content of a document (for example, a text file, etc.) used by a certain application program that is being executed on a computer. In addition, if the content displayed by the browser changes as a result of processing executed in response to user operations on GUI elements related to patent application related information, all or part of the changed display content may be changed by the user. Depending on the operation, it can be selected and can be copied and pasted.

In addition, the important word/phrase extraction unit 1051 retrieves the same important word/phrase with which the corresponding translation word is associated in the important word/phrase setting dictionary acquired by the important word/phrase setting dictionary acquisition unit 103 and retrieves it by the document acquiring unit 102. The additional information related to the important word/phrase extracted from the invention document, for example, represents the corresponding translated word. Further, when the important word/phrase extraction unit 1051 extracts important words other than the important words to which corresponding translation words are associated in the important word/phrase setting dictionary from the invention document acquired by the document acquisition unit 102, the extracted important words/phrases are also included. For example, the additional information related to the key phrase may be displayed as a GUI element so that any element of the set of translation word candidates for the important word can be selected in response to an operation. Contains display control information for In this case, the display control information is selected according to the operation from, for example, translation word candidate information representing a set of translation word candidates for important words and phrases, and a set of translation word candidates for important words and phrases displayed as GUI elements. A script for executing a determination process that determines an element as a corresponding translation word corresponding to its important term, and a corresponding correspondence between the important term and each important term for which a corresponding translation term is determined as a result of the execution of the determination process. It also includes a script for executing an update process for updating the important word/phrase setting dictionary acquired by the important word/phrase setting dictionary acquisition unit 103 so as to include information associated with translated words.

The additional information related to an important term in the patent application related information may be, for example, the importance of the important term, attribute information of the important term, or the like. Further, the additional information related to the important word/phrase may be display control information for executing some kind of processing related to the important word/phrase in response to the operation of a GUI element. In this case, the display control information includes, for example, information indicating the content displayed as a GUI element, information indicating an operation on the GUI element (for example, a mouse click operation), and a process that is started in response to the operation. It is associated with a script.

Examples of ways to output information such as patent application related information in the output unit 108 include writing the information to a recording medium, writing the information as a file, displaying the information, and printing the information. , transmitting that information, outputting audio that represents that information, and causing other physical phenomena that represent that information.

For example, the output unit 108 may have all or part of a function equivalent to a browser, and may output the patent application related information by displaying the patent application related information. The output unit 108 may generate a display document (for example, an HTML document) containing patent application related information, and then display the patent application related information in accordance with the display document.

If the invention document acquired by the document acquisition unit 102 includes a claim, the output unit 108 may include claim information representing the content of the claim in the patent application related information, for example. In patent application related information as a display document, claim information is one of the contents to be displayed.

In the display document, the claim information representing the contents of the claim of the invention document is a display control included in the additional information related to each of the plurality of important words in the claim extracted by the important word extraction unit 1051. The information includes information for controlling the display mode of the important word/phrase, for example, information indicating the color to be added to the important word/phrase. For example, display control information for a certain important term in patent application related information as an HTML document can be obtained by using a <span> tag that describes the important term as the content in the HTML file (that is, a <span> tag preceding the important term as the content). <span> tag) as an example of identification information specified to distinguish the important word from other important words, and the contents of the <span> tag of that class name in the CSS file. It is configured to include information that defines the display color as a display mode. In this case, the CSS file may include, for example, information indicating a color as a display mode of the content of the <span> tag for each of the plurality of pieces of identification information (for example, plural class names). The information indicating the color to be added to the important word/phrase may be, for example, the font color, character background color, etc. for the character string of the important word/phrase, or at least one side of the top, bottom, left, or right of the minimum rectangular border surrounding the character string of the important word/phrase. This is color information that expresses the line color of one side when a line width is given, for example, using RGB values. For example, the output unit 108 adds different colors (i.e., display colors) to each of a plurality of important phrases in the complaint information whose character strings that constitute the important phrases extracted by the important phrase extraction unit 1051 are different from each other. and display it. For example, for a plurality of important terms whose character strings are different from each other, class names are determined to be different from each other, and the display color as the display mode of the contents of the <span> tag for each class name is set. This can be achieved by determining the color information to indicate different colors for each class. Further, the output unit 108 may output any one of the above-mentioned key terms and phrases in which the character strings constituting the key terms extracted by the key term extraction unit 1051 in the complaint information are different from each other. Important words and phrases in the claim information that are composed of the same character string may be displayed in the same color (that is, the same display color) as that one of the above-mentioned important words and phrases. Adding the same display color between such important words can be achieved by using classes with the same class name related to display colors. If the number of important terms that differ from each other exceeds a predetermined number (for example, 500, etc.), they will be sorted in a fixed order based on the importance of the important terms or attribute information (for example, first line number), etc. Important words and phrases that differ up to a certain number may be expressed in colors with different RGB values, and important words and phrases that exceed a certain number may be expressed in the same color using one predetermined RGB value.

When the invention document acquired by the document acquisition unit 102 includes a specification, the output unit 108 may include, for example, specification information representing the content of the specification in the patent application related information. good. In patent application related information as a display document, specification information is one of the contents to be displayed.

The specification information representing the content of the specification of the invention document in the display document is included in the additional information related to each of the plurality of important terms in the specification extracted by the important term extraction unit 1051. The displayed display control information includes information for controlling the display mode of the important term, for example, information indicating the color to be added to the important term.

For example, for each of the plurality of important words and phrases extracted by the important word and phrase extraction unit 1051 in the specification information, the output unit 108 determines whether the important word is one of the plurality of important words and phrases extracted by the important word and phrase extraction unit 1051 in the claim information. If the key words and phrases in the description information are the same, the key words and phrases in the description information are displayed in the same display mode as the same key words and phrases in the claim information (for example, displayed with the same color added), and the key words and phrases in the description information are is different from any of the plurality of important words extracted by the important word extraction unit 1051 in the claim information, the important word in the specification information is displayed in a manner different from any of the plurality of important words in the claim information. May be displayed. Note that, as described above, the output unit 108 can use a browser or the like to display in an identifiable manner whether or not the important words in the specification information are the same important words as the important words in the claim information. It is also possible to configure and output a display document in which display control information for each important phrase in the specification information is appropriately determined, as shown in FIG. The display control information for one important word in the specification information in the display document is, for example, if that one important word is the same as any of the important words extracted by the important word extraction unit 1051 in the claim information, The first important phrase in the specification information is displayed in the same display mode as the same important phrase in the claim information, and the first important phrase in the specification information is extracted by the important phrase extraction unit 1051 in the claim information. If it is different from any of the important words/phrases, it includes information for controlling the important word/phrase in the specification information to be displayed in a display mode different from any of the extracted important words/phrases in the claim information.

When the output unit 108 includes claim information and specification information in the patent application related information generated as a display document based on the analysis result of the invention document by the analysis unit 105, the important term extraction unit 1051 extracts the information from the invention document. Among a plurality of extracted important words and phrases in the claim information or specification information, the same important words and phrases as the words and phrases representing the element names described with the addition of codes in the invention document are related to the important words and phrases. The display control information as additional information is, for example, a key word or phrase in a drawing document (for example, a searchable PDF file) acquired by the document acquisition unit 102 corresponding to the invention document according to the operation. An operation-compatible GUI element (i.e., an operation-compatible GUI element for displaying the part where the code is written (for example, the page, position, etc. where the code is written, identified by searching for the code in a drawing document) Contains information to be displayed as a display element (with a GUI function).

In the example of displaying patent application related information, the output unit 108 converts the display mode of a plurality of key words extracted by the key word extraction unit 1051 in the claim information and specification information in the patent application related information to the display mode of the plural key words and phrases. It is also possible to generate a display document that includes display control information and the like that includes highlighting information for displaying words in a manner different from that of other words and phrases, and to perform display according to the display document. As an example of realizing this display by a browser or the like that performs display based on an HTML document as an example of a display document, the output unit 108 outputs, for example, a plurality of important information regarding patent application related information as an HTML document. As display control information for controlling the display mode of words other than phrases, a character color, a character background color, etc. are determined as default values for a character string. Then, in the display control information that controls the display mode of each of the plurality of important words and phrases, a <span> tag whose content is the important word of the class name defined so that each of the plurality of important words can be identified as described above is added. Includes emphasis display information such as color information that determines the font color, font background color, etc. as a display mode of the content to be other than the default value.

In the example of displaying patent application related information, the output unit 108 extracts each of the plurality of important terms extracted by the important term extraction unit 1051 from the claim information in the patent application related information, and extracts the important terms according to the operation. It may also be displayed as a GUI element for displaying additional information related to the phrase. As an example of realizing this display by a browser or the like that performs display based on the display document, the output unit 108 controls the display of each important word in the patent application related information as the display document. The information includes, for example, information for displaying additional information regarding the important phrase, and information for displaying the important phrase as a GUI element for executing additional display of information regarding the important phrase in response to an operation. include.

Adding information related to important words and phrases is a process of adding and displaying additional information related to the important words and phrases, such as the importance level, attribute information, and translated information about the important words. (For example, corresponding translation words, translation word candidate set), etc., or displaying search results related to the important words and phrases. The importance level, attribute information, and search results regarding important words and phrases may be specified by any method, and may be the same as those shown in Embodiment 2, which will be described later.

An example of the claim information to be displayed when the output unit 108 outputs patent application related information including claim information representing the contents of the claims of the invention document acquired by the document acquisition unit 102 as a display document is In addition to the contents, for each of the plurality of important words and phrases extracted by the important words and phrases extraction unit 1051 in the content of the claim, the part of the important word in the content of the claim is converted into a corresponding translated word or translated word corresponding to the important word. The candidate set may include replaced content. In this case, the claim information includes, for example, each part (for example, one line of description) expressing a claim in the invention document, followed by important words and phrases in the description copied from that part, and the corresponding translated words. Alternatively, a substituted word may be inserted into the set of translation word candidates.

Further, the specification information to be displayed when the output unit 108 outputs patent application related information including specification information representing the content of the specification of the invention document acquired by the document acquisition unit 102 as a display document. For example, in addition to the content of the specification, for each of a plurality of important terms extracted by the important term extraction unit 1051 in the content of the specification, the part of the important term in the content of the specification is added to the important term. The replaced content may be included in the corresponding translation word or translation word candidate set. In this case, the specification information includes, for example, each part (for example, one line of description) representing the specification of the invention document, followed by important words and phrases in the description copied from that part. It may also be an embodiment in which a substituted word is inserted into the translated word or the set of translated word candidates.

For each of the plurality of conditions defined in the invention document standard provisions for translation acquired by the invention document standard provisions for translation acquisition unit 104, the inappropriate character string pattern information related to the condition indicates the inappropriate character string pattern information. When a character string corresponding to the character string pattern exists in a claim in an invention document acquired by the document acquisition unit 102, the patent application related information outputted by the output unit 108 is, for example, a combination of the claim and the translation It may include modified version claim information generated by modifying the claim to satisfy the conditions based on the modification information associated with the inappropriate character string pattern information in the invention document standard provisions. Some of the conditions among the plurality of conditions defined in the invention document standard regulations for translation may be determined, for example, based on the selection information defined in the setting information, or based on the selection operation of the user of the invention document analysis system 10. can be selected by. In addition, for each of the plurality of conditions defined in the standard provisions of invention documents for translation acquired by the standard provisions of invention documents for translation acquisition unit 104, inappropriate character string pattern information related to the condition indicates If a character string that corresponds to the inappropriate character string pattern exists in the specification of the invention document acquired by the document acquisition unit 102, the patent application related information output by the output unit 108, for example, and the modification information associated with the inappropriate character string pattern information in the standard provisions for invention documents for translation, and amended version specification information generated by modifying the specification to satisfy the conditions. may include.

When the invention document acquired by the document acquisition unit 102 includes a claim, a predetermined section of the claim (for example, one claim) cites multiple sections (for example, multiple claims). When the output unit 108 outputs the patent application related information generated based on the analysis result of the invention document by the analysis unit 105, the patent application related information is generated based on the analysis result of the invention document by the analysis unit 105. Amended claim information representing the result of amending the claim so as to include content that has been changed to cite only the first paragraph, and the citation relationship of each clause of the claim represented by the amended claim information, An expression in which the item numbers of items that have a mutual citation relationship are associated using lines or symbols, and may include a modified claim tree as the expressed information.

The patent application related information output by the output unit 108 and generated based on the analysis result of the invention document by the analysis unit 105 is, for example, between each of a plurality of important words extracted from the invention document by the important word extraction unit 1051. It may also include important word/phrase relationship information indicating the relationship between. This important term/phrase relationship information includes, for example, inclusion relationship information indicating a structural inclusion relationship between elements represented by each of a plurality of important terms/phrases. For example, if an invention document describes an "information processing device" that includes "input section" and "display section," "information processing device," "input section," and "display section" are extracted as important terms. , the inclusion indicates a relationship in which the element "information processing device" represented by the important word includes the "input section" and "display section" as the element group represented by the important word. The output unit 108 can output patent application related information including related information.

Further, the patent application related information output by the output unit 108 may include, for example, the result of machine translation of the invention document by the translation unit 1054.

The output unit 108 includes, for example, a learning model output unit 1081.

A display document (for example, an HTML document) containing display control information for displaying each of the plurality of important words outputted by the output unit 108 as a GUI element that can be selected from a set of translation word candidates is displayed in a browser or the like. When the key word is displayed, a script included in the display control information determines an element selected according to the operation from a set of translation word candidates for the key word as a corresponding translated word corresponding to the key word. , and a process of generating and outputting translation auxiliary information for each important word/phrase for which a corresponding translation word has been determined as a result of the determination process.

As described above, the learning model output unit 1081 outputs each of the plurality of translation auxiliary information generated as a result of displaying each of the display documents outputted for each of the plurality of invention documents acquired by the document acquisition unit 102. Outputs a learning model for important word translation, which is a learning model generated by machine learning, using pairs of important words and corresponding translated words in , as training data for machine learning for important word translation. do. This machine learning is realized, for example, by using a general machine learning algorithm and performing arithmetic processing using sets of important words and phrases having a certain correlation and their corresponding translated words as training data. Examples of general machine learning algorithms include neural networks that are composed of an input layer, an output layer, and a hidden layer, with weights representing the degree of information transmission between the layers, and deep learning (deep learning) that includes many hidden layers. (Learning) etc. In deep learning, learning is performed by adjusting the weights of each layer of the neural network in order to bring the pair of input and output of the neural network closer to the training data (that is, to reduce the error). As a method for adjusting the weights, for example, back propagation may be used.

Furthermore, the learning model output unit 1081 may have a function of outputting the translation learning model generated by the machine learning unit 107, for example. The output format of the learning model (important term translation learning model or translation learning model) by the learning model output unit 1081 may be in any format that can be referenced by the translation unit 1054 or other processing system for machine translation, for example. There may be. For example, the learning model output unit 1081 outputs the learning model as the contents of a file, database data, etc. in a recording medium inside or outside the invention document analysis system 10.

[2-2. Example of invention document analysis processing according to Embodiment 1]
An example of an invention document analysis process executed by the invention document analysis system 10 having the above-described configuration will be described.

FIG. 7 is a flowchart illustrating an example of invention document analysis processing.

The invention document analysis process includes, for example, a setting information acquisition step S1, a document acquisition step S2, an analysis step S3, and an output step S4.

In a setting information acquisition step S1, the invention document analysis system 10 acquires setting information (see FIG. 5) including condition information indicating conditions for extracting important words.

The invention document analysis system 10 acquires an invention document in document acquisition step S2.

In an analysis step S3, the invention document analysis system 10 analyzes the invention document acquired in the document acquisition step S2 based on the setting information acquired in the setting information acquisition step S1. The analysis step S3 is a process that includes an important phrase extraction step S31.

In an important term extraction step S31, the invention document analysis system 10 extracts important terms from the invention document acquired in the document acquisition step S2, based on the condition information of the setting information acquired in the setting information acquisition step S1.

In an output step S4, the invention document analysis system 10 outputs information generated based on the analysis result in the analysis step S3 for the invention document acquired in the document acquisition step S2. In the output step S4, the invention document analysis system 10 outputs information generated based on each of the plurality of important phrases extracted in the important phrase extraction step S31 from the invention document acquired in the document acquisition step S2. An example of the information output by the invention document analysis system 10 in the output step S4 is patent application related information (see FIG. 6). The invention document analysis system 10 acquires an invention document and a corresponding translation result document in a document acquisition step S2, and performs machine learning using a pair of an invention document and a translation result document that meet a certain condition as training data in an output step S4. A learning model for translation or a learning model for important phrase translation that is generated as a result may be output.

In the important term extraction step S31, for example, important term candidate character string pattern information indicating an important term candidate character string pattern, reference character string pattern information indicating a reference character string pattern, and a code indicating a code string pattern in the condition information are extracted. By referring to the character string pattern information and the specific particle string pattern information indicating the specific particle string pattern, the first character string corresponding to the reference string pattern is included in the invention document acquired in the document acquisition step S2. If there is a second character string that corresponds to the important word candidate character string pattern following , the same character string as the second character string is extracted as an important word and the claim is included in the invention document. In this case, if there is a period character following the third character string that corresponds to the important word candidate character string pattern in the claim, the same character string as the third character string is extracted as the important word and the invention If a fourth character string that corresponds to the important word candidate character string pattern is followed by a fifth character string that corresponds to the code string pattern in the document, the same character string as the fourth character string is marked as important. Extracted as a word or phrase, if there is a seventh character string that corresponds to the specific particle character string pattern following the sixth character string that corresponds to the important word/phrase candidate character string pattern in the invention document, the sixth character Extract the same character string as the column as an important phrase.

Note that the example in FIG. 7 is only an example, and the invention document analysis system 10 may execute a process in which some of the steps shown in FIG. 7 are omitted or changed as the invention document analysis process, or It is also possible to add some processing steps to the plurality of steps shown in FIG. All or some of the steps may be performed in parallel.

[2-3. Example of hardware configuration or network configuration of invention document analysis system]
As described above, the invention document analysis system 10 is realized by, for example, a computer equipped with a memory, a microprocessor, an input device, and a display device. The invention document analysis system 10 may be realized by a device such as one computer, or may be realized by a plurality of devices such as computers. Furthermore, each of the plurality of invention document analysis systems 10 can be used by one or more users.

8 to 10 show an example of the hardware configuration or network configuration of the invention document analysis system 10.

In the example of FIG. 8, the invention document analysis system 10 is configured with one terminal device 10a. The terminal device 10a is, for example, a computer having a browser function that realizes display based on a display document. The terminal device 10a is, for example, a computer equipped with a touch panel or the like and running an OS, and can run an application program for realizing a browser function under the OS. A computer has a configuration in which, for example, a microprocessor (eg, CPU), memory (eg, ROM, RAM), communication interface, connection interfaces such as input/output devices, etc. are connected via a bus. A computer may include input/output devices, for example. An OS, an application program, etc. including a group of instructions is stored in a memory, for example, and the instructions of the group of instructions are read from the memory, interpreted, and executed by a microprocessor. The terminal device 10a may be, for example, a desktop PC, a notebook PC, a smartphone, or the like. The invention document analysis processing program, which is a program for executing the invention document analysis processing (see FIG. 7), may be, for example, one or more application programs executed on the terminal device 10a, or may be one or more application programs executed on the terminal device 10a. It may also be a program (for example, a plug-in module built into the browser) that constitutes a part of the browser executed in the browser. The invention document analysis processing program executed on the terminal device 10a may be written in any language. All or a part of the invention document analysis processing program may be, for example, an executable program written in C language, C++ language, etc. and compiled, or written in Python (registered trademark) language, etc. and executed by an interpreter. It may also be a program that is executed with, etc. In addition, all or part of the invention document analysis processing program may be implemented using macros (for example, Visual Basic (registered trademark), etc. It may also be a program written in

When the invention document analysis processing program is implemented as an application program executed on the terminal device 10a, the application program has, for example, a function of generating and displaying information (for example, patent application related information) based on the analysis results. It may have.

When the invention document analysis processing program is implemented as an application program executed on the terminal device 10a, the application program generates patent application related information as a display document (for example, an HTML document), and displays the patent application related information. The document for display may be displayed by a browser (for example, an application program having a browser function for interpreting and displaying a document for display). For example, the browser can display patent application related information by accessing display documents as a so-called local file group. For example, the application program as the invention document analysis processing program executes the function of a web server and starts a browser, so that the generated display document is accessed by the started browser as a resource on the so-called local host. It may also be displayed.

An example of a display document as an output mode of patent application related information output by the invention document analysis system 10 is an HTML document. In this disclosure, the display document as an output mode of patent application-related information is mainly explained using an example of an HTML document. It may be a document in which the display mode (display color, etc.) of each display element such as a word or phrase can be controlled by adding some display control information, and specifically, for example, Microsoft (registered trademark) It may be a document (so-called DOC file) in a format normally handled by an application program having a document creation and editing function such as Word, etc., and display of GUI elements etc. that can be operated as display control information related to display elements such as important words. It may also be a DOC file or the like with an accompanying macro for control.

Further, the invention document analysis processing program executed on the terminal device 10a may output patent application related information and the like as the contents of a file in any format (for example, a text file).

In the example of FIG. 9, the invention document analysis system 10 includes, for example, one server device 10b and a plurality of terminal devices 10c to 10e connected to the server device 10b via a LAN (Local Area Network). . The terminal devices 10c to 10e are similar to the terminal device 10a, and are used by different users (for example, users located in the same local environment such as the same company, the same company, the same office, the same division, the same department, etc.). It is assumed that it will be used by users. Furthermore, in the example of FIG. 9, it is also possible to operate the LAN as a WAN (Wide Area Network). The server device 10b has, for example, a file server function, and includes, for example, a recording medium (for example, a memory, a hard disk, etc.) for recording files.

For example, the server device 10b holds a file on a recording medium that includes setting information for each of the terminal devices 10c to 10e to access and obtain. For example, the server device 10b holds in a recording medium a file containing as its content the standard provisions for translation invention documents for each of the terminal devices 10c to 10e to access and obtain. For example, the server device 10b collects and integrates translation auxiliary information generated in each of the terminal devices 10c to 10e, holds a file containing the auxiliary translation information as a result of the integration, and stores the auxiliary translation information. It may have a function of distributing file translation auxiliary information to each terminal device connected via LAN.

For example, the server device 10b holds a file containing one bilingual dictionary information or each file containing bilingual dictionary information for each of a plurality of groups (for example, technical fields), and is connected via a LAN. The bilingual dictionary information required by any of the terminal devices 10c to 10e can be distributed to that terminal device.

For example, the server device 10b stores an invention document and information outputted based on the results of an invention document analysis process based on the invention document (for example, a patent application including invention document evaluation information) in each of the terminal devices 10c to 10e. It may have a function of associating and collecting and storing (that is, retaining) related information) and a translation result document created by the user using that information. In the case where the server device 10b has this function, each of the terminal devices 10c to 10e uses a set of some invention documents held by the server device 10b and a translation result document corresponding to the invention documents as training data. It becomes possible to generate a learning model through machine learning and utilize the learning model for, for example, machine translation of an invention document to be analyzed in invention document analysis processing.

In the example of FIG. 10, the invention document analysis system 10 is composed of a plurality of computers distributed in various locations. The invention document analysis system 10 in the example of FIG. 10 includes, for example, one server device 10f, and a plurality of terminal devices 10g and 10h connected to the server device 10f via a LAN or a WAN such as the Internet. The number of terminal devices may be one. The server device 10f may be able to communicate with the server device 10i via a LAN or WAN. For example, data to be communicated between devices connected via a WAN is encrypted using existing encryption technology, and then the data is communicated. The server device 10i may be responsible for executing part of the invention document analysis process, for example.

The terminal devices 10g and 10h are similar to the terminal device 10a, and are assumed to be used by different users. The server device 10f has, for example, a web server function, and includes, for example, a recording medium (for example, a memory, a hard disk, etc.) that can store a file of a display document. The server device 10f includes, for example, a CGI (Common Gateway Interface) function implemented by a program written in Perl language or the like for a part of the invention document analysis process.

Either of the terminal devices 10g and 10h transmits the invention document to the server device 10f using, for example, a browser. The server device 10f that has received the invention document executes an invention document analysis process by requesting the server device 10f to perform some processing as necessary, and as a result, information ( For example, a display document (for example, an HTML document) regarding patent application related information is generated. The server device 10f outputs information based on the analysis result of the invention document by transmitting the display document to the terminal device that is the source of the invention document that is the basis of its generation. Either terminal device 10g or 10h that has received the display document uses a browser to control the display of information based on the display document. This allows the browser to output information based on the analysis results of the invention document. For example, when patent application related information is displayed as a display document by a browser, important words in claim information and important words in specification information are displayed in a display manner (for example, font color, font background color, etc.). Further, the browser may display warning information including indications, evaluations, etc. regarding the invention document. When the warning information is displayed, the user can confirm whether the invention document has a problem. When an operation-compatible GUI element, etc. is displayed on the display device of the terminal device as a result of information display control by the browser, the display control information specified in the display document's display control information is displayed according to the user's operation. Processing is performed and, for example, displayed content may change. Note that, if necessary, the user can freely edit the content displayed by the browser by copying and pasting the content into an application program (for example, a text editor) having a document creation and editing function.

For example, the server device 10f authenticates the access source terminal device (that is, the client terminal) using identification information (ID), password, etc. related to the user, user group, terminal, or terminal group, and registers it in advance. , it is possible to upload invention documents only to predetermined managed client terminals, and it is also possible to display a list of patent application related information as the analysis results of invention documents uploaded in the past. It may also have a management function that allows viewing (that is, downloading), deletion, etc. of patent application related information. In addition, the server device 10f allows multiple client terminals that have been authenticated as the same group based on the designation of the same user group or terminal group ID to view patent application related information as an analysis result of the same invention document. You can also do it. Further, a predetermined client terminal authenticated by ID, password, etc. and the server device 10f may appropriately share and execute each processing step constituting the invention document analysis process.

The server device 10f may have the function of, for example, a search site (so-called search engine) that performs a search in response to a request from a client terminal. In this case, the server device 10f may execute the invention document analysis process using the invention documents acquired as a result of searching for patent documents as the analysis target.

The server device 10f or the server device 10i, for example, based on patent publications published on the Internet by the government offices and international organizations of each country, provides an invention document indicated in the publication and a translation result document corresponding to the invention document. It may have a function of collecting multiple sets of.

[2-4. Effects of invention document analysis system according to Embodiment 1]
The invention document analysis system 10 extracts important words from an invention document created by an inventor or a patent attorney using an important word extraction unit 1051 based on setting information, and generates a document based on each extracted important word. Since the information is outputted by the output unit 108, translators and the like may be able to better understand the invention document by referring to the outputted information. Facilitating the understanding of invention documents will be useful for translators to appropriately or efficiently translate invention documents.

For example, when there is an expression that refers to a word in an invention document, the important word/phrase extraction unit 1051 treats the referenced word as an important word under certain conditions, and extracts a word written immediately before a period character in a claim of the invention document. Since they are treated as important words under certain conditions, words that have been noted and words that express the subject matter of the claim can be treated as important words. For example, the important word/phrase extraction unit 1051 treats a word (for example, a noun phrase, etc.) that precedes a particle in an invention document written in Japanese as an important word under certain conditions, and treats the word/phrase that precedes a particle (such as a noun phrase) as an important word in an invention document written with a code. Since words and phrases are treated as important words and phrases under certain conditions, relatively important words and phrases such as subject and object, as well as words and phrases that represent elements presumed to be illustrated in drawings related to invention documents, can be treated as important words and phrases. . As a result, the invention document analysis system 10 outputs information from the output unit 108 based on important words and phrases that are estimated to be important in the invention document, so the output information can be useful.

For example, the output unit 108 outputs, for each of the plurality of important words extracted from the invention document by the important word extraction unit 1051, the important words and related additional information (for example, the importance of the important words, attribute information, a set of translation word candidates). By outputting patent application related information including Based on the additional information, it may be possible to efficiently or appropriately determine a translated word that corresponds to the important word and has the same meaning as the important word in the destination language. The patent application related information outputted by the output unit 108 includes translated word information (for example, information representing the corresponding translated word itself, a set of translation word candidates, etc.) corresponding to the important word/phrase. By referring to the translated word information, the translator may be able to efficiently or appropriately determine the translated word/phrase corresponding to the important word/phrase. When the output unit 108 outputs patent application related information including a set of translation word candidates corresponding to an important word/phrase, each element of the set of translation word candidates has a value indicating the degree of validity of the corresponding translation word. When it is added, the translator may be able to efficiently or appropriately determine the translated word corresponding to the important word from the set of translated word candidates.

In the invention document analysis system 10, for example, the classification unit 1053 performs group classification based on the important word group extracted from the invention document by the important word extraction unit 1051, and selects bilingual dictionary information for each group (for example, technical field). , selectively uses the bilingual dictionary information of the group specified to which the invention document belongs to specify a set of translation word candidates for the important words and phrases of the invention document. As a result, a group of bilingual dictionary information suitable for the invention document is used, so that an appropriate set of translation word candidates can be specified for the important word/phrase.

The output unit 108 outputs patent application related information as a display document such as an HTML document, and the display document is a display for making display elements of a set of translation word candidates corresponding to important words and phrases into operation-compatible GUI elements. This GUI element includes control information, and can execute a process of determining one element of the set of translation word candidates selected by the operation as a corresponding translation word, and outputting the important word and phrase and the corresponding translation word as translation auxiliary information. . Thereby, the translator can easily determine the corresponding translated word corresponding to the important word/phrase from the set of translated word candidates by selecting the GUI element of the display document. Furthermore, the translator may be able to edit the determined corresponding translated words using a text editor or the like, for example, by copying and pasting from the display content of the display document using a browser or the like. Furthermore, the result of the selection by the translator may subsequently improve the validity of the invention document analysis system 10's identification of a set of translation word candidates corresponding to the important word/phrase via translation auxiliary information.

In the invention document analysis system 10, for example, the learning model output unit 1081 generates pairs of important words and corresponding translated words in the translation auxiliary information by machine learning using them as training data for machine learning for important word phrase translation. Output the learning model for important word/phrase translation. Thereby, machine translation of the important words and phrases in the invention document can be realized through arithmetic processing using the important words and phrases translation learning model and using the important words and phrases as input. By referring to the results of this machine translation, the translator may be able to efficiently or appropriately determine translated words and phrases that are translated important words and phrases in the invention document.

For example, the invention document analysis system 10 extracts important words from an invention document according to the important word setting dictionary, and outputs patent application related information including the important words and the corresponding translated words of the important words determined according to the important word setting dictionary. . This allows inventors, etc. or patent attorneys to create an important word setting dictionary when they want to define a specific word as an important word, or when they want to define a corresponding translation for an important word. , the invention document analysis system 10 outputs appropriate patent application related information. Therefore, the translator may be able to efficiently or appropriately translate the invention document by referring to the patent application related information.

The invention document analysis system 10 identifies, for example, a set of translation word candidates corresponding to important words in an invention document using important words extracted from a plurality of invention documents belonging to the same group as the invention document and corresponding translation result documents. Machine translation is performed through arithmetic processing using a learning model for important phrase translation that is generated as a result of machine learning using pairs of phrases and corresponding translated words as training data for machine learning. Thereby, the invention document analysis system 10 can appropriately specify the set of translation word candidates.

The invention document analysis system 10 outputs patent application related information including claim information representing the claims of the invention document as a display document such as an HTML document, and controls the display mode of important words in the claims in the claim information. Includes display control information. This may make it possible to identify important words and phrases whose display mode is controlled in complaint information displayed by a browser or the like. In the invention document analysis system 10, when the display mode of important words is controlled, for example, by highlighting them (for example, displaying them in bold), it may be possible to quickly grasp important words in claim information. The invention document analysis system 10 uses, for example, a display element of an important phrase in claim information of a display document as an operation-compatible GUI element, and displays the importance level, attribute information, corresponding translation word, or translation of the important phrase according to the operation. Display control information for displaying the word candidate set or the search results of the important words can be included in the claim information. As a result, translators who view claim information in display documents can obtain various information such as importance and attribute information by operating the displayed important words and phrases. In addition, it may be possible for the translator to efficiently or appropriately determine the translated words for the important words.

The invention document analysis system 10 outputs patent application related information including specification information representing the specification of the invention document as a display document such as an HTML document, and includes important words and phrases in the specification in the specification information. A display control that includes display control information for controlling the display mode and displaying in a distinguishable manner whether or not the important word is stated in a claim, and also controlling the display mode of different important words and phrases to be different. May contain information. As a result, translators, etc. who view the display document displayed on a browser etc. can identify whether or not each important term stated in the specification information is an important term stated in the claim. , it may be possible to quickly understand whether a plurality of important words and phrases are the same or different depending on the display mode of each important word and phrase. The invention document analysis system 10 uses, for example, colors as display modes for mutually different important words (for example, the font color of the important words, the character background color, all sides of a rectangular frame surrounding the important words and having a line width, or any one of them). The display document may include display control information for displaying each important word or phrase in a different color (color of a border line, etc.). As a result, translators who view documents displayed on a browser, etc. can quickly determine whether multiple important words are the same or different based on the color added to each important word. obtain.

For example, the invention document analysis system 10 converts important words and phrases in each content of claim information and specification information in patent application related information output based on the analysis results of invention documents into corresponding translation words or translation word candidate sets. Include the replaced content. As an example, the invention document analysis system 10 may insert, immediately after each line of claim information and specification information, a line in which important words and phrases in the content of that line are replaced with corresponding translated words or a set of translation word candidates. The translator may be able to efficiently translate the invention document by editing the content in which the important words and phrases are replaced with corresponding translation words or a set of translation word candidates.

For example, the invention document analysis system 10 is configured such that key terms in claim information or specification information of patent application-related information as a display document output based on the analysis result of an invention document are detected by a document acquisition unit in a manner corresponding to the invention document. If the description of an element name with a code included in the drawing in the drawing document obtained by step 102 is a word or phrase representing the element name included in the invention document, the display element of the important word or phrase is Accordingly, a display document is generated so as to add display control information for providing a function as a GUI for displaying the portion of the drawing document in which the code is written. As a result, translators who have viewed a display document displayed on a browser, etc. can manipulate key words to display and confirm the part of the drawing document that corresponds to the element represented by the key word. This makes it possible to understand the important words and phrases efficiently and appropriately.

For example, the invention document analysis system 10 analyzes the invention document based on the determination result of whether or not it satisfies some of the conditions related to the invention document standard regulations for translation, which define a plurality of conditions that the invention document to be translated should satisfy. Outputs patent application related information including warning information pointing out potentially inappropriate parts of the document. By referring to this patent application related information, the inventor or patent attorney who created the invention document can modify the invention document if necessary. In addition, the translator can, for example, obtain the invention documents to be translated and the patent application related information from the inventor or patent attorney, so that even if the warning information in the patent application related information remains, the translator can It can be assumed that the inventor or patent attorney has determined that there is no need to take particular action on the indication, so it is possible to obtain an appropriate invention document to be translated. It may be possible to do so efficiently or properly. For example, the invention document analysis system 10 can narrow down some of the conditions used in the judgment for pointing out warning information among the plurality of conditions of the standard provisions for invention documents for translation, according to the specifications in the setting information, etc. You can do it like this. As a result, inventors, etc. or patent attorneys can make the invention document analysis system 10 output patent application related information that does not include unnecessary indications in warning information by setting the setting information, and the patent application By providing relevant information to the translator, it is possible to prevent the translator from being misled by the warning information.

For example, the invention document analysis system 10 adds invention document evaluation information indicating the conformity of the invention document to the invention document standard provisions for translation to warning information in the patent application related information generated based on the analysis result of the invention document. may be included. The suitability is expressed, for example, by an evaluation value as a result of the suitability evaluation. By referring to this invention document evaluation information, an inventor or a patent attorney may be able to appropriately modify the invention document. Furthermore, by checking the invention document evaluation information, the translator may be able to recognize whether the invention document is suitable for translation. Invention documents containing inappropriate descriptions, for example, are prone to translation errors, and a situation may arise in which a translator modifies an invention document before translating it in order to obtain appropriate rights. Therefore, if the warning information in the patent application related information generated based on the invention document includes invention document evaluation information indicating that the conformity with the standard provisions of invention documents for translation is relatively low, It can be determined that there may be discrepancies between the invention document and the translation result document that is the result of translation based on the invention document, in addition to the difference in language. Because of the generation, it may be possible to determine that the data is not suitable for use as training data for machine learning.

The invention document analysis system 10 generates revised claim information in which potentially inappropriate descriptions of the claims of the invention document are corrected, and the specification of the invention document, based on, for example, correction information in the standard provisions for invention documents for translation. Outputs patent application related information including revised specification information that corrects potentially inappropriate descriptions. An inventor or patent attorney who has created an invention document may be able to efficiently modify the invention document by referring to the revised claim information and revised specification information. By referring to the revised claim information and revised specification information and being provided with the revised invention document, the translator can efficiently or appropriately translate the invention document.

The invention document analysis system 10 may, for example, change a clause in a claim of an invention document that cites multiple clauses to a clause that cites only one clause with the first item number among the multiple clauses; Patent application related information including a revised claim tree representing the citation relationship of each clause of the claim represented by the revised claim information is output. The invention document analysis system 10, for example, displays a claim tree representing the citation relationship between each clause of the claims of an invention document in patent application related information as a display document, including an operation-compatible GUI element that can change the citation relationship. The revised claim tree and revised claim information are updated to match the citation relationships that have been changed in response to operations on the claim tree displayed based on the display document in a browser, etc. obtain. In this case, by manipulating the claim tree, the inventor, etc. or the patent attorney can provide the translator with a modified claim tree that reflects the manipulation and patent application related information including the modified claim information. A translator who receives information related to a patent application from an inventor, etc. or a patent attorney, and who is informed that the claim portion of an invention document should be translated based on the revised claim information, cites multiple clauses in the invention document. It may be possible to translate efficiently without having to worry about whether or not to revise a section, and if so, how to revise it. In addition, a translator or the like can quickly grasp the citation relationship between each clause of a claim expressed in the revised claim information using the revised claim tree.

The invention document analysis system 10 outputs, for example, patent application related information including important word/phrase relationship information indicating the relationship between each of a plurality of important words/phrases extracted from the invention document by the important word/phrase extraction unit 1051. This important word/phrase relationship information includes, for example, inclusion relationship information indicating a structural inclusion relationship between elements represented by each of a plurality of important words/phrases. As a result, translators who have obtained patent application-related information can relatively quickly understand the relationships between multiple important words in invention documents, and by understanding those relationships, they can efficiently select translated words that correspond to each important word. Alternatively, it may be possible to determine the invention appropriately, and as a result, it may be possible to translate the invention document efficiently or appropriately.

In the invention document analysis system 10, for example, the classification unit 1053 performs group classification of invention documents based on the correlation between important word groups extracted by the important word extraction unit 1051 from each of a plurality of invention documents, and the machine learning unit In step 107, a learning model for translation is generated for each group by machine learning using a plurality of sets of an invention document and a corresponding translation result document as training data. As a result, in a machine translation system, for example, it becomes possible to machine translate an invention document through arithmetic processing using a translation learning model suitable for the group to which the invention document belongs, among translation learning models for each group. In addition, the invention document analysis system 10 may machine translate at least a portion of the invention document through arithmetic processing using, for example, a translation learning model suitable for the group to which the invention document belongs among the translation learning models for each group. , patent application related information including the machine translation results may be output. By referring to the results of machine translation of such an invention document, a translator may be able to efficiently or appropriately translate the invention document.

In the invention document analysis system 10, for example, the document evaluation unit 106 selects an invention document that satisfies the conditions related to the standard provisions of invention documents for translation to a degree exceeding a predetermined degree, and the machine learning unit 107 selects the invention document that satisfies the conditions related to the standard provisions for invention documents for translation. A learning model for translation for machine translation is generated by machine learning using multiple sets of an invention document and a translation result document corresponding to the invention document as training data, and the learning model for translation is generated by the learning model output unit 1081. Output. As a result, a combination of an invention document containing inappropriate descriptions and a translation result document whose contents are inconsistent with the invention document due to translation errors etc. generated based on the invention document is difficult to be used as training data. An appropriate translation learning model can be output from the invention document analysis system 10. Thereby, for example, in a machine translation system or the like, it becomes possible to machine translate an invention document through arithmetic processing using an appropriate translation learning model output from the invention document analysis system 10. By referring to the machine translation result of the invention document, the translator may be able to efficiently or appropriately translate the invention document.

For example, the invention document analysis system 10 uses the machine learning unit 107 to generate a revised invention document in which inappropriate descriptions in the invention document are corrected based on correction information included in the invention document standard regulations for translation, and a modified version of the invention document in which inappropriate descriptions in the invention document are corrected based on correction information included in the invention document standard regulations for translation. A translation learning model for machine translation is generated by machine learning using a set with a corresponding translation result document as training data, and the learning model output unit 1081 outputs the translation learning model. Although the invention document and the translation result document generated by correcting inappropriate descriptions in the invention document at the translation stage have language differences, the discrepancies in their content have been reduced. Therefore, there is a possibility that the translation learning model output from this invention document analysis system 10 is appropriate. Thereby, for example, in a machine translation system, it becomes possible to machine translate an invention document through arithmetic processing using a potentially appropriate translation learning model output from the invention document analysis system 10. By referring to the machine translation result of the invention document, the translator may be able to efficiently or appropriately translate the invention document.

The invention document analysis system 10 is configured, for example, by installing an invention document analysis processing program for executing an invention document analysis process as illustrated in FIG. 7 into a device including a microprocessor. When the invention document analysis processing program is started in the invention document analysis system 10, invention document analysis processing is executed. In the invention document analysis process, for example, an invention document created by an inventor or a patent attorney, acquired in document acquisition step S2, is analyzed in analysis step S3, based on the setting information acquired in setting information acquisition step S1. This is a process of analyzing the information and outputting the information generated as a result of the analysis in an output step S4. The analysis step S3 includes an important word extraction step S31 that extracts important words from the invention document. For translators and the like, understanding of the invention document may be facilitated by referring to the information output by the invention document analysis process. Facilitating the understanding of invention documents will be useful for translators to appropriately or efficiently translate invention documents. Note that the invention document analysis system 10 can output, for example, a display document for displaying patent application related information in the output step S4. In the invention document analysis process performed by the invention document analysis system 10, in step S4, a GUI screen is displayed that allows selection of one of the one or more display documents that have been output in the past, and the selected document is selected. The display document may be interpreted and displayed by a browser in a device including a microprocessor.

(Embodiment 2)
Another example of the embodiment of the invention document analysis system will be described below with reference to the drawings.

The invention document analysis system shown in this embodiment is basically a system that executes invention document analysis processing for analyzing invention documents, similar to the invention document analysis system 10 shown in the first embodiment. The invention document analysis system shown in this embodiment corresponds to a partial modification of the invention document analysis system 10 shown in Embodiment 1, and the invention document analysis process in this embodiment is the invention document analysis system 10 shown in Embodiment 1. This corresponds to a partially modified version of the process. The invention document analysis system according to this embodiment may include all the components of the invention document analysis system 10 shown in Embodiment 1, as long as they do not conflict with the matters described in this embodiment. Regarding the invention document analysis system according to this embodiment, descriptions of matters similar to those related to the invention document analysis system 10 shown in Embodiment 1 will be omitted as appropriate.

[3-1. Configuration example of invention document analysis system according to Embodiment 2]
FIG. 11 shows an example of the configuration of the invention document analysis system 20. Further, FIG. 12 shows an example of main data types handled by the invention document analysis system 20. The invention document analysis system 20 can acquire, for example, setting information, plug-in modules, invention document standard provisions for translation, important word/phrase setting dictionary, bilingual dictionary information, translation auxiliary information, invention documents, drawing documents, translation result documents, etc. . The invention document analysis system 20 includes, for example, a setting information display document, patent application related information, important word group information, training data used in machine learning for generating various learning models, a learning model for translation, and important word phrase translation. A learning model, a learning model for specifying a technical field, etc. can be output.

The invention document analysis system 20 is a system that executes an invention document analysis process that acquires an invention document and analyzes the invention document.

The invention document analysis system 20 is a system for analyzing invention documents and outputting information such as patent application related information (see FIG. 12) based on the analysis results, and includes, for example, a microprocessor that executes a program. This is realized by equipment, etc. For example, similar to the invention document analysis system 10 shown in the first embodiment, the invention document analysis system 20 outputs patent application related information as a display document. Further, the invention document analysis system 20 may output the patent application related information as a file in various formats such as a text file.

Similar to the invention document analysis system 10 shown in the first embodiment, the invention document analysis system 20 may have, for example, the hardware configuration or network configuration shown in any one of FIGS. 8 to 10. The invention document analysis system 20 may be realized by a device such as one computer, or may be realized by a plurality of devices such as computers. Furthermore, each of the plurality of invention document analysis systems 20 can be used by one or more users. Here, the description will be continued assuming an example in which the invention document analysis system 20 is realized by a computer equipped with an OS including a file system and other programs.

Users of the invention document analysis system 20 are assumed to be, for example, creators of invention documents (eg, patent attorneys, inventors, etc.), translators who perform processing using invention documents, and the like. Here, the translator etc. is a translator who translates the invention document into a foreign language and creates a translation result document for foreign application, and a translator who performs research, analysis, correction, evaluation, opinion statement, etc. regarding the invention document. This may include a person who performs various processes using invention documents.

As shown in FIG. 11, the invention document analysis system 20 includes a setting information acquisition section 201, a document acquisition section 202, an important word/phrase setting dictionary acquisition section 203, an invention document standard provision acquisition section for translation 204, an analysis section 205, and a machine learning section. 207 and an output section 208. The functions of the setting information acquisition unit 201, document acquisition unit 202, important word/phrase setting dictionary acquisition unit 203, invention document standard provision acquisition unit for translation 204, analysis unit 205, machine learning unit 207, and output unit 208 are configured to, for example, It is realized by one or more computers that execute all or part of the analysis processing program. The analysis unit 205 includes an important term extraction unit 2051 that executes an important term extraction process for extracting important terms from an invention document, a warning information generation unit 2059 that can generate warning information based on the invention document, and the like.

[3-1-1. Setting information acquisition section]
The setting information acquisition unit 201 is a partial modification of the setting information acquisition unit 101 shown in the first embodiment. Setting information acquisition section 201 and setting information acquisition section 101 may be the same. The setting information acquisition unit 201 acquires setting information that the invention document analysis system 20 uses as a basis for analyzing an invention document.

The setting information may exist as any data such as data in a database or data as an environment variable of the OS, and may exist as the contents of a file, for example. The file is, for example, provided inside a device (for example, a computer) as one component of the invention document analysis system 20, or provided in an external device that can communicate with the device constituting the invention document analysis system 20. , a memory, a hard disk, various optical disks, a magneto-optical disk, etc., or may be recorded in a memory card, a USB (Universal Serial Bus) memory, etc., which is detachable from any of these devices.

The configuration information acquisition unit 201 may acquire configuration information, for example, by reading a configuration information file that is a file that includes configuration information as content. The configuration information acquisition unit 201 may acquire the configuration information by reading a plurality of configuration information files and integrating the contents of the multiple files according to a predetermined priority order for inconsistencies. The setting information file may define general setting information regardless of the group such as the technical field. Furthermore, within one setting information file, multiple types of setting information may be defined for each group, such as by technical field or by the department to which an inventor belongs, or for each group, separate types of setting information containing different setting information may be defined. A configuration information file may also exist. A specific example of the department to which the inventor belongs in the present disclosure is the division, department, etc. to which the inventor belongs within one company or within one company, or the company, company, etc. to which the inventor belongs. .

The setting information acquisition unit 201 specifies the group to which the invention document acquired in the invention document analysis system 20 belongs based on general setting information, and from the setting information including the setting information corresponding to the specified group, Setting information corresponding to the identified group may be acquired. The configuration information acquisition unit 201 also acquires configuration information in a configuration information file specified by the user among multiple configuration information files for each group, or configuration information specified by the user among multiple types of configuration information in one configuration information file. The configuration information may also be acquired.

FIG. 13 shows an example of the contents of the setting information.

The setting information may include, for example, the invention document standard provisions for translation itself or information for specifying the invention document standard provisions for translation. The information for specifying the standard provisions for invention documents for translation is, for example, information necessary to access the standard provisions for invention documents for translation. The invention document standard provisions for translation may exist as any data such as data in a database, data as an environment variable of an OS, or, for example, as the contents of a file. Specific information for specifying the standard provisions for invention documents for translation includes, for example, information such as a file path necessary to access the file containing the standard provisions for invention documents for translation, or These are URLs (Uniform Resource Locators) and the like necessary for accessing data as standard specifications for invention documents.

The setting information includes condition information indicating conditions for extracting important words from the invention documents to be analyzed by the invention document analysis system 20. The condition information may be the content shown in Embodiment 1, or may be partially modified. The condition information includes, for example, important word/phrase candidate character string pattern information, reference character string pattern information, code character string pattern information, specific particle character string pattern information, and the like. Various character string patterns related to condition information (for example, important word candidate character string patterns indicated by important word candidate character string pattern information) are expressed, for example, in a manner in which one or more corresponding character strings of one or more characters are listed. It may be a character string pattern, or it may be a character string pattern expressed in a manner that limits the character code range in which all or part of the relevant character string is a character string in a certain range, It may also be a character string pattern expressed in a manner with a so-called wild card designation in which a part of the relevant character string is a character string of the entire range.

Important word candidate character string pattern information includes predetermined endings specific to technical terms (for example, "device", "board", "system", "means", "part", "step", "medium") , "agent," "file," etc.) may include information representing a character string pattern that corresponds to a technical term-specific ending character string. The key term candidate character string pattern information is a predetermined word or phrase that is presumed to be a technical term (for example, "reception part", "reception part", "reception part", "screw", "sieve", etc.) or It may also include information representing a character string pattern that corresponds to a character string that includes the word at the end. Further, the setting information may include technical term-specific ending information indicating one or more technical term-specific ending character strings.

Basically, a character string corresponding to the important word/phrase candidate character string pattern indicated by the important word/phrase candidate character string pattern information may be extracted as an important word/phrase by the important word/phrase extraction unit 2051. However, the condition information is an extraction exception that indicates a string pattern for a string that should not be extracted as an important word even if it corresponds to the important word candidate character string pattern indicated by the important word candidate character string pattern information. It may include information related to conditions. For example, a character string that corresponds to the character string pattern indicated by the information related to the extraction exception condition is not extracted as an important word by the important word/phrase extraction unit 2051.

The important word/phrase candidate character string pattern information may be information indicating a plurality of important word/phrase candidate character string patterns, for example. For example, the important word candidate character string pattern information is a predetermined character string in which the first character is an alphabetic character, a kanji, or a katakana, and the last character is an alphabetic character, a katakana, a kanji, or a predetermined hiragana. It is determined to correspond to one important word/phrase candidate character string pattern. The important word candidate character string pattern information is one of the predetermined hiragana (for example, "e", "ki", "ke", "shi", "mi", "ri", "re", etc.) immediately after the kanji. A character string including one or more sets in which one character is arranged may be determined to correspond to one important word/phrase candidate character string pattern.

A character string pattern related to a character string that is a candidate for an important word/phrase indicated by the important word/phrase candidate character string pattern information is expressed, for example, by a regular expression. For example, when the important word extraction process executed by the important word phrase extraction unit 2051 is realized by a computer that executes a program written in the Python (registered trademark) language, various types of The string pattern is, for example, the standard library "re" for Python (registered trademark) regular expression operations (for example, string search), or a library provided by a third party that extends the functionality of "re". This is a character string pattern that can be expressed using regular expressions, such as "regex".

In important word extraction processing, one example of a specific reference string pattern for extracting reference strings from invention documents by performing a string search using the function of the library "re" is "' It is a string pattern that lists reference strings such as (said|above|that|that|mentioned|... (hereinafter omitted)').You can also search for strings using the function of the library "regex". One example of a specific important word candidate character string pattern for extracting important word candidate character strings from invention documents by performing the following is generally "r'([A-Za-zA-Za-z]+ ([ //::][A-Za-zA-Za-z]+)+)*[A-Za-zA-Za-z]*[ー・\p{Katakana}\p{Han}]+ )...(hereinafter omitted)' is a character string pattern. Here, "\p{Katakana}" indicates katakana, and "\p{Han}" indicates kanji. There may be multiple key word candidate character string patterns. For example, some specific examples are "r'([A-Za-zA-Za-z]*[ー・\p{Katakana}\p {Han}]*[A-Za-zA-Za-z]*[----・／/]?[A-Za-zA-Za-z]+)', "r'([A-Za -zA-Za-z]*[ー・\p{Katakana}\p{Han}]+)', "r'([A-Za-zA-Za-zー・\p{Katakana}]* [＼p{Han}]+[Iuekikukeshisusechitsutenanehimimumerirururegigejizejibibe][ー・＼p{Katakana}＼p{Han}]*[A-Za -zA-Za-z]+[---・／/]?[A-Za-zA-Za-z]+)', "r'[A-Za-zA-Za-zー・\p {Katakana}]*[\p{Han}]+[Iuekikukeshisusechitsutenanehimimummerirururegigejizejibibe][A-Za-zA-Za-z]*[ー・\p{Katakana}\p{Han}]+)', ``r'([A-Za-zA-Za-zー・\p{Katakana}]*[\p{Han}]+[I [\p{Han}] ][ー・\p{Katakana}\p{Han}]*[A-Za-zA-Za-z]+[---・／/]?[A-Za-zA-Za-z]+) ''', 'r'([A-Za-zA-Za-zー・\p{Katakana}]*[\p{Han}]+[Iuekikukeshisusechitsutanehimimumerirure Gigejizejibibe][\p{Han}][Iuekikukeshisusechitsutenanehimimummerirururegigejizejibibe][A-Za-zA-Za-z]*[ -\p{Katakana}\p{Han}]+)' etc. can be used to search for important word candidate character strings. In addition, the code string pattern indicated by the code string pattern information is, for example, a pattern determined to correspond to a character string that can be used as a code, and furthermore, the code string pattern that follows the opening parenthesis is used as a code. It may be a pattern that is determined to correspond to a character string in which one or more characters that can be used are arranged, and a closing parenthesis is placed after the one or more characters. Characters that can be used as codes include, for example, Arabic numerals, English letters, Greek letters, Roman numerals, and some symbols. One example of a specific code string pattern for extracting a character string used as a code from an invention document by performing a string search using the function of the library "re" is "'([0 -90-9A-Za-zA-Za-z][----・／/0-90-9A-Za-zA-Za-z]*)...(hereinafter omitted)' etc. Patterns It is. In addition, the specific particle string pattern indicated by the specific particle string pattern information can be used to extract a string that is a predetermined particle from an invention document by performing a string search using the function of the library "re". One example is a pattern such as "'(to|to|to|and|to|to|is|...(partially omitted)...)'".

The setting information may include, for example, information for specifying the important word/phrase setting dictionary (for example, information used to access the important word/phrase setting dictionary file). Further, the setting information may include information for specifying translation auxiliary information (that is, information that associates an important word with a corresponding translated word for each of one or more important words), and the invention document analysis system 20 uses the setting information Translation auxiliary information may be acquired based on the translation auxiliary information, and a group of important words in the translation auxiliary information may be used as an important word/phrase setting dictionary. Further, the setting information may include, for example, identification information of bilingual dictionary information. In the case where the invention document analysis system 20 is composed of the devices shown in FIG. This may be path information including the host name of the server device 10b indicating the location of the server device 10b.

The setting information may include, for example, important word extraction range information for limiting the range of lines in the invention document from which important words are to be extracted. Examples of key word extraction range information include the portion of the invention document that substantially corresponds to the statement of claims (for example, claims), and the portions of the invention document that substantially correspond to the claims (for example, claims), and the portions of “Detailed Description of the Invention” and “Examples” in the specification. An example of this is to limit the range of lines from which important words and phrases are to be extracted to the portion that substantially corresponds to the description. In normal operation, it is assumed that claims should be in the range from which important words and phrases are extracted. It may also be done.

The setting information may include, for example, a non-code criterion that is a criterion for determining that a character string of an invention document is not a code. This non-coded standard may be included in the coded string pattern indicated by the coded string pattern information, or may exist separately from the coded string pattern information. If a non-coded standard exists separately from the coded string pattern information, the important phrase extraction unit 2051 of the analysis unit 205, etc., may detect a character string that corresponds to the coded string pattern indicated by the coded string pattern information. It is possible that the character string may be considered to be inapplicable based on the non-coding criteria.

The setting information may include, for example, comment specifying information for specifying a comment part or a comment line in the invention document. For example, when an invention document is a document in the preparation stage of a patent application, comments are matters included in the invention document for the purpose of alerting, memorandum, annotation, etc., and may be changed ( (for example, deletion). A comment line is a line whose entire contents are comments. Examples of comment identification information include information that indicates that a line containing "■" (that is, a so-called solid block) or "*" (that is, a so-called American stamp) at the beginning is a comment line, and information that expresses that a line containing "■" (that is, a so-called solid block) or "*" (that is, a so-called American stamp) is a comment line, and "■ { " (that is, a so-called solid block and opening curly brace) to "}" (that is, a closing curly brace) is information that expresses that the part is a comment.

The setting information may include, for example, importance standard information indicating a criterion for determining the importance of the important words extracted by the important word extracting unit 2051 from the invention document to be low under certain conditions. This certain condition is, for example, a condition that both the use following the reference character string and the use preceding the code are important words that are not found in the invention document. An example of this importance standard information is low importance ending character string pattern information indicating a low importance ending character string pattern of the ending character string of a word with low importance. Examples of low-importance word-suffix character strings that are word-suffix character strings related to low-importance word-suffix character string patterns include "number," "etc.," "example," and the like.

For example, if there is a plug-in module provided from the outside as an executable program module to replace or expand some of the functions of the invention document analysis system 20, the setting information may be used to access and execute the plug-in module. This may include plug-in module identification information, which is information necessary to do so. As a plug-in module, for example, a program module responsible for processing for determining whether or not an invention document to be analyzed by the invention document analysis system 20 conforms to each of a plurality of items of the invention document standard provisions for translation. , a program module that handles character string patterns to be added for expansion of important word extraction etc. and important word extraction processing using the character string patterns; Examples include a program module for executing a process of outputting a result of arithmetic processing. Specific examples of plug-in module identification information include information such as the file path required to access the file containing the plug-in module, or information necessary to download the plug-in module via the network. This is a URL, etc. The invention document analysis system 20 executes a plug-in module, for example, when executing a function that is replaced or expanded by the plug-in module. If the plug-in module has version information (for example, version number, etc.) attached to it for version control, the invention document analysis system 20 identifies a plug-in module with a newer version (for example, version number) than the plug-in module that has already been acquired. If an in-module is available, delete the already acquired plug-in module and acquire and use the new version of the plug-in module. The invention document analysis system 20 uses an API (Application Programming Interface) for making available information on important words, codes, etc. and various other information obtained by analyzing invention documents by the analysis unit 205 from a plug-in module. etc. may be provided. The invention document analysis system 20 can execute the acquired plug-in module as appropriate to achieve the purpose of the plug-in module.

The setting information may include, for example, category classification standard information indicating criteria for distinguishing to which category the element represented by the important phrase extracted from the invention document by the important phrase extraction unit 2051 belongs. Examples of categories include things, methods, programs, and data. Category classification criteria information is, for example, when an important phrase includes a character string such as "program" or "subroutine" at the end, it indicates that the element represented by the important phrase belongs to the program category. This is the information. Category classification standard information includes, for example, when an important phrase includes a character string such as "method," "step," or "procedure" at the end, the element represented by the important phrase belongs to the method category. This is information that expresses things. Category classification standard information, for example, if an important phrase includes a character string such as "data" or "information" at the end, it indicates that the element represented by the important phrase belongs to the data category. This is the information. The category classification standard information is, for example, information expressing that an element represented by an important term that does not belong to any of the categories of method, program, and data belongs to the category of things.

The setting information includes, for example, the content, insertion position, insertion conditions, etc. of a fixed phrase used when the analysis unit 205 generates specification information representing the content of an incomplete specification based on the claims of an invention document. may include one or more boilerplate information to indicate. An example of standard text information is, for example, in the line next to the heading "Details of Carrying Out the Invention" in the specification, "The drawings used in this disclosure are schematic diagrams and are strictly illustrated. An example of this is information that expresses the insertion of a fixed phrase with the content "It is not...". Examples of insertion conditions in each form information include the condition that the claims of the invention document include a claim that claims a program, a claim that claims a product, and a claim that claims a recording medium. Examples include conditions for inclusion.

The setting information is, for example, information related to the explanation of one or more elements used when the analysis unit 205 generates specification information representing the contents of an incomplete specification based on the claims of the invention document. , may include incomplete message information indicating the content of the incomplete message for writing support. The incomplete message for writing support may include, for example, a character string listing examples of items that should generally be written. The unfinished message information is, for example, a descriptive support that can be employed as an explanation of the element indicated by the important phrase in the claim-based specification information generated by the analysis unit 205 based on the important phrase extracted from the claim of the invention document. This is information that can define different contents of an unfinished message depending on which category the element belongs to, such as object, method, program, data, etc.

The setting information is used, for example, when the analysis unit 205 generates modified claim information representing a claim that has been modified for the region where the application is planned (for example, the country where the application is filed) based on the claims of the invention document. It may include modification information that defines the modification method to be used. For example, if a claim of an invention document has a clause that cites multiple clauses, this modification information expresses that the clause is changed to cite only the one with the smallest item number among the multiple clauses. Information, etc. The modification information may be included in, for example, the standard provisions for invention documents for translation.

The configuration information may include information specifying the format of claims in the revised claim information.

[3-1-2. Document acquisition department]
The document acquisition unit 202 is a partially modified version of the document acquisition unit 102 shown in the first embodiment. The document acquisition unit 202 and the document acquisition unit 102 may be the same. The document acquisition unit 202 includes an invention document acquisition unit 2021 , a drawing document acquisition unit 2022 , and a translation result document acquisition unit 2023 .

The invention document acquisition unit 2021 acquires an invention document to be analyzed by the invention document analysis system 20, for example, according to a specification by a user of the invention document. The invention document acquisition unit 2021 can acquire multiple invention documents. The invention document acquired by the invention document acquisition unit 2021 can be used by the analysis unit 205 and the machine learning unit 207.

The drawing document acquisition unit 2022 acquires the drawing document corresponding to the invention document acquired by the invention document acquisition unit 2021, for example, according to the user's designation of the drawing document corresponding to the invention document. The drawing document acquired by the drawing document acquisition unit 2022 can be used by the analysis unit 205.

The translation result document acquisition unit 2023 acquires a translation result document that is the translation result of the invention document acquired by the invention document acquisition unit 2021, for example, in accordance with the user's designation of the translation result document that is the translation result of the invention document. The translation result document acquired by the translation result document acquisition unit 2023 can be used by the machine learning unit 207.

Each of the invention documents, drawing documents, and translation result documents is provided by database data, data as environment variables of the OS, data stored in a memory area used as a clipboard, a web server on a network such as the Internet, etc. For example, it may exist as the contents of a file.

The invention document acquisition unit 2021, the drawing document acquisition unit 2022, and the translation result document acquisition unit 2023 each generate a file containing an invention document based on the designation of each file path by the user of the invention document analysis system 20, for example. Invention documents, drawing documents, and translation result documents can be obtained by accessing invention document files, drawing document files that contain drawing documents, and translation result document files that contain translation result documents. . Invention document files, drawing document files, and translation result document files comply with predetermined file naming rules (for example, rules that specify that the first character string of the file name, extension, etc. are different for each file type). In this case, the drawing document file or the translation result document file may be located in the same folder as the invention document file, thereby saving the user the trouble of specifying the file path. may be reduced. Note that the predetermined file name assignment rule may be defined in the configuration information acquired by the configuration information acquisition unit 201, for example. For example, the drawing document acquisition unit 2022 retrieves the invention document acquired by the invention document acquisition unit 2021 and a file that matches a predetermined file naming rule (for example, an invention document and a file other than the extension) in the same or related folder. A drawing document may be obtained from a file with the same name and a pdf extension, or a corresponding drawing document may be obtained from a website on the Internet based on the publication number of the invention document, etc. Good too. Further, for example, the translation result document acquisition unit 2023 acquires a translation result document from a file that matches a predetermined file naming rule in the same or related folder as the invention document acquired by the invention document acquisition unit 2021. Alternatively, the corresponding translation result document may be obtained from a website on the Internet based on the publication number of the invention document.

The drawing document file is, for example, a PDF file. The invention document file and the translation result document file are, for example, text files. The invention document file and the translation result document file may be, for example, a PDF file or a file in a format mainly used for various application programs having document creation and editing functions such as Microsoft (registered trademark) Word, or an XML file. (Extensible Markup Language) file. Further, the invention document file and the translation result document file may be composed of one or more files in the same format as a display document (for example, an HTML document) output by the invention document analysis system 20, for example. The invention document acquisition unit 2021 receives, for example, claim information and specification information in patent application related information as display documents output by the invention document analysis system 20 in the past, or revised claim information and revised specification information, In accordance with the user's instructions, the invention document may be obtained as an invention document including claims and specifications. Note that the PDF file may be a searchable PDF file that includes character strings in the content in a searchable state. Further, the drawing document acquisition unit 2022 converts a drawing document file that is not a searchable PDF file into a searchable PDF file by character recognition technology using machine learning (e.g., deep learning) or character recognition using OCR technology. You may also use it by converting it to .

The invention document acquisition unit 2021, the drawing document acquisition unit 2022, and the translation result document acquisition unit 2023 edit (for example, format conversion, partial deletion, etc.) the contents of the invention document file, the drawing document file, and the translation result document file, respectively. In the above, an invention document, a drawing document, and a translation result document may be acquired. As a specific example, the invention document acquisition unit 2021 reads the invention document, which is the content of the invention document file, in a method according to the encoding format of the invention document file (for example, UTF-8, EUC, Shift_JIS, CP932, etc.). By performing format conversion as necessary, for example, obtain an invention document as character code string data in a predetermined encoding format (for example, UTF-8) that can represent a relatively large number of characters. do. The encoding format of the invention document file may be specified, for example, depending on whether an error occurs when reading the invention document file through trial and error, or the encoding format may be determined by reading the invention document file by trial and error, or by converting the invention document file into a binary byte format. ) column and inspect the contents to identify the encoding format.

Note that the invention document analysis system 20 may be configured to have a document creation and editing function (for example, it may be configured to include an application program related to the document creation and editing function, such as a text editor). In this case, the invention document acquisition unit 2021 acquires the invention document, which is data generated or edited by the process executed in response to the user's input operation, in the document creation and editing function of the invention document analysis system 20. You can also do it. Similarly, in the document creation and editing function of the invention document analysis system 20, the translation result document acquisition unit 2023 acquires a translation result document that is data generated or edited by a process executed in response to a user's input operation. You can also do it. Further, the invention document analysis system 20 may be configured to have a drawing creation and editing function. In this case, the drawing document acquisition unit 2022 acquires the drawing document, which is data generated or edited by the process executed in response to the user's input operation, in the drawing creation and editing function of the invention document analysis system 20. You can also do it.

When invention documents, drawing documents, and translation result documents exist as data provided by a server device such as a web server on the Internet, for example, an invention document acquisition unit 2021, a drawing document acquisition unit 2022, and a translation result document acquisition unit 2023 accesses the server device via the WAN based on the URL related to the server device, information specifying the invention document, etc. (for example, the publication number related to the gazette issued regarding the invention document) specified by the user. By downloading each data, an invention document, a drawing document, and a translation result document can be respectively obtained. In addition, invention documents, drawing documents, and translation result documents may be files or data that can be output by a device equipped with a file server, database, or scanner function connected via LAN to a device as one component of the invention document analysis system 20. For example, the invention document acquisition unit 2021, the drawing document acquisition unit 2022, and the translation result document acquisition unit 2023 each receive the content or data of each file via the LAN. Documents, drawing documents, and translation result documents can be obtained. Therefore, the invention document acquisition unit 2021 reads a handwritten memo of a technical idea, for example, in a case where the memo has substantial content that can be converted into the content of a claim or specification of an invention document. An invention document can be obtained by receiving data output from a device equipped with a scanner function.

For example, when the invention document acquisition unit 2021 acquires an invention document that is the content of an international application in Japanese for which international publication publications etc. are provided on the Internet, the translation result document acquisition unit 2023 is configured to make the translation result document available on the Internet. Using the information indicating the related applications related to the international application, the content of the application written in a foreign language to the government office of the specified country, which is presumed to correspond to the translated content corresponding to this international application, is Translation result documents may be obtained by accessing published patent publications and the like.

[3-1-3. Important word/phrase setting dictionary acquisition part]
The important word/phrase setting dictionary acquisition unit 203 is a partial modification of the important word/phrase setting dictionary acquisition unit 103 shown in the first embodiment. The important word/phrase setting dictionary acquisition unit 203 and the important word/phrase setting dictionary acquisition unit 103 may be the same.

The important word/phrase setting dictionary acquisition section 203 can obtain the important word/phrase setting dictionary based on the information for specifying the important word/phrase setting dictionary in the setting information acquired by the setting information obtaining section 201 . The information for specifying the important word/phrase setting dictionary in the setting information may be information that identifies and prioritizes a plurality of important word/phrase setting dictionaries. In this case, the important word/phrase setting dictionary acquisition unit 203 By acquiring the word/phrase setting dictionaries and integrating them in a memory or the like according to their priority order, they can be used as one important word/phrase setting dictionary.

The contents of the important word/phrase setting dictionary can be generated and edited by the user.

The important word/phrase setting dictionary may exist as any data such as data in a database, data provided by a web server on a network such as the Internet, and may exist as the contents of a file, for example.

The important word/phrase setting dictionary may be included as part of the invention document file. For example, a predetermined heading such as <Important word/phrase setting dictionary> or <Important word/phrase list> may be added to the invention document file, and data corresponding to the important word/phrase setting dictionary may be written after the heading. In this case, the important word/phrase setting dictionary acquisition unit 203 can obtain the important word/phrase setting dictionary based on the information for specifying the important word/phrase setting dictionary in the setting information, which indicates the predetermined heading.

The important word/phrase setting dictionary acquired by the important word/phrase setting dictionary acquisition unit 203 can be used by the important word/phrase extraction unit 2051 in the analysis unit 205 to extract important words. The important word/phrase setting dictionary is information that lists a plurality of important words to be extracted. In the important words and phrases setting dictionary, important words may be classified and listed according to some group, such as the department to which the inventor who created the invention in the invention document belongs, or the technical field to which the invention belongs. Among the set of important words for each of the plurality of groups in the important word/phrase setting dictionary, the set of important words for one group based on the setting information, selection by the user, etc. is extracted by the important word/phrase extraction unit 2051. It may also be used.

The important word/phrase setting dictionary acquisition unit 203 refers to the identification information of the important word/phrase setting dictionary in the setting information and identifies a predetermined folder in the file system, thereby acquiring the important word/phrase settings stored as a file in the predetermined folder. You may also obtain a dictionary. For example, if a predetermined folder stores a plurality of important word setting dictionary files that are classified into some groups such as technical fields and can be identified by file names, the important word setting dictionary acquisition unit 203 out of the plurality of files, acquires an important term setting dictionary from an important term setting dictionary file corresponding to a group (for example, a group specified by the user) that is related to the invention document acquired by the invention document acquisition unit 2021. Good too. In addition, the important word/phrase setting dictionary acquisition unit 203 uses technology that has a higher correlation with the important word group extracted from the invention document by the important word/phrase extraction unit 2051 based on condition information etc. without using the important word/phrase setting dictionary. The important word/phrase setting dictionary may be obtained from a file of the important word/phrase setting dictionary corresponding to a group such as a field.

In the important word/phrase setting dictionary, a corresponding translated word may be added to the important word/phrase. Note that for each of all important words and phrases in the important word/phrase setting dictionary, it can be said that information in which the corresponding translated word of the important word is appended is translation auxiliary information. Further, in the important word/phrase setting dictionary, information including a set of important words/phrases and corresponding translated words for important words/phrases to which corresponding translation words are associated can be used as translation auxiliary information. For important words and phrases that are associated with corresponding translation words in the important word and phrase setting dictionary, for example, the translation unit 2054 of the analysis unit 205 converts the important words and phrases included in the claim information based on the invention document or the specification information, etc. Corresponding translation words are determined, and claim translation information, specification translation information, etc. can be generated based on the determination result. When a user (for example, an inventor or a patent attorney) defines an important word/phrase setting dictionary to associate a number of specific important words with corresponding translation words, the invention document analysis system 20 can create an important word/phrase setting dictionary. Outputs information that reflects the corresponding translated words corresponding to the important words based on the . For example, a translator who receives the output information can clearly recognize the designation of translated words for important words and phrases, and may be able to translate the invention document appropriately or efficiently. An example of the use of the important word/phrase setting dictionary is, for example, if a claim of an invention document created by a user includes important words such as "information display timing control section," "information transmission timing control section," or "equipment control section." In such cases, the user may use the key words "information display timing control section," "information transmission timing control section," and "device control section" in the key word setting dictionary in order to simplify the terminology of the claim in the target language. For example, "first controller," "second controller," and "third controller" are associated with "first controller," "second controller," and "third controller" as corresponding translated words. Note that the user can efficiently generate the important word setting dictionary by, for example, editing the important word list in the patent application related information that is output as a result of having the invention document analysis system 20 analyze the invention document. , updates, etc. may be possible.

In the important word/phrase setting dictionary, a sentence explaining the important word/phrase using the important word/phrase as the subject may be added to the important word/phrase.

For example, as in the invention document analysis system 10 shown in FIG. 9, when there are a plurality of terminal devices 10c to 10e connected to a server device 10b having a file server function via LAN, each of the terminal devices performs the invention document analysis. It may function as system 20. In this case, each terminal device functioning as the invention document analysis system 20 records an important word/phrase setting dictionary in a recording medium (for example, memory, hard disk, etc.) included in the terminal device, and stores the important word/phrase setting dictionary at any time on the server device 10b. You may also send it to The important word/phrase setting dictionary in each terminal device may be created and edited by the user of each terminal device. Then, the server device 10b distributes a new important word setting dictionary generated as a result of integrating the important word setting dictionaries received from each terminal device to each terminal device, and each terminal device receives the delivered important word setting dictionary. An existing important word/phrase setting dictionary may be updated based on the dictionary.

Note that when the output based on the analysis results of invention documents by the invention document analysis system 20 (for example, output of patent application related information) is realized as a display on a display device included in a terminal device, equipment, etc. (for example, a computer), , the invention document analysis system 20 may add a word or phrase to the important word/phrase setting dictionary in response to a user operation on an operation-compatible GUI element regarding a part of the word or phrase in the displayed content. When a word is added to the important word/phrase setting dictionary in this way, when the invention document analysis system 20 subsequently performs invention document analysis processing, the important word/phrase extraction unit 2051 changes the important word/phrase settings after the addition of the word. Extract important words based on the dictionary. In addition, the invention document analysis system 20 adds a word to the important word/phrase setting dictionary in response to a user operation for displaying information on the display device based on the analysis result of the invention document, and then adds the word/phrase to the important word/phrase setting dictionary after the addition. Based on the previous analysis, GUI elements for specifying reanalysis, that is, GUI elements for analyzing the invention document again (for example, extracting important words, etc.) (for example, a button labeled "Reanalysis") It may be displayed when displaying information based on the results on a display device. The invention document analysis system 20 may display the GUI element for specifying reanalysis on the display screen, for example, as appropriate, or may display it at all times.

The important word/phrase setting dictionary acquisition unit 203 may also acquire a non-important word/phrase setting dictionary, which is information listing a plurality of words that should not be extracted as important words (in other words, information including a plurality of non-important words). good. In this case, the information for specifying the non-important word/phrase setting dictionary can be defined in the setting information acquired by the setting information obtaining unit 201, and the important word/phrase setting dictionary obtaining unit 203 can specify the non-important word/phrase setting dictionary in the setting information. A non-important word/phrase setting dictionary can be obtained based on the identification information.

[3-1-4. Invention document standard provision acquisition department for translation]
The invention document standard provision acquisition unit 204 for translation is a partial modification of the invention document standard provision acquisition unit 104 shown in the first embodiment. The invention document standard provision acquisition unit for translation 204 and the invention document standard provision acquisition unit for translation 104 may be the same.

The standard provision acquisition unit for invention documents for translation 204 determines the criteria that the invention document to be translated should satisfy based on the content of the setting information (for example, information for specifying the standard provision for invention documents for translation) acquired by the setting information acquisition unit 201. Obtain standard provisions for invention documents for translation that specify multiple conditions. Each condition in the standard regulations for invention documents for translation is determined, for example, so that the invention document is a Japanese manuscript for translation that is sufficiently suitable as a subject for translation. The invention document standard provisions for translation acquired by the invention document standard provisions for translation acquisition unit 204 in the invention document analysis system 20 are used, for example, in the warning information generation unit 2059 of the analysis unit 205 to generate warning information based on the invention document. .

The standard provisions for invention documents for translation are assumed to be provisions established by an organization for the purpose of standardizing invention documents for translation, but they may be established by any other party. Furthermore, this organization related to standardization may be a division of a company or company, or may be composed of multiple companies or companies, regardless of its size.

For example, if the information for specifying the standard provisions for invention documents for translation in the setting information is file path information, the standard provisions for invention documents for translation acquisition unit 204 accesses the file according to the file path. Read and obtain the invention document standard provisions for translation from the file. For example, if the information for specifying the standard provisions of invention documents for translation in the setting information is URL information, the standard provisions of invention documents for translation acquisition unit 204 acquires a server device or the like on the network according to the URL. Access the Invention Document Standard Regulations for Translation as a provided resource and download and obtain the Invention Document Standard Regulations for Translation.

The invention document standard provisions for translation may include version information such as a version number indicating the version of the invention document standard provisions for translation, for example. The standard provisions for invention documents for translation may be updated under a predetermined version control system whenever the contents of the provisions are changed. The invention document standard provisions for translation acquisition section 204 acquires information for specifying the standard provisions of invention documents for translation in the setting information acquired by the setting information acquisition section 201 (for example, information for specifying the standard provisions of invention documents for translation via a network). If it is possible to access the standard provisions of invention documents for translation based on the URL) and obtain a version of the standard provisions of invention documents for translation that is newer than the version (for example, version number) of the standard provisions of invention documents for translation that has already been obtained. , for example, delete the already acquired standard provisions for invention documents for translation, and obtain a new version of the standard provisions for invention documents for translation.

The standard provisions for invention documents for translation may, for example, include provisions in association with the target language (that is, include provisions for each target language), or may include provisions for only a specific language (for example, English). It is also possible to have specific regulations. If the invention document standard regulations for translation include regulations for each translation destination language, the translation destination language may be specified, for example, by setting information or by a user operation on the invention document analysis system 20.

The invention document standards for translation may include, for example, data or an inspection program for processing to determine whether the invention document satisfies the conditions. This inspection program is, for example, a program responsible for determining whether or not an invention document conforms to each of a plurality of items (i.e., condition items indicating conditions) of the standard provisions for invention documents for translation. , for example, may be a plug-in module. For example, the warning information generation unit 2059 in the analysis unit 205 generates information that is acquired by the invention document acquisition unit 2021 by executing an inspection program included in the invention document standard regulations for translation (that is, by causing a microprocessor to execute the inspection program). It is also possible to determine whether or not the invented invention document conforms to each of a plurality of conditions defined in the standard provisions for invention documents for translation, and to generate warning information by using the results of the determination.

The standard provisions for invention documents for translation include, for example, character string patterns for parts of invention documents that do not meet the conditions and may be inappropriate, as information that defines conditions that invention documents to be translated should satisfy. Contains inappropriate character string pattern information indicating an inappropriate character string pattern representing . The standard provisions for invention documents for translation, for example, can be used to identify inappropriate character string patterns in the invention document in association with all or part of inappropriate character string pattern information that indicates inappropriate character string patterns related to description parts that may be inappropriate. Contains correction information on how to correct the written section where appropriate.

The invention document standard provisions for translation classify condition items, which are items of conditions that an invention document must satisfy, based on, for example, the handling of character strings that may be included in an invention document or differences in characteristics, and for each condition item, For example, it includes condition item names and pattern information (for example, inappropriate character string pattern information) for extracting inappropriate descriptions in invention documents. An inappropriate statement in an invention document may be a potentially inappropriate statement, and if a patent right is obtained by submitting a translation result document based on the invention document to a predetermined region (for example, the United States). It may be a statement that may cause some kind of disadvantage. Pattern information for extracting inappropriate descriptions may be other than patterns for specifying character strings. For example, if a claim has a clause that cites multiple clauses, the cited clause It is also possible to include a program related to an algorithm for extracting descriptions in which a person cites multiple terms. All or some of the condition items of the invention document standard provisions for translation may further include modification information. The modification information may include a program related to an algorithm for modifying the invention document, such as deleting a character string in the invention document that corresponds to the inappropriate character string pattern indicated by the inappropriate character string pattern information. . In addition, corrections are made to correspond to pattern information for extracting as inappropriate statements where a claim has a clause that cites multiple clauses and the cited clause further cites multiple clauses. The information may include, for example, a program for modifying a term that cites a plurality of terms so that it cites a singular term.

In addition, all or some of the condition items of the invention document standard provisions for translation further include an explanation to the user of the inappropriate reason for the inappropriate description in the invention document that corresponds to the pattern indicated by the pattern information; It may also include a condition item explanation message that explains to the user the disadvantage caused by the description or explains to the user how to appropriately correct it. Note that the explanation of the inappropriate reason for the inappropriate description in the condition item explanation message may include, for example, information such as provisions of laws and regulations, judicial precedents, etc. that serve as the basis for the inappropriate reason.

FIG. 14 shows an example in which a condition item name and a pattern indicated by pattern information are defined for each condition item, and modification information is added to some of the condition items, as an example of the standard definition of invention documents for translation. The example shown in the same figure is just a simple example, and condition items may be added, or further conditions may be added in response to more detailed classification of condition items, patterns indicated by more detailed classification of pattern information, etc. Appropriate correction information may be added.

Condition item names such as “lack of subject in each sentence of the specification” shown in Figure 14 are used when the pattern indicated by the pattern information for extracting inappropriate descriptions that correspond to the condition exists in the invention document. , for example, can be expressed in the warning information generated by the warning information generation unit 2059 and included in the patent application related information. Note that "each sentence in the specification" is separated, for example, by a line feed code immediately after a period character or at the end of a line. Note that a plurality of pattern information (for example, character string pattern information) may be associated with one condition item. For example, regarding the condition item name "Character string that may be desirable to change to another expression in U.S. applications" shown in FIG. 14, the description of the invention document states that "(1) in , etc., and (2) in Figure 2 is a diagram representing ....'' It is possible that there are multiple separate figures in one drawing. Further, character string pattern information indicating a character string pattern to which a character string indicating gender (for example, "(1) in FIG. 2) corresponds" may be associated. For example, this character string pattern may include a number following the character "Figure", followed by a predetermined character such as "no", "no", etc. It is determined that descriptions that include alphanumeric characters between parentheses and closing parentheses are applicable.

As shown in Figure 14, the condition item name “Among important words and phrases in the claim, the importance level is not stated in the specification.” As a specific example, the pattern information of "pattern for detecting the absence of important words/phrases" is the importance level in the setting information among the important words extracted by the important word/phrase extraction unit 2051 from the claims of the invention document. This is information indicating a set of character string patterns for detecting important words and phrases from the specification of an invention document for each important word and phrase that is not determined to be of low importance based on the standard information.

Another example of a condition item in the standard provisions for invention documents for translation is, for example, the condition item name "Multiple parentheses" and the inappropriate description of multiple parentheses containing parentheses. Examples of condition items include pattern information indicating a pattern for extracting descriptions and a condition item explanation message such as that it is desirable to reduce parentheses. By using this condition item, if multiple parentheses are included in an invention document, the warning information generated by the warning information generation unit 2059 and included in the patent application related information will be clearly described. Points regarding the use of multiple parentheses may be included to call attention to the following.

Another example of a condition item in the standard provisions for invention documents for translation is, for example, the condition item name "description of functional means in claims" and the condition item name "description of functional means in claims" and a condition item for extracting the description of functional means as an inappropriate description. Examples of condition items include pattern information indicating a pattern. By using this condition item, if a claim of an invention document includes a description of a functional means, the description of the functional means will be included in the warning information generated by the warning information generation unit 2059 and included in the patent application related information. Such indications may be included. The pattern indicated by this pattern information is, for example, a pattern that corresponds to descriptions of so-called functional blocks such as "... means", "... section", etc. The description of functional means in a claim is not necessarily inappropriate, and there may be cases where functional means are strategically described, but if a claim contains a description of functional means, it may be disadvantageous in the US application. Since there is a possibility that this may occur, this indication in the warning information may be useful for calling attention.

The invention document standard provisions for translation are used, for example, by the warning information generation unit 2059 of the analysis unit 205. In addition, the modification information as an example of the content of the invention document standard provisions for translation is used, for example, to generate modified claim information and modified specification information that may be included in the patent application related information output by the output unit 208. It is used in the analysis section 205.

[3-1-5. Analysis Department]
The analysis unit 205 is a partially modified version of the analysis unit 105 shown in the first embodiment, and is the same as the analysis unit 105 except for points not particularly shown here.

The analysis unit 205 analyzes the invention document acquired by the invention document acquisition unit 2021 based on the setting information acquired by the setting information acquisition unit 201.

The analysis unit 205 includes an important phrase extraction unit 2051, a classification unit 2053, a translation unit 2054, an importance determination unit 2055, an attribute identification unit 2056, a claim citation relationship detection unit 2057, an inter-element relationship determination unit 2058, a warning information generation unit 2059, It includes a claim information generation unit 205a, a specification information generation unit 205b, and a claim tree generation unit 205c.

The analysis result of the invention document by the analysis unit 205 is used, for example, in the output unit 208 to generate patent application related information to be output, and is also used in the machine learning unit 207 to evaluate the invention document. .

FIG. 15 shows an example of the content of patent application related information generated by the analysis unit 205 and output by the output unit 208. FIG. 16 shows a configuration example of an HTML document, which is an example of a display document as one expression format of patent application related information. FIG. 17 shows an example of a display screen displayed by a browser or the like based on a display document (for example, an HTML document) expressing patent application related information. The display screen illustrated in FIG. 17 includes a menu column displaying a menu for selecting various information items in patent application related information, a content display column for patent application related information items according to the menu selection, etc. There is. In the display screen illustrated in FIG. 17, for example, the above-mentioned GUI element for specifying reanalysis may be displayed in the operation column for specifying reanalysis of an invention document.

[3-1-5-1. Important word extraction section]
The important word/phrase extraction unit 2051 is a partial modification of the important word/phrase extraction unit 1051 shown in the first embodiment. The important word/phrase extraction unit 2051 may be similar to the important word/phrase extraction unit 1051. Extraction of important words from an invention document by the important word/phrase extraction unit 2051 is useful if it is accurate to some extent, and does not necessarily need to be completely accurate.

The important phrase extraction unit 2051 may include the same functional configuration as the first extraction unit 10511, the second extraction unit 10512, the third extraction unit 10513, and the fourth extraction unit 10514, for example. For example, the important phrase extraction unit 2051 extracts references indicated by the reference character string pattern information in the condition information in the setting information acquired by the setting information acquisition unit 201 in the claims and specifications of the invention document acquired by the document acquisition unit 202. If a character string that corresponds to the character string pattern is followed by a character string that corresponds to the important word candidate character string pattern indicated by the important word candidate character string pattern information in the condition information, the key word candidate character string pattern The first extraction unit may be provided with a function of extracting a character string that is the same as the relevant character string as an important phrase. For example, the important word/phrase extraction unit 2051 extracts the important word/phrase indicated by the key word/phrase candidate character string pattern information in the condition information in the setting information acquired by the setting information acquisition unit 201 into the claim of the invention document acquired by the document acquisition unit 202. a second extraction unit having a function of extracting the same character string as the character string corresponding to the important word/phrase candidate character string pattern as an important word/phrase when there is a period character following the character string corresponding to the candidate character string pattern; may be provided. For example, the important word/phrase extraction unit 2051 extracts important word/phrase candidate characters indicated by the key word/phrase candidate character string pattern information in the condition information in the setting information acquired by the setting information acquisition unit 201 into the invention document acquired by the document acquisition unit 202. If a character string that corresponds to the column pattern is followed by a character string that corresponds to the code string pattern indicated by the code string pattern information in the condition information, the character string that corresponds to the key word candidate string pattern and A third extraction unit may be provided that has a function of extracting the same character string as an important phrase. Further, the important word/phrase extraction unit 2051 extracts, for example, an important word/phrase indicated by the important word/phrase candidate character string pattern information in the condition information in the setting information acquired by the setting information acquiring unit 201 in the invention document acquired by the document acquiring unit 202. If a character string that corresponds to the candidate character string pattern is followed by a character string that corresponds to the specific particle string pattern indicated by the specific particle string pattern information in the condition information, it corresponds to the important word candidate character string pattern. The fourth extraction unit may be provided with a function of extracting the same character string as the important word/phrase.

The important term extraction unit 2051 executes important term extraction processing to extract important terms from the invention document acquired by the invention document acquisition unit 2021 based on the condition information included in the setting information acquired by the setting information acquisition unit 201. do. By appropriately determining the condition information, the important word extraction unit 2051 can, for example, extract words from individual patent publications among several patent publications published by the Japan Patent Office, as an example of an invention document. A plurality of important words can be extracted from the claims or specifications.

The important words and phrases extracted by the important word and phrase extraction unit 2051 are used, for example, to generate information such as patent application related information output by the invention document analysis system 20. Here, using an example in which an invention document includes claims and a specification, the important word extraction unit 2051 that extracts important words from each of the claims and the specification of the invention document will be described in detail. The important word extraction unit 2051 executes a claim important word extraction process that extracts important words from the claims of an invention document and a specification important word extraction process that extracts important words from the specification of an invention document. do.

In the analysis unit 205, for example, the important phrase extraction unit 2051 manages the extracted important phrases by distinguishing from which of the claims and the specification the important phrases are extracted. The analysis unit 205 can hold important phrase management information for managing important phrases. The important term management information is information in which each important term extracted from the invention document is associated with identification information, importance, attribute information, corresponding translated words, a set of translated word candidates, and the like.

The important word extraction unit 2051 extracts important words in the invention document acquired by the invention document acquisition unit 2021 in a range according to the important word extraction range information in the setting information. Here, in the setting information, the range of all claims and the range in which "Detailed Description of the Invention" and one or more "Examples" are described in the specification are extracted from important words. The following explanation assumes an example where the range is set. The scope of each claim can be determined, for example, by the heading of one claim, "Claim n" (n is a natural number), which is sandwiched between opening and closing corner brackets in the line order of the invention document. from the line to the first of either the heading represented by the next string of characters enclosed in corner brackets (for example, the string starting with "Claims", "Document name", etc.) or the end of the invention document It is. Furthermore, the scope of the description of "Details for Carrying Out the Invention" in the specification can be, for example, from the heading line "Details for Carrying Out the Invention" enclosed in corner brackets to the following: This is the first of the headings represented by character strings enclosed in corner brackets (for example, "Example", "Explanation of symbols", etc.) or the end of the invention document. In addition, the range in which one "Example" is described in the specification is, for example, from the heading line "Example" or "Example m" (m is a natural number) enclosed in corner brackets. , up to either the heading represented by the next string of characters enclosed in corner brackets or the end of the invention document, whichever occurs first.

The important word/phrase extraction unit 2051 may extract important words/phrases from the invention document using, for example, a plug-in module that performs important word/phrase extraction processing.

[3-1-5-1-1. Extraction of important words in complaints]
The important word extraction unit 2051 performs claim important word extraction processing, for example, by searching for important word candidates for each claim line of an invention document, and after extracting the important word candidates, based on predetermined criteria (for example, extraction exception conditions). ), and the words determined to be important words are extracted as important words. Note that the important word/phrase extraction unit 2051 searches for important word/phrase candidates by excluding comment lines or comment portions from the search target range, for example, based on the comment specifying information of the setting information.

For example, the important word extraction unit 2051 converts the character format of each line of the claims in the invention document for the search before searching for important word candidates for each line of the claims. may be used for exploration. Examples of character format conversion include conversion of half-width characters to full-width characters, conversion of lowercase letters (for example, "a") to uppercase letters (for example, "A"), and the like. By performing such character format conversion, it is possible to simplify the character string patterns required to search for important word candidates (for example, important word candidate character string patterns) and speed up the processing related to the extraction of important words. etc. may be realized.

In one line of a claim, multiple key word candidates that are determined to be potentially important words are extracted, and if the description ranges of the character strings of the multiple key word candidates partially overlap, the key word extraction unit Step 2051 selects one of the plurality of important word candidates based on predetermined selection criteria and extracts it as an important word. For example, this selection criterion is applicable if one of the strings of multiple key word candidates is included in the string of other key word candidates. This is a standard in which a character string that is a candidate for one important word is excluded, and an important word or phrase is selected from the remaining candidate character strings. For example, the important word/phrase extraction unit 2051 may use one or a small number of relatively complex important word/phrase candidate character string patterns to extract the important word/phrase, or may use a relatively concise important word/phrase candidate character string pattern. It is also possible to use more than one. For example, in order to extract important words, if a plurality of character strings match the important word candidate character string pattern, the important word/phrase extraction unit 2051 extracts important words. When a character string is included in the description of the invention document, the remaining character strings after excluding one of the plurality of character strings are selected as selection targets, and important words are selected and extracted.

For example, if the same important words and phrases included in the important words and phrases setting dictionary acquired by the important words and phrases setting dictionary acquisition unit 203 are included in the claims of the invention document, the important words and phrases extraction unit 2051 extracts the words from the claims. The word/phrase is extracted as a candidate for an important word/phrase. For example, the important word/phrase extraction unit 2051 does not extract words included in the non-important word/phrase setting dictionary acquired by the important word/phrase setting dictionary acquisition unit 203 as important word/phrase candidates.

For example, the important word/phrase extraction unit 2051 refers to each character string pattern information indicated by the condition information of the setting information and extracts words in the claims of the invention document as important word/phrase candidates.

As a specific example, the important word/phrase extraction unit 2051 extracts an important word/phrase candidate character string in the condition information following a character string corresponding to the reference character string pattern indicated by the reference character string pattern information in the condition information in the complaint. When a character string corresponding to the important word/phrase candidate character string pattern indicated by the pattern information exists, the same character string as the character string corresponding to the important word/phrase candidate character string pattern is extracted as an important word/phrase candidate. If the same character string as the character string corresponding to the important word candidate character string pattern is to be extracted as an important word from among the important word candidates, the important word/phrase extraction unit 2051, for example, When there is a character string in the claim that includes the character string extracted as the important word at the end of the word and that corresponds to the important word candidate string pattern indicated by the important word candidate string pattern information, the important word is Character strings that correspond to the candidate character string pattern may be further extracted as candidates for important words. As a result, for example, when the character string corresponding to the reference character string pattern is "said", when "control device" is extracted as an important phrase from the description of "said control device" in the claim, Furthermore, "equipment control device" in the claim, which has "control device" at the end of the word, can be further extracted as a candidate for an important phrase. Note that the important phrase extraction unit 2051 detects that a predetermined prefix character string exists following the character string corresponding to the reference character string pattern indicated by the reference character string pattern information in the condition information, and that the predetermined prefix character string is present in the complaint. If the character string is followed by a character string that corresponds to the important word candidate character string pattern indicated by the important word candidate character string pattern information in the condition information, then the character string that corresponds to the important word candidate character string pattern is the same as that character string. The character string may be extracted as a candidate for an important phrase. Here, the predetermined prefix character string is an ordinal number character string that includes one or more ordinal numbers such as "first", "first", "first, third to fifth", etc. , or a quantity character string expressing a number such as "plurality", "one", "a set", "one or more", or "predetermined", "predetermined", "one", etc. These are limited character strings such as "No", "Other", etc. If the ordinal number string contains multiple ordinal numbers, the ordinal number string must be separated by "~", "-", ",", comma character ",", or the part of speech is a conjunction. Contains character strings, etc. Conjunctions may be identified using conventional morphological analysis techniques, or by comparing strings of conjunctions (such as "and", "or", "and", etc.) against a list of conjunctions. Good too. In the conjunction list, for example, descriptions such as "and/or" and "and/or" may be included.

For example, the important word/phrase extraction unit 2051 extracts specific particle character string pattern information in the condition information following a character string corresponding to the important word/phrase candidate character string pattern indicated by the important word/phrase candidate character string pattern information in the condition information in a complaint. When a character string corresponding to the specific particle character string pattern indicated by is present, the same character string as the character string corresponding to the important word/phrase candidate character string pattern is extracted as an important word/phrase candidate. Note that, for example, the important word/phrase extraction unit 2051 extracts a predetermined character string (for example, "each", "each", etc.) exists, and following that predetermined character string, there is a character string that corresponds to the specific particle string pattern indicated by the specific particle string pattern information in the condition information. , the same character string as the character string corresponding to the important word/phrase candidate character string pattern may be extracted as an important word/phrase candidate.

For example, the important word/phrase extraction unit 2051 extracts the code string pattern information in the condition information following the character string corresponding to the important word/phrase candidate character string pattern indicated by the important word/phrase candidate character string pattern information in the condition information in the complaint. If there is a character string that corresponds to the indicated code string pattern (for example, a code between opening parentheses and closing parentheses), the same character string as the important word candidate character string pattern exists. are extracted as important word candidates.

The important word/phrase extraction unit 2051 detects, in a claim, a character other than a kanji or a predetermined kanji (for example, "each") that corresponds to the key word/phrase candidate character string pattern indicated by the key word/phrase candidate character string pattern information in the condition information. If there is a character string that corresponds to the important word/phrase candidate character string pattern, the same character string as the character string corresponding to the important word/phrase candidate character string pattern may be extracted as a candidate for the important word/phrase.

For example, the important word/phrase extraction unit 2051 extracts the important word/phrase candidate character string pattern information in the condition information at the beginning of a line, after a punctuation mark, or after a verb, auxiliary verb, adjective, conjunction, adverb, or predetermined hiragana in a claim. If there is a character string that corresponds to the word candidate character string pattern, the same character string as the character string that corresponds to the important word candidate character string pattern may be extracted as an important word candidate. Note that each line in an invention document can be composed of multiple character strings, and the part of speech of each character string (for example, the type of character string classified into verbs, adjectives, conjunctions, etc.) can be determined using existing morphological analysis technology. It can be determined to some extent using . Therefore, part of the condition information (for example, important word/phrase candidate character string pattern information) may indicate a character string pattern that distinguishes parts of speech. Character string patterns with different parts of speech can be created by, for example, performing morphological analysis of each word in an invention document using morphological analysis technology, and then classifying the words that are each of the predetermined parts of speech selected as a result of the morphological analysis using morphological analysis technology. When extracting candidates for important words or phrases on the premise that a specific symbol string is added to at least one of immediately before or after the word, or that the word is replaced with a specific symbol string for each part of speech, This is a character string pattern expressed by including specific symbol strings for each part of speech. An example of a program for performing morphological analysis is the open source morphological analysis engine "Mecab" by Taku Kudo, developed by the Kyoto University Graduate School of Informatics and Nippon Telegraph and Telephone Corporation's Communication Science Research Institute joint research unit project. can be mentioned. For example, it may be determined that a noun corresponds to one important word/phrase candidate character string pattern indicated by the important word/phrase candidate character string pattern information in the condition information. In addition, when determining whether or not a character string corresponding to the specific particle string pattern indicated by the specific particle string pattern information exists in an invention document (for example, in a claim), the character string is determined by morphological analysis. A determination as to whether a character string whose part of speech is a particle exists or not may be added.

For example, when there is a character string extracted as an important word in the specification of an invention document based on condition information, the important word extraction unit 2051 detects whether the same character string exists in the claim of the invention document. When doing so, the character string may be extracted as a candidate for an important word or phrase, either unconditionally or under certain conditions. An example of this certain condition is that the importance of the important words extracted from the specification of the invention document is equal to or higher than a predetermined level. As a specific example, the important word/phrase extraction unit 2051 may, for example, in the specification of the invention document, follow the character string corresponding to the important word/phrase candidate character string pattern indicated by the important word/phrase candidate character string pattern information in the condition information, If there is a character string that corresponds to the code string pattern indicated by the code string pattern information in the condition information, the same character string as the key word candidate character string pattern is included in the claim of the invention document. If such a character string exists, the character string may be extracted as a candidate for an important word or phrase.

For example, when a patent document is indicated as prior art document information in the specification of an invention document, the important phrase extraction unit 2051 extracts important phrases from the claims and specifications of the patent document based on condition information. , if there is a word or phrase in the claim of the invention document that is the same as an important word or phrase in the patent document, it may be extracted as a candidate for the important word or phrase.

For example, when an "explanation of symbols" is described in the specification of an invention document, the important phrase extraction unit 2051 extracts words that are associated with the symbols in the "explanation of symbols" that are included in the invention document. If the word/phrase is present in a claim, the word/phrase may be extracted as a candidate for an important word/phrase. For example, among the words in a claim extracted as a candidate for an important word, if that word in the claim is followed by a predetermined non-successive character string, the important word/phrase extraction unit 2051 selects that word as an important word. It is also possible not to extract it. The predetermined non-successive character string is, for example, "Suru" or "Shi". For example, if a predetermined unnecessary word is included at the beginning of a word in a claim that has been extracted as a candidate for an important word, the important word/phrase extraction unit 2051 extracts a word from which the predetermined unnecessary word has been removed. , it may be regarded as extracted as a candidate for an important word or phrase. Examples of the predetermined unnecessary words include "so-called", "substance", and "various".

For example, if a character string in a claim extracted as an important word candidate is a character string that corresponds to an extraction exception condition, the important word extraction unit 2051 may decide not to extract that character string as an important word. good. The extraction exception condition may be defined as part of the condition information, for example. Extraction exception conditions include, for example, character strings that are not important words on their own (e.g., "plural", "upper", etc.), character strings that are not important words either alone or as a ending (e.g., "actual", "hundred"). etc.), character strings that indicate units and are not important words (for example, "km", "cm", "MB", "kΩ", etc.), etc.

For example, the important word/phrase extraction unit 2051 extracts an opening parenthesis and characters corresponding to the important word/phrase candidate character string pattern indicated by the important word/phrase candidate character string pattern information in the condition information, following the important word/phrase extracted from the claims of the invention document. When a string and a closing parenthesis are consecutively included in this order, a character string that corresponds to the important word/phrase candidate character string pattern may be extracted as an important word/phrase candidate. If the character string that corresponds to the important word candidate character string pattern is multiple words separated by a colon ":", and each of the multiple words corresponds to the important word candidate character string pattern, the important word The extraction unit 2051 may extract each of the plurality of words as candidates for important words.

The important word/phrase extracting unit 2051 extracts, as important words, words representing elements described as the subject of claims (so to speak, the elements to be claimed) in each claim section of the invention document, for example. For example, if the important word/phrase extraction unit 2051 includes a period character following a word that is the same as an important word included in the important word/phrase setting dictionary acquired by the important word/phrase setting dictionary acquisition unit 203 in a claim of an invention document, extracts the word as an important word representing the billing element. For example, if a claim of an invention document includes a period character following a character string that corresponds to the important word candidate character string pattern indicated by the important word candidate character string pattern information in the condition information, , the same character string as the character string corresponding to the key word candidate character string pattern is extracted as the key word/phrase representing the billing element. Whether or not the important phrase extracted by the important phrase extraction unit 2051 represents an element to be claimed in a claim may be one of the items of attribute information in the important phrase management information managed by the analysis unit 205, for example.

The key word candidate character string pattern used to search for the claimed element in the claims of the invention document is the same as the key word candidate character string pattern used to search for key word candidates related to elements other than the claimed subject matter. or may be partially different. As an example, the key word candidate character string pattern used to search for the claimed element is a predetermined character string that includes a hiragana "no" between them, which is sometimes used to express a loose combination of words; Used to search for important word candidates related to elements other than the billing subject, assuming that both the character string and another predetermined character string that do not include the hiragana "no" are the corresponding character string pattern. The important word/phrase candidate character string pattern may be a character string pattern that does not include character strings that include the hiragana "no" between them.

[3-1-5-1-2. Extraction of important words in the specification]
The important word extraction unit 2051 performs the specification important word extraction process, for example, in the specification of the invention document, in accordance with the important word extraction range information in the setting information. Important words and phrases are extracted in the following sections: ``Form for purposes'' and ``Examples''). The extraction of important words in the specification by the important words and phrases extraction unit 2051 is generally the same as the extraction of important words in the claims described above, except for some differences. Hereinafter, a similar explanation will be omitted as appropriate, but in order to extract important words in the extraction range of important words in the specification of an invention document, it is necessary to A similar method to word extraction may be used.

For example, the important word extraction unit 2051 searches for important word candidates for each line in the important word extraction range in the specification of the invention document, and after extracting the important word candidates, searches for important word candidates based on predetermined criteria. It is determined whether the words can be important words or phrases, and the words determined to be possible important words are extracted as important words or phrases.

Note that the important word/phrase extraction unit 2051 searches for important word/phrase candidates by excluding comment lines or comment portions from the search target range, for example, based on the comment specifying information of the setting information. Further, the important word/phrase extraction unit 2051 may search for important word/phrase candidates, for example, excluding the range of quoted descriptions from the search target range. The range of the cited description is, for example, the range between two corresponding separate character codes for the quotation, surrounded by quotation marks or parentheses. When expressed in dots, examples include "...", "...", "...", etc.

For example, the important word/phrase extraction unit 2051 may convert the character format of each line in the important word/phrase extraction range in the specification of the invention document before searching for key word candidates for each line.

In one line of the extraction range of important words in the specification of the invention document, a plurality of candidates for important words that are determined to be potentially important words are extracted, and the description range of the character strings of the candidates for the plurality of important words is extracted. If there are some overlaps, the important word/phrase extraction unit 2051 selects one of the plurality of important word candidates based on predetermined selection criteria and extracts it as the important word/phrase. Note that, for each line in the extraction range of important words in the specification of the invention document, the important word/phrase extraction unit 2051 extracts all the parts written in parentheses in that line (that is, the parts sandwiched between the opening parentheses and the ending parentheses). In this method, candidates for important words are extracted by separately targeting the part written outside the parentheses, which is formed by concatenating the remaining character strings after removing the written part), and the part written inside each parenthesis. Good too.

For example, the important word/phrase extraction unit 2051 detects whether the same words as the important words included in the important word/phrase setting dictionary acquired by the important word/phrase setting dictionary acquisition unit 203 are included in the important word/phrase extraction range in the specification of the invention document. If so, that word within the important word/phrase extraction range in the specification is extracted as an important word/phrase candidate. For example, the important word/phrase extraction unit 2051 does not extract words included in the non-important word/phrase setting dictionary acquired by the important word/phrase setting dictionary acquisition unit 203 as important word/phrase candidates.

For example, if the same word or phrase extracted as an important word or phrase in the claim of the invention document is included in the extraction range of important words or phrases in the specification of the invention document, the important word/phrase extraction unit 2051 extracts Words within the extraction range of important words in the book are extracted as important word candidates.

The important word/phrase extracting unit 2051, for example, refers to each character string pattern information indicated by the condition information of the setting information and extracts words in the important word/phrase extraction range in the specification of the invention document as important word/phrase candidates.

As a specific example, the important word/phrase extraction unit 2051 may include, for example, an important word/phrase candidate character string pattern indicated by the key word/phrase candidate character string pattern information in the condition information in the extraction range of important words/phrases in the specification of the invention document. If a code string exists following the relevant character string, the same character string as the character string that corresponds to the important word/phrase candidate character string pattern is extracted as an important word/phrase candidate. The code string is, for example, a character string that corresponds to the code string pattern indicated by the code string pattern information in the condition information. In addition, if the code string pattern information does not reflect the non-code standard in the setting information, the code string is a character string that corresponds to the code string pattern and is not a code based on the non-code standard. This is a character string that cannot be judged as For example, in a case where the "explanation of the code" is described in the specification of the invention document, the code string may be a character string that is presumed to be the code in the "explanation of the code".

For example, the code string pattern represented by the code string pattern information using a regular expression or the like is expressed so that a string consisting of characters that can be used as a code (for example, alphanumeric characters and predetermined symbols) corresponds to the code string pattern. . The code string pattern may be, for example, a string pattern in which a plurality of codes are connected by "~", "-", ",", comma character ",", etc. . If the non-code standard is not included in the code string pattern, the non-code standard is a character string that corresponds to the code string pattern, for example, a character string that is rarely used as a code and is not a code. This is information that determines that character examples that are estimated to have a sufficient possibility are determined not to be codes. Examples of character strings that are determined to be non-code based on the non-code criteria are character strings in which the first three or more characters are uppercase letters, and the first one or more characters in which the number is a predetermined unit character string (for example, "km", " Examples include character strings followed by "cm", "MB", "kΩ", etc.). The non-sign criterion may indicate a character string that is determined to be not a code in combination with the ending of the element name immediately before the code. For example, if the ending of the element name immediately before the code is "voltage". In certain cases, a code consisting of the first character or more of a number followed by a unit character string such as "V" or "mV", or the first character when the element name immediately before the code ends in "distance" Examples include codes made up of the above numbers followed by unit character strings such as "m", "cm", and "km". A non-coded standard is, for example, a string of characters that can be used as a code, followed by a predetermined character that rarely follows a code (e.g., ``tsu'', ``year'', ``second''). ", "times," etc.), a character string composed of characters that can be used as a code may be determined so that it is determined not to be a code.

For example, from the extraction range of important words in the specification of an invention document, the important word/phrase extraction unit 2051 selects candidate words with the same content from among the words and phrases that are extracted as a result of the subsequent presence of a code character string. A plurality of words and phrases that are extracted by the presence of subsequent code character strings, and which have a common ending but different parts other than the ending (for example, "equipment control section" and "first control section") ), the plurality of words may be excluded from the important word candidates, and the common ending part (for example, "control section") may be newly extracted as the important word candidate. In addition, since code character strings with the same content continue to exist from different positions in the extraction range of important words in the specification of the invention document, the important words and phrases extraction unit 2051 temporarily selects them as important words and phrases. The plural words and phrases extracted and excluded from the candidates (for example, "equipment control section" and "first control section") may be used as the same word in the warning information generated by the analysis section 205 for alerting, etc. It can be pointed out as a group of words with different expressions that are marked with a symbol.

For example, the important word extraction unit 2051 extracts an opening parenthesis and a condition following an important word extracted by the presence of a code character string in the extraction range of important words in the specification of an invention document. Important word/phrase candidate character string pattern in information Characters that correspond to the important word/phrase candidate character string pattern when a character string that corresponds to the important word/phrase candidate character string pattern indicated by the information and a closing parenthesis are included consecutively in this order. A column may be extracted as a candidate for an important term. If the character string that corresponds to the important word candidate character string pattern is a plurality of words separated by a colon ":", the important word extraction unit 2051 selects each of the plurality of words as a candidate important word, for example. You can also extract it as

The important word/phrase extraction unit 2051 extracts, for example, a character string whose ending includes a character string extracted as an important word in a claim of an invention document, into the important word/phrase candidate character string pattern indicated by the important word/phrase candidate character string pattern information in the condition information. If the relevant character string exists in the extraction range of important words in the specification of the invention document, the character string that corresponds to the important word candidate character string pattern can also be extracted as an important word candidate. good.

For example, the important word/phrase extraction unit 2051 extracts, in the extraction range of important words in the specification of the invention document, a character string that follows a character string corresponding to the important word/phrase candidate character string pattern indicated by the important word/phrase candidate character string pattern information in the condition information. When there is a character string that corresponds to the specific particle string pattern indicated by the specific particle string pattern information in the condition information, the same character string that corresponds to the important word candidate string pattern is added to the important word. Extract as a candidate. Note that the important word/phrase extraction unit 2051 includes, for example, a character string that follows the key word/phrase candidate character string pattern indicated by the key word/phrase candidate character string pattern information in the condition information in the extraction range of important words/phrases in the specification. Then, there is a predetermined character string (for example, "each", "each", etc.), and the predetermined character string is followed by a specific particle string pattern indicated by the specific particle string pattern information in the condition information. If a corresponding character string exists, the same character string as the character string corresponding to the important word/phrase candidate character string pattern may be extracted as an important word/phrase candidate.

The important term extraction unit 2051 detects, for example, that a character string corresponding to the important term candidate character string pattern indicated by the important term candidate character string pattern information in the condition information exists in the important term extraction range in the specification of the invention document. In this case, character strings corresponding to the important word/phrase candidate character string pattern may be extracted as important word/phrase candidates. In addition, characters extracted as important words simply because they correspond to the important word candidate character string pattern without satisfying additional conditions such as preceding a character string that corresponds to a code character string or a specific particle character string pattern. Regarding a string, the analysis unit 205 can determine the importance of the important term that is the character string to be relatively low under certain conditions, such as the number of entries of the important term in the invention document.

[3-1-5-1-3. Management of extracted important words]
The analysis unit 205 manages each important term extracted by the important term extraction unit 2051 using important term management information.

FIG. 18 shows an example of important phrase management information.

The important word management information shown in the example of FIG. 18 is information that associates each important word extracted by the important word extraction unit 2051 with identification information, importance, attribute information, a set of translation word candidates, and a corresponding translation word. . Each item of the important phrase management information shown in the example of FIG. 18 will be explained below.

The identification information of the important word/phrase is, for example, an identification code such as a number that is determined to be different for each of a plurality of important words/phrases having different contents (that is, character strings). Specifically, the identification information of important words and phrases includes, for example, an identification symbol that distinguishes important words and phrases extracted from the claims of the invention document and important words and phrases that were not extracted from the claims of the invention document, and an identification information of the same identification symbol. This is an identification code that includes a serial number for an important phrase. The important words and phrases extracted from the claims of the invention document can also be extracted from the specification of the invention document, and are hereinafter also referred to as important words and phrases of the claims. In addition, important words and phrases that were not extracted from the claims of the invention document are words that were not extracted from the claims of the invention document but were extracted from the specification, and are hereinafter also referred to as important words and phrases of the specification. When the number of important words and phrases of claims with mutually different content exceeds a predetermined number (for example, 500), the analysis unit 205 analyzes the important words of the claims that do not exceed the predetermined number based on the priority order of the important words. A mutually different identification code (for example, one of "A1" to "A500") is set for each important phrase in One predetermined identification code (for example, "A999") is determined. Similarly, when the number of important words in specifications with mutually different contents exceeds a predetermined number, the analysis unit 205 determines that the number of important words in the specifications does not exceed a predetermined number based on the priority order of the important words. A mutually different identification code (for example, "B1" to "B500") is set for each important word/phrase in the minutes, and each important word/phrase in the minutes exceeding a predetermined number is given a predetermined number, without being distinguished from each other. An identification code (for example, "B999") is determined. The priority of important words and phrases is determined, for example, based on the importance of the important words or attribute information (for example, first appearance number), and the identification code of the important words and phrases includes serial numbers in order according to the priority. The analysis unit 205 may employ, for example, a serial number determined in descending order of the first line number of the important phrase as the serial number part in the identification code for the important phrase extracted by the important phrase extraction unit 2051.

The importance level of the important phrase is determined by the importance level determination unit 2055 of the analysis unit 205.

The attribute information of the important phrase is determined by the attribute specifying unit 2056 of the analyzing unit 205.

The set of translation word candidates for important words and phrases is, for example, a set of words and phrases that are words and phrases translated from important words and phrases into a predetermined translation destination language and are candidates for corresponding translation words, and is determined by the translation unit 2054 of the analysis unit 205.

The corresponding translation word of the important word/phrase is, for example, a word/phrase selected by the user of the invention document analysis system 20 from the set of translation word candidates of the important word/phrase.

Each important term extracted from the invention document by the important term extraction unit 2051 is used in various processes in the invention document analysis system 20 by, for example, referring to important term management information. For example, the invention document analysis system 20 outputs patent application related information including information regarding each important phrase extracted from the invention document. The analysis unit 205 converts various information included in the patent application related information to be outputted by the output unit 208 into a translation unit 2054, an inter-element relationship determination unit 2058, a warning information generation unit 2059, a claim information generation unit 205a, and a specification. The complaint tree information generation unit 205b, claim tree generation unit 205c, etc. generate the information.

[3-1-5-2. Attribute identification part]
The attribute identifying unit 2056 identifies attributes for each important term extracted from the invention document by the important term extracting unit 2051. For example, the attribute specifying unit 2056 updates the attribute information of the important term management information held by the analyzing unit 205 so as to reflect the attribute specified for each important term. For example, the attribute specifying unit 2056 updates each item corresponding to the various attributes of the attribute information of the important phrase management information by specifying various attributes of the important phrase. Each part of the analysis unit 205 can utilize the attribute information of each important word/phrase extracted from the invention document by, for example, referring to the attribute information of the important word/phrase management information.

FIG. 19 shows an example of items of attribute information of important words and phrases in the important words and phrases management information. The attribute information of the important phrase illustrated in Figure 19 includes the item "first appearance line number", the item "first appearance claim number", the item "definition claim number set and reference claim number set", and the item "element to be claimed in the claim". "Representation or not" item, "Symbol" item, "Number of uses in the entire invention document, claim part, specification part, and invention embodiment part" item, "Whether or not it is used as a subject" item, " Items such as ``Constitutional inclusion relationship between the element expressed by an important term and elements expressed by other important terms'', ``Whether or not it has a technical term-specific suffix character string'', and ``Is the word included in the important term setting dictionary?'' ``Is it not?'', ``Number of characters in important word/phrase,'' and ``Whether it is a low importance word or phrase and whether it includes a character string ending in a low importance word.''

Each item of the attribute information of the important word/phrase illustrated in FIG. 19 will be explained below. This attribute information is defined for each important term extracted from the invention document by the important term extraction unit 2051, and each item indicates one of the attributes of the corresponding important term. The attribute specifying unit 2056 performs processing such as searching for descriptions related to the important word from the invention document for each important word, thereby identifying each attribute of the important word, that is, each attribute information of the important word. Specify the content of the item.

The item "first appearance line number" as an attribute of an important phrase indicates the number of the first line in which the important phrase is written in the invention document among the serial numbers determined in the order of description for each line of the invention document. As another example, the item "First appearance line number" indicates the number of the first line in which the important word or phrase is stated in the claim of the invention document, and if the important word or phrase is not stated in the claim, The key phrase may indicate the number of the first line in the specification of the invention document.

The "first claim number" item, which is an attribute of an important term, indicates the number of the first claim (for example, the claim number) of the claim section of the invention document in which the important term is stated, and If a word or phrase is not stated in a claim of an invention document, a number other than the claim item number (for example, 0) is indicated, for example. Here, the explanation will be continued assuming that when an important term is not stated in a claim of an invention document, the "first appearance claim number" as an attribute of the important term is 0. Thereby, it is possible to identify whether the important word/phrase is an important word/phrase of the claim (that is, whether or not it is stated in the claim) based on the contents of the item "First Appearance Claim Number". The important words and phrases of the claim (that is, the important words that are stated in the claim) may be relatively more important than the important words and phrases of the specification (that is, the important words that are not stated in the claim). Furthermore, if the first item of a claim in which an important phrase is stated has a relatively small item number other than 0, there is a slight possibility that the important phrase indicates a relatively high-level matter conceptually.

The item "definition claim number set and reference claim number set" as an attribute of an important phrase indicates a set whose elements are the item numbers of claim sections of the invention document in which the important phrase is stated. The reference claim number set is a reference character string (for example, a character string corresponding to the reference character string pattern indicated by the reference character string pattern information in the condition information) followed by the item number of the claim section that includes the description of the important phrase. It is a set of elements. The definition claim number set is a set whose elements are the item numbers of claim items that include descriptions of the important words that do not follow the reference character string. If the number of elements in the reference claim number set for an important term is 0, there is a possibility that the important term is a term that is not stated subsequent to the reference character string in the claim.

The item "Whether or not it represents an element to be claimed in a claim" as an attribute of an important phrase is information indicating whether the element represented by the important phrase is an element to be claimed in any one of the claims. It is.

The item "code" as an attribute of an important word or phrase indicates the code if there is a description in which a code is added to the important word in the invention document. There may be a plurality of codes. If a description in which a code is added to the important word cannot be detected in the invention document, it is indicated that the code is not included. For example, if the item "code" is expressed as a character string, the content of the item "code" is, for example, the description of the code itself if there is a description with a code added to the important word. If no description in which a code is added to the important word can be detected, the string will be a string containing only null characters. For example, if the item "code" is expressed as a set of character strings of codes, if a description in which a code is added to the important word cannot be detected, the item "code" The number of elements in the set of character strings that are the contents of is 0, and if there is a code corresponding to that important word, the number of elements in the set of character strings that are the contents of the "code" item is 1 or more. . Thereby, it is possible to identify whether or not a code corresponding to the important word/phrase exists based on the contents of the "code" item. An important word or phrase that is described with a code added in an invention document may be relatively more important than an important word or phrase that is not described with a code added. The item "Sign" as an attribute of an important term further indicates the code if there is a statement in which a code is added to the important term in the specification of the invention document; , when a description in which a code is added to the important word or phrase and a description in which a code is not added coexist, this may be indicated to indicate the mixture. For each important term extracted by the important term extracting unit 2051, the attribute specifying unit 2056 searches the invention document or the specification of the invention document for a portion where the important term is described, and determines where the important term is written. By determining whether or not a code follows an important phrase in a portion, the content of the item "code" as an attribute of the important phrase can be specified.

The item "number of uses in the entire invention document, claim part, specification part, and invention embodiment part" as an attribute of an important term indicates the number of times the important term is stated in the entire invention document, and its importance in the claims of the invention document. It shows the number of times a word has been written, the number of times the important word has been written in the specification of the invention document, and the number of times the important word has been written in the "Details for carrying out the invention" in the specification of the invention document. The attribute specifying unit 2056 specifies the content of this item by counting the number of times each important term is written. Relatively speaking, important words and phrases that have been described many times may have high importance.

The item "whether or not the important term is used as a subject" as an attribute of the important term indicates, for example, that the statement in which the important term is used as a subject is included in the invention document or in the specification of the invention document. Indicates whether or not there is. For example, when the attribute identification unit 2056 detects that an invention document includes a description in which the important word is placed immediately before the character "wa", the attribute identification unit 2056 determines that the important word is used as the subject. It is determined that the description is included. This determination result is expressed as the content of the item "Used as subject". Note that the attribute identifying unit 2056 determines, for example, that the important word includes a description placed immediately before the character "wa", and the character "ha" is determined to be a particle by morphological analysis. Only in this case, it may be determined that a description in which the important word or phrase is used as the subject is included. Further, the attribute specifying unit 2056 also determines that, for example, in an invention document or the like, the important word/phrase is ``ha'' with a predetermined character string (for example, ``each'', ``each'', ``each'', etc.) in between. Even when it is detected that a description placed before a character is included, it may be determined that a description in which the important word or phrase is used as a subject is included. In addition, the character "wa" written after the subject shown here can be a predetermined particle (e.g. "ga", "mo", etc.) or a predetermined continuous particle (e.g. "toha", " A predetermined character string that can follow a word corresponding to the subject, such as "ni", "de", etc., may be treated in the same way as the character "wa". In this case, one important word indicated in the important word candidate character string pattern information of the condition information is preceded by a predetermined character string that can follow the word corresponding to the subject, which is treated in the same way as the character "wa". When a character string corresponding to the candidate character string pattern is described in the invention document, the character string corresponding to the important word/phrase candidate character string pattern is extracted as an important word/phrase candidate by the important word/phrase extraction unit 2051. good. An important word or phrase that has an attribute indicating that it is used as a subject in the item "Used as a subject" is assumed to be a word or phrase to which some explanatory statement has been added in the invention document. Therefore, it can mean that the word is likely to have a relatively high degree of importance.

The item "compositional inclusion relationship between an element expressed by an important term and an element expressed by another important term" as an attribute of an important term indicates, for example, that the element expressed by the important term in an invention document is one of the other elements. In the case where each of the above key words structurally includes the element expressed, the set of key words expressing the included element, and the element expressed by the key word in the invention document are included in one or more other A set of important words and phrases representing elements contained in the element represented by each important word is shown. If the number of elements in each set is 0, this indicates that the element represented by the important word has no structural relationship with the element represented by one or more other important words. This item can be used for outputting information to make it easier for the user of the invention document analysis system 20 to understand the relationships between important words.

Items as one attribute of the attribute information of the important word/phrase identified by the attribute specifying unit 2056 are not limited to those mentioned above.For example, the attribute information may include "a plurality of" immediately before the key word/phrase in the specification of the invention document. , "one or more", "one", etc., when one or more descriptions with character strings representing numbers are added, a set whose elements are character strings representing numbers from each description. An item related to "number" indicating a set of numbers may be included as one attribute. In addition, the attribute information is an item indicating the category of an important term (for example, indicating to which of multiple types such as a product, a method, a program, etc., an important term belongs), which is determined based on the category classification standard information in the setting information. item) may be included as one attribute.

The attribute specifying unit 2056 causes the inter-element relationship determining unit 2058 to execute inclusion relationship determining processing, thereby determining the structural inclusion relationship that exists between the elements expressed by each of the plurality of important words and phrases, and determining the written content of each line of the invention document. Make a judgment based on the following. Note that there may not only be a structural inclusion relationship between elements expressed by multiple important terms classified as things, but also between elements expressed by multiple important terms classified as other than things, such as methods, programs, data, etc. It is also possible that there may be a structural inclusion relationship between the two. For example, a "control method" as an element of a method expressed by an important term may include "first control procedure", "second control procedure", etc. as elements of a method expressed by an important term. Good too.

Then, the attribute specifying unit 2056 determines, for each important term, a set of important terms related to the included side, as one attribute of the important term, according to the determination result by the inclusion relationship determining process of the inter-element relationship determining unit 2058. Then, a set of important words related to the inclusion side is specified.

The item "Whether or not it has a technical term-specific ending character string" as an attribute of an important term indicates that the important term includes a technical term-specific ending character string that includes a predetermined ending unique to technical terms. Indicates whether it is a character string. The attribute specifying unit 2056 specifies the content of this item by, for example, searching for a predetermined technical term-specific character string from the important term. The technical term-specific ending character string can be indicated by, for example, technical term-specific ending information in the setting information acquired by the setting information acquisition unit 201. Assuming that technical term-specific ending character strings are appropriately defined by technical term-specific ending character strings, important words and phrases with technical term-specific ending character strings have a good possibility of having technical importance. It means that it is a certain word.

The item “Whether or not the word is included in the important word/phrase setting dictionary” as an attribute of the important word/phrase indicates whether the important word/phrase is included in the important word/phrase setting dictionary acquired by the important word/phrase setting dictionary acquisition unit 203. shows. The attribute specifying unit 2056 specifies the content of this item, for example, by searching for the important term from the important term setting dictionary acquired by the important term setting dictionary acquisition unit 203. Although the accuracy of extraction of important words in the invention document analysis system 20 may not necessarily be high, when it is assumed that the important words and phrases setting dictionary is in a state that defines important words and phrases selected by the user, for example, When the content of the item "Whether or not it is included in the important word/phrase setting dictionary" of an important word/phrase indicates that it is included in the important word/phrase setting dictionary, the important word/phrase is a word that does not need to be newly registered in the important word/phrase setting dictionary. This also means that the word is likely to be relatively important.

The item "Number of characters of important word/phrase" as an attribute of the important word/phrase indicates the number of characters of the important word/phrase. For example, if a word with one character that does not follow the reference string and does not have a code added is extracted as an important word, the important word does not follow the reference string. There is a possibility that it is less important than an important phrase with two or more characters without a code. Also, for example, if a word with two characters that does not follow the reference string and has no code added is extracted as an important word, then that important word will be added to the reference string. It may be less important than an important word or phrase that has three or more characters that are not followed by a code.

The item "Whether or not it is a low importance word or phrase and whether it contains a character string ending in a low importance word" as an attribute of an important word or phrase indicates that the important word is a predetermined low importance word or phrase (for example, "morphology"). , "power," etc.), and whether the word includes a predetermined predetermined low-importance ending character string at the end. If the content of this item indicates that the important word or phrase is the same as a less important word or phrase, or is a word that includes a less important ending character string at the end, then the importance of the important word is relative. There is a low possibility that

[3-1-5-3. Importance determining section]
The importance determining unit 2055 determines the importance of each important term extracted from the invention document by the important term extracting unit 2051. For example, the importance determining unit 2055 determines the importance of each important term extracted from the claim part and the specification part of the invention document based on the attribute information of the important term specified by the attribute specifying unit 2056. . The importance determination unit 2055 may, for example, determine the importance of an important word representing an element to be claimed in a claim to be relatively high, or determine the importance of an important word that is described as being used as a subject in an invention document. may be determined to be relatively high, or the importance of an important word or phrase that is the same as a predetermined low importance word or phrase or that includes a predetermined low importance word ending character string may be determined to be relatively low. The importance level determination unit 2055 may determine the importance level to be relatively low for important terms extracted simply by matching the important term candidate character string pattern.

For example, the importance determining unit 2055 updates the important term management information held by the analyzing unit 205 to reflect the importance determined for each important term.

Below, as an example of determining the importance of each important term by the importance determining unit 2055, important terms in the claim (that is, important terms stated in the claim) and important terms in the specification (that is, important terms not stated in the claim) are shown below. An example of determining the importance of each important word/phrase by distinguishing it from important words (important words/phrases described in the specification) will be described. Here, the explanation will be given assuming that the larger the importance value, the higher the importance.

[3-1-5-3-1. Determination of importance of important words stated in claims]
First, the importance determining unit 2055 sets a first predetermined value (for example, “3”) as the initial value of the importance for each important term extracted by the important term extracting unit 2051 from the claims of the invention document. Then, the importance determining unit 2055 increases or decreases the set importance value for each important word based on the attribute information of the important word. When the importance determining unit 2055 increases or decreases the importance value of the important phrases of a claim, an upper limit value (for example, 5) and a lower limit value (for example, 0) may be determined in advance.

For example, if the content of the item "First Appearance Claim Number" as an attribute of the important term is other than 0 and smaller than a constant value (for example, 5), the importance determining unit 2055 determines the value of the set value for the important term. Increase the importance value by 1. This constant value may be a variable value, such as half of the total number of claims, for example. Further, the amount by which the importance value is increased for each of a plurality of different constant values may be varied, for example, such that the smaller the constant value, the larger the amount.

For example, if the number of elements of the reference claim number set in the content of the item "definition claim number set and reference claim number set" as one attribute of the important phrase is 1 or more, the importance determination unit 2055 determines the importance Increase the preset importance value for the word by 1.

For example, when the content of the item "Used as a subject" as an attribute of the important word indicates that it is used as a subject, the importance determination unit 2055 determines the set value for the important word. Increase the importance value by 1.

For example, if the content of the item "whether or not it has a technical term-specific ending character string" as an attribute of the important term indicates that it has a technical term-specific ending character string, the importance determining unit 2055 determines whether Increase the set importance value of the important word by 1.

For example, when the content of the item "Whether or not the word is included in the important word/phrase setting dictionary" as an attribute of the important word/phrase indicates that the word/phrase is included in the important word/phrase setting dictionary, , increases the set importance value for that important term by 1.

It should be noted that the amount by which the importance determination unit 2055 increases the set importance value of the important word in response to the attribute information item of the important word does not need to be unified to 1 for each item. Different amounts may be used.

The importance determining unit 2055 may reduce the set importance value of the important word/phrase by a predetermined amount corresponding to the item of the attribute information of the important word/phrase.

For example, the importance determining unit 2055 determines that the content of the item "Whether or not it is a low importance word or phrase and whether it includes a character string ending in a low importance word" as one attribute of the important word is a low importance word or phrase. When indicating that a word-suffix character string is included, the preset importance value for the important word/phrase is decreased by 1.

The importance determining unit 2055 updates (for example, increases or decreases) the importance according to several items of the attribute information of the important phrase as described above, and the final result is the importance of the important phrase. This means that the degree of importance is determined as follows.

[3-1-5-3-2. Determining the importance of important words that are not stated in the claims but are stated in the specification]
Among the important words and phrases extracted by the important word and phrase extracting unit 2051 from the specification of the invention document, the importance determining unit 2055 first sets a second predetermined value as the initial value of the importance for the important words and phrases that are not stated in the claims. (for example, "1"). The second predetermined value is, for example, a value that is less than or equal to the first predetermined value as the initial value of the importance of the important words and phrases described in the claim. Then, the importance determining unit 2055 increases or decreases the set importance value for each important word based on the attribute information of the important word. When the importance determining unit 2055 increases or decreases the importance value for important words in the specification, for example, it sets an upper limit value (for example, 4) and a lower limit value (for example, 0) that may be different from important words in the claim. It may be determined in advance.

For example, if the content of the item "code" as an attribute of the important word indicates that there is a code corresponding to the important word, the importance determination unit 2055 determines the set value for the important word. Increase the importance value by 1.

For example, when the content of the item "Used as a subject" as an attribute of the important word indicates that it is used as a subject, the importance determination unit 2055 determines the set value for the important word. Increase the importance value by 1.

For example, when the content of the item “whether or not it has a technical term-specific ending character string” as an attribute of the important term indicates that it has a technical term-specific ending character string, the importance determining unit 2055 determines whether Increase the set importance value of the important word by 1.

For example, when the content of the item "Whether or not the word is included in the important word/phrase setting dictionary" as an attribute of the important word/phrase indicates that the word/phrase is included in the important word/phrase setting dictionary, , increases the set importance value for that important term by 1.

The importance determination unit 2055 determines, for example, the specification of the invention document as the content of the item "number of uses in the entire invention document, claim part, specification part, and invention embodiment part" as an attribute of the important word. If the number of times an important term has been described in the "Detailed Description of the Invention" of the book is greater than a predetermined value (for example, "2"), the preset importance value for that important term. Increase by 1.

For example, when the content of the item "Number of characters of important word/phrase" as an attribute of the important word/phrase is more than a predetermined value (for example, "2"), the importance determination unit 2055 determines the number of characters of the important word/phrase. Increase the configured importance value by 1.

For example, the importance determination unit 2055 determines that the content of the item "Whether or not it is a low importance word or phrase and whether it includes a character string ending in a low importance word" as one attribute of the important word is a low importance word or phrase. When indicating that a word-suffix character string is included, the preset importance value for the important word/phrase is decreased by 1. Note that the importance determination unit 2055 determines, for example, that the content of the item "Whether or not it is a low importance word or phrase and whether it includes a character string ending in a low importance word" as one attribute of the important word is a low importance word or phrase; When indicating that a low-importance ending character string is included, the set importance value for the important word/phrase may be updated to "0".

The importance determining unit 2055 updates (for example, increases or decreases) the importance according to several items of the attribute information of the important phrase as described above, and the final result is the importance of the important phrase. This means that the degree of importance is determined as follows.

[3-1-5-3-3. Update of importance determined for important words]
For example, the importance determining unit 2055 may make a final determination by further updating the importance once determined for each important term.

An example of further updating of the once-determined importance by the importance-determining unit 2055 will be described below.

After once determining the importance of each important term, the importance determining unit 2055 determines the importance of the important term related to the included element with respect to the structural inclusion relationship between the elements represented by each important term. , further updated so that the value is the same or higher than the highest importance value among the importance levels for all the included elements included in the element that the important word represents. do. Note that the structural inclusion relationship between elements is determined by the inclusion relationship determination process of the inter-element relationship determination unit 2058.

In this example, for example, an element represented by the important word "sensor" whose importance was once determined to be "4" and an element represented by the important word "operation reception section" whose importance was once determined to be "2" In the case where the importance of the important word "input part" representing an element that includes two elements in the structure is once determined to be "2", the importance determining unit 2055 determines the importance of the important word "input part". The degree will be updated from "2" to "4". As a result, among the important words and phrases related to the group of elements on the included side between elements that have a structural inclusion relationship, the importance of the important word that has been determined to have the highest importance will eventually change to It will be propagated to the importance of the important phrase related to.

[3-1-5-4. Inter-element relationship determination section]
The inter-element relationship determination unit 2058 performs inter-element relationship determination processing to determine the relationship between elements expressed by each important phrase extracted from the invention document by the important phrase extraction unit 2051. The inter-element relationship determination unit 2058 generates important word/phrase relationship information indicating the determination result of the relationship between elements represented by each important word/phrase. In response to the determination result by the inter-element relationship determination unit 2058, the attribute identification unit 2056 identifies attribute information for each important phrase related to the determination result.

[3-1-5-4-1. Determination of multiple elements having a compositional containment relationship]
The inter-element relationship determination unit 2058 executes inclusion relationship determination processing to determine whether the elements represented by each of the important terms extracted by the important term extraction unit 2051 from the invention document have a structural inclusion relationship. Based on the result of the inclusion relationship determination process by the inter-element relationship determining unit 2058, the attribute specifying unit 2056 determines, as an attribute of the important term, “the structural inclusion of the element expressed by the important term and the element expressed by other important terms. Specify the content of the “Relationship” item.

The inclusion relationship determination process by the inter-element relationship determination unit 2058 includes, for example, a description that corresponds to a predetermined inclusion description pattern indicating that an element represented by one important term includes an element represented by another important term from the invention document. This is achieved by searching. The predetermined inclusive description pattern may be included in the configuration information acquired by the configuration information acquisition unit 201.

An example of a description that corresponds to the predetermined inclusive description pattern is that in a sentence where one important word or phrase precedes the case particle "wa" as the subject, another important word or phrase is a predetermined inclusive description character that indicates inclusion. string (e.g., "comprises," "includes," "contains," etc.) or other important words immediately precede a predetermined inclusive description character string, and An example is a description in which an important word or phrase exists immediately after the predetermined inclusive description character string. Any character string assumed to represent inclusion may be defined as the predetermined inclusion description character string. In addition, other important words and phrases in the example of the description that corresponds to this predetermined inclusive description pattern include conjunctions that express parallel or selection, the predetermined particle "to" that connects elements, and commas characters "," It may be enumerated with etc. in between. The predetermined inclusion description character string may be an expression such as a passive voice (for example, "provided for", "included in", etc.), and a further predetermined inclusion description pattern is defined in response to this expression. It's okay.

To give a specific example, the inter-element relationship determination unit 2058 determines that the statement "input section including a sensor and an operation reception section" in the invention document corresponds to a predetermined inclusive description pattern, and therefore determines the key word "input section". It is determined that the element represented by includes the element represented by the important word "sensor" and the element represented by the important word "operation reception section" due to its structure. In response to this determination result, the attribute specifying unit 2056 determines whether the attribute of the item "compositional inclusion relationship between the element represented by the important word and the element represented by other important words" is one of the attributes of the important word "input part". The important words "sensor" and the important words "operation reception unit" are included as elements of the set of important words representing the included elements.

To give another specific example, the inter-element relationship determination unit 2058 determines that the statement “the information processing device includes an input unit, a storage unit, a display information generation unit, and a display unit” in the invention document is Since it corresponds to the inclusive description pattern, the element represented by the important word "information processing device" is structurally divided into the element represented by the important word "input part", the element represented by the important word "storage part", and the element represented by the important word "display information". It is determined that the element represented by "generation part" and the element represented by the important term "display part" are included. In response to this determination result, the attribute specifying unit 2056 selects, for example, the item "structural inclusion relationship between the element represented by the important word and the element represented by other important words" as an attribute of the important word "input part". The key term "information processing device" is included as an element of the set of key terms representing the containing element. In addition, the attribute specifying unit 2056 determines, for example, the included side of the item "structural inclusion relationship between an element expressed by an important term and an element expressed by another important term" as an attribute of the important term "information processing device". The important words "input section", the important words "storage section", the important words "display information generation section", and the important words "display section" are included as elements of the set of important words representing the elements.

In the inclusion relationship determination process, the inter-element relationship determination unit 2058 determines, in each line of the invention document, the purpose of a verb (e.g., "detect", "accept", etc.) that corresponds to a character string other than the predetermined inclusion description character string representing the inclusion relationship. A word may be excluded from the search target range of descriptions that correspond to a predetermined inclusion description pattern. That is, in each line of the invention document, the inter-element relationship determination unit 2058 excludes the search start part of the line up to the verb part corresponding to other than the predetermined inclusive description character string, and determines that one important word or phrase is found from the remaining part. A description that corresponds to a predetermined inclusion description pattern indicating that the represented element includes the element represented by another important word (for example, another important word exists immediately before a predetermined inclusion description character string, and one important word is The inclusion relationship between the elements represented by the plurality of important words may be determined by searching for a description that exists immediately after the predetermined inclusion description character string. As a result, the inter-element relationship determination unit 2058 determines, for example, "sound", " Even if words such as ``user'' and ``operation'' are important terms, by excluding them, the element represented by the important word ``input part'' is structurally divided into the element represented by the important word ``sensor'' and the element represented by the important word ``operation receiving part.'' ” can be determined to include the element represented by “.

For example, when the inter-element relationship determination unit 2058 performs an inclusion relationship determination process on claims in an invention document, if a subject is not stated before a predetermined inclusive description character string in a part of a claim term, , unless a subject preceded by the case particle "wa" is clearly specified before that part, the key word or phrase representing the claimed element in that one claim is considered to be the subject. It may be determined that the element represented by the subject includes the element represented by the important word immediately before the predetermined inclusive description character string. In addition, the inter-element relationship determining unit 2058 may detect whether a subject is stated before a predetermined inclusive description character string in a part of a claim term, for example, in an inclusion relationship determination process performed for claims in an invention document. If the subject preceding the case particle "wa" is clearly specified before the part of the part, the element represented by the subject includes the element represented by the important word immediately before the predetermined inclusive description character string. It may be determined that there is one.

[3-1-5-4-2. Determination of multiple elements in the relationship between a recording medium and the contents of that recording medium]
For example, the inter-element relationship determination unit 2058 performs record relationship determination to determine whether the elements represented by each important phrase extracted from the invention document by the important phrase extraction unit 2051 have a relationship between a recording medium and the contents of the recording medium. Processing can be performed. Here, the recording medium may include what is called a storage medium. The attribute specifying unit 2056 adds, for example, the item “recorded content of the element represented by the important term” to the attribute information of the important term, and determines the content based on the result of the recording relationship determination process by the inter-element relationship determining unit 2058. It is also possible to specify the contents of the item.

The record relationship determination process by the inter-element relationship determination unit 2058 is performed, for example, when an invention document corresponds to a predetermined record description pattern indicating that an element represented by one important term has recorded content as an element represented by another important term. This is achieved by searching the description. The predetermined record description pattern may be included in the setting information acquired by the setting information acquisition unit 201.

An example of a description that corresponds to a predetermined record entry pattern is a sentence in which one important word/phrase precedes the case particle "wa" as the subject, and another important word/phrase follows a predetermined record entry character string (for example, "Record "," "remembers," "stores," etc.), or other important words exist immediately before the predetermined record entry character string, and one Examples include a description in which an important word/phrase exists immediately after the predetermined record-described character string. Any character string that is assumed to represent that information, data, etc. is recorded or stored may be defined as the predetermined record description character string. In addition, other important words and phrases in the example of the description that corresponds to this predetermined record writing pattern include conjunctions that express parallel or selection, the predetermined particle "to" that connects elements, and commas characters "," It may be enumerated with etc. in between. The predetermined record entry character string may be an expression such as a passive voice (for example, "recorded in", "stored in", etc.), and a further predetermined record entry pattern corresponding to this expression may be used. It may be determined.

To give a specific example, the inter-element relationship determination unit 2058 determines that the statement "a recording device that records a configuration file" in the invention document corresponds to a predetermined record description pattern, and therefore the element represented by the important phrase "configuration file" is is determined to be the recorded content of the recording medium represented by the key word "recording device". In response to this determination result, the attribute specifying unit 2056 includes, for example, the important term "configuration file" as the content of the item "recorded content of the element represented by the important term" as one attribute of the important term "recording device". .

To give another specific example, the inter-element relationship determining unit 2058 determines that the statement "The memory stores operation images, operation information, and device control information" in the invention document is based on a predetermined Since this corresponds to the record description pattern, the element represented by the important word "memory" is the element represented by the important word "operation image", the element represented by the important word "operation information", and the element represented by the important word "device control information". is determined to be the recorded content. In response to this determination result, the attribute specifying unit 2056 selects, for example, the important word "image for operation" as the content of the item "recorded content of the element represented by the important word" as an attribute of the important word "memory". The important words "operation information" and the important words "equipment control information" are included.

In the record relationship determination process, the inter-element relationship determination unit 2058 selects the object of the verb that corresponds to a character string other than the predetermined record description character string in each line of the invention document from the search target range of the description that corresponds to the predetermined record description pattern. May be excluded. That is, in each line of the invention document, the inter-element relationship determination unit 2058 excludes the search start part of the line up to the verb part corresponding to a character string other than the predetermined recorded character string, and identifies one important word from the remaining part. A statement that corresponds to a predetermined record description pattern indicating that the element represented is the recorded content of the element represented by another important word (for example, another important word exists immediately before a predetermined record description character string, It is determined whether the elements expressed by each of the plurality of important words have a relationship with the recording medium and the contents of the recording medium by searching for descriptions in which the word exists immediately after the predetermined record description character string. You may.

[3-1-5-4-3. Determination of multiple elements in the relationship between a program and its realization contents]
The inter-element relationship determination unit 2058 performs, for example, program realization relationship determination that determines whether the elements represented by each important phrase extracted by the important phrase extraction unit 2051 from the invention document have a relationship between a program and the content realized by the program. Processing can be performed. The attribute specifying unit 2056 adds, for example, the item “realization content of the element represented by the important term” to the attribute information of the important term, and based on the result of the program realization relationship determination process by the inter-element relationship determining unit 2058, It is also possible to specify the contents of the item.

The program implementation relationship determination process by the inter-element relationship determination unit 2058 is performed by, for example, identifying one important word or phrase that substantially indicates a program (for example, all or ending with "program", "routine", "software", etc.) from the invention document. The element expressed by the word (phrase) is realized by the element expressed by other important words indicating a method, process, function, etc. (for example, words ending in "method", "process", "means", "function", etc.) This is achieved by searching for a description that corresponds to a predetermined program description pattern that indicates that there is a program. The predetermined program description pattern may be included in the setting information acquired by the setting information acquisition unit 201.

An example of a description that corresponds to a predetermined program description pattern is a sentence in which one important word or phrase precedes the case particle "wa" as the subject, and another important word or phrase follows a predetermined program realization description character string (for example, "wa"). ``to realize'', ``to realize'', ``to function as'', etc.), or other important words exist immediately before the predetermined program realization description character string, and the first Examples include a description in which an important phrase exists immediately after the predetermined program implementation description character string. As the predetermined program realization description character string, any character string that is assumed to represent the realization of a function, process, etc., or that it is intended to realize a function, process, etc. may be defined. In addition, other important words and phrases in the example of the description that corresponds to this predetermined program realization description pattern include a conjunction that expresses parallel or selection, a predetermined particle "to" that connects elements, a comma character "," '', etc. may be listed. The predetermined program realization description character string may be an expression such as a passive voice (for example, "realized by", etc.), and a further predetermined program realization description pattern may be defined in response to this expression. good.

To give a specific example, the inter-element relationship determination unit 2058 determines that the statement "conversion program for realizing a character code conversion function" in the invention document corresponds to a predetermined program realization description pattern, and therefore It is determined that the element represented by "code conversion function" has a relationship such that it is the implementation content of the program represented by the important term "conversion program". In response to this determination result, the attribute specifying unit 2056 selects, for example, the important word "character code conversion function" as the content of the item "realization content of the element expressed by the important word" as an attribute of the important word "conversion program". Include.

To give another specific example, the inter-element relationship determination unit 2058 determines that the statement "The conversion routine causes the computer to function as a character code conversion means" in the invention document corresponds to a predetermined program implementation description pattern. Therefore, it is determined that the element represented by the important phrase "conversion routine" is realized by the element represented by the important phrase "character code conversion means." In response to this determination result, the attribute specifying unit 2056 selects, for example, the important word "character code conversion means" as the content of the item "realization content of the element represented by the important word" as an attribute of the important word "conversion routine". include.

In the program realization relationship determination process, the inter-element relationship determination unit 2058 determines, in each line of the invention document, the object of a verb that corresponds to a character string other than a predetermined program realization description character string that indicates that a function, process, etc. is realized by a program. , descriptions corresponding to a predetermined program implementation description pattern may be excluded from the search target range. That is, in each line of the invention document, the inter-element relationship determination unit 2058 extracts one important word from the remaining portion, excluding the search start portion of the line to the verb portion corresponding to a character string other than the predetermined program implementation description character string. A description that corresponds to a predetermined program realization description pattern that indicates that the element expressed by is for realizing the element expressed by another important word (for example, if another important word is placed immediately before a predetermined program realization description character string) By searching for descriptions such that one important word exists immediately after the predetermined program realization description character string), the elements expressed by each of the plurality of important words can be compared with the program and the content realized by that program. It may also be determined that there is a relationship.

[3-1-5-4-4. Determination of multiple elements in an equivalence relationship]
The inter-element relationship determination unit 2058 may, for example, execute an equivalence relationship determination process that determines whether there is an equivalence relationship between elements represented by a plurality of important terms extracted from the invention document by the important term extraction unit 2051. The attribute specifying unit 2056 adds, for example, the item “element having an equivalence relationship with the element represented by the important term” to the attribute information of the important term, and based on the result of the equivalence relationship determination process by the inter-element relationship determining unit 2058. It is also possible to specify the contents of that item.

The equivalence relationship determination process by the inter-element relationship determination unit 2058 is performed, for example, from an invention document to a predetermined key phrase in which one important phrase is followed by another important phrase that is presumed to express the same meaning as that one important phrase. This is achieved by searching for a description that corresponds to an equivalent description pattern. The predetermined equivalent description pattern may be included in the setting information acquired by the setting information acquisition unit 201.

An example of a description that corresponds to a predetermined equivalent description pattern is, for example, the case particle ``wa'' that follows one important word is followed by another important word, and the other important word is ``is''. Examples include descriptions followed by character strings such as ".", "Also referred to as.", etc.

Another example of a description that corresponds to the predetermined equivalent description pattern is a description in which an opening parenthesis, another important word, and a closing parenthesis are consecutively placed immediately after one important word or phrase. In this example, for example, a plurality of important words separated by colons ":" may be included between the opening parenthesis and the closing parenthesis as other important words. To give a specific example, the inter-element relationship determination unit 2058 determines that since the description "display device" in the invention document corresponds to a predetermined equivalent description pattern, the element represented by the important term "display device" is It is determined that the key word "display" has an equivalence relationship with the element expressed. To give another specific example, the inter-element relationship determination unit 2058 determines that the statement "Graphical User Interface (GUI)" in the invention document corresponds to a predetermined equivalent description pattern, so It is determined that the element represented by "Graphical User Interface" has an equivalence relationship with the element represented by each of the important term "GUI" and the important term "Graphical User Interface". In response to the determination result in this example, the attribute specifying unit 2056 selects the important term "GUI" and Include the key phrase "Graphical User Interface".

[3-1-5-4-5. Discrimination of multiple elements in a conceptual hierarchical relationship]
The inter-element relationship determination unit 2058 executes, for example, a hierarchical conceptual relationship determination process that determines that there is a conceptual hierarchical relationship between the elements represented by the plurality of important words extracted from the invention document by the important word extraction unit 2051. obtain. The attribute specifying unit 2056 adds, for example, the item “subordinate conceptual element of the element represented by the important term” to the attribute information of the important term, and based on the result of the hierarchical conceptual relationship determination process by the inter-element relationship determining unit 2058. It is also possible to specify the contents of that item.

The inter-element relationship determination unit 2058 performs the upper-lower concept relationship determination process, for example, from an invention document to identify other important words that are assumed to represent subordinate concepts (for example, concrete examples) following one important word or phrase. This is achieved by searching for a description that corresponds to a predetermined subordinate concept description pattern in which the word/phrase is written. The predetermined subordinate concept description pattern may be included in the setting information acquired by the setting information acquisition unit 201.

Examples of descriptions that correspond to the predetermined subordinate concept description pattern include, for example, after one important word, "is a specific example of", "is an example of", "is an example", "is a specific example of", etc. ”, etc. are placed, and the character string leading to the example expressions is followed by other important words and phrases, such as “dere” and “de”, which express assertions. An example is a description followed by a character string such as "There is." Note that comma characters "," may be included in the character string leading to this example expression. In addition, other important words and phrases in the example of the description that corresponds to this predetermined subordinate concept description pattern include a conjunction that expresses parallel or selection, a predetermined particle "to" that connects elements, a comma character "," '', etc. may be listed.

Another example of a description that falls under the predetermined subordinate concept description pattern is an opening parenthesis immediately after one important word, a character string representing an example such as "for example," another important word, and a closing parenthesis. Examples include descriptions arranged consecutively. A character string representing an example such as "for example" may include a comma character "," at the end of the word.

To give a specific example, the inter-element relationship determining unit 2058 determines that the statement "An example of a display device is a liquid crystal display" in the invention document corresponds to a predetermined subordinate concept description pattern, It is determined that the element as a subordinate concept of the element represented by "" is the element represented by the important term "liquid crystal display". In response to this determination result, the attribute specifying unit 2056 selects, for example, the important term "liquid crystal display" as the content of the item "lower conceptual element of the element expressed by the important term" as an attribute of the important term "display device". Include.

[3-1-5-4-6. Generation of important word related information]
The inter-element relationship determination unit 2058 outputs the results from the output unit 208 based on the results of the inclusion relationship determination process, the recording relationship determination process, the program realization relationship determination process, the equivalence relationship determination process, or the superior-superior concept relationship determination process. Generate important word/phrase related information included in patent application related information.

In the invention document analysis system 20, when the output unit 208 outputs patent application related information including important term related information as a display document (for example, an HTML document), the inter-element relationship determination unit 2058 outputs the patent application related information including important phrase related information as a display document. Generate important phrase related information included in patent application related information. The important term/phrase relationship information portion of the patent application related information as a display document, which is generated by the inter-element relationship determining unit 2058, includes display elements corresponding to each important term, for example. For example, a browser may interpret a document for display and display each display element. The inter-element relationship determining unit 2058 may, for example, add display control information to the display element corresponding to each important word or phrase for differentiating the display mode for each different important word or phrase.

[3-1-5-4-6-1. Generation of inclusion relationship information]
The inter-element relationship determining unit 2058 includes, for example, structural information between elements expressed by each of the plurality of important terms extracted by the important term extracting unit 2051 from the invention document, based on the important term relationship information. Includes inclusion relationship information that expresses the inclusion relationship. The structural inclusion relationship between elements in the inclusion relationship information may be expressed in any manner. FIG. 20 shows a display example of the inclusion relation information part in the important word/phrase relation information of the patent application related information as a display document.

As a specific example, the inter-element relationship determination unit 2058 uses, as inclusion relationship information, a plurality of terms connected by the citation relationship detected by the claim citation relationship detection unit 2057, for example, among the terms of the claims of the invention document. For each group constituted, the important words and phrases on the including side and the included side shall be connected with lines, ruled line symbols "└", etc. based on the inclusion relationship between the important words stated in each clause of the claims in that group. A description representing a tree structure (that is, a hierarchical structure) formed by is generated. The inter-element relationship determining unit 2058 may simplify the inclusion relationship information by omitting the second and subsequent keywords so that one important term expressed in the hierarchical structure appears only once in the hierarchical structure. As an example of this case, for example, when one important term expressed in a hierarchical structure appears again in the same hierarchy or a lower hierarchy, the inter-element relationship determination unit 2058 may Omit the expression. At this time, the inter-element relationship determining unit 2058 may add a symbol "..." or the like to the important phrase to express that the re-description of one important phrase is omitted.

For example, in the inclusion relationship information including descriptions of a plurality of important words, the inter-element relationship determining unit 2058 determines whether an important word representing a claimed subject in any section of a claim of an invention document is used to express the claimed subject in the important word. It is also possible to express the word so that it can be distinguished from other important words by adding a symbol (for example, "★") defined in . According to such inclusion relationship information, it becomes easy to understand the structural inclusion relationship between the elements represented by the important phrases claimed in the claim and other important phrases.

For example, the inter-element relationship determination unit 2058 determines whether a certain group is described in inclusion relationship information that includes descriptions of a hierarchical structure representing multiple important terms for each group composed of multiple terms connected by citation relationships. Each important word/phrase may be appended with a section number indicating in which section within the group the important word/phrase is written.

[3-1-5-4-6-2. Generation of record-related information]
The inter-element relationship determining unit 2058 determines, based on the important term relationship information, for example, that the elements represented by each of the plurality of important terms extracted by the important term extracting unit 2051 from the invention document are related to the recording medium based on the result of the recording relationship determination process. If there is a relationship with the contents of the recording medium, recording relationship information expressing the relationship may be included. The relationship between elements in recording-related information can be expressed, for example, by associating important words and phrases related to the recording medium with important words and phrases related to the content, and by placing important words and phrases related to the recording medium at the top of the list. The expression may be expressed using any method, such as an expression expressed in a hierarchical structure with one or more important words and phrases related to the content at a lower level.

[3-1-5-4-6-3. Generation of program realization related information]
The inter-element relationship determining unit 2058 uses the important term relationship information to determine, for example, the relationship between the elements represented by each of the plurality of important terms extracted by the important term extracting unit 2051 from the invention document based on the result of the program realization relationship determination process. If there is a relationship with the content realized by the program, program implementation relationship information expressing the relationship may be included. The relationship between elements in the program implementation related information may be expressed using any method, such as, for example, an expression in which a program and implementation content are associated with each other.

[3-1-5-4-6-4. Generation of equivalence relationship information]
The inter-element relationship determination unit 2058 includes, in the important term relationship information, an equivalent relationship between elements represented by each of the plurality of important terms extracted from the invention document by the important term extraction unit 2051, based on the result of the equivalence relationship determination process, for example. Equivalence relation information expressing the relation in a certain case can be included. The expression of the relationship between elements in the equivalence relationship information may be expressed by any method, such as an expression in which important words and phrases related to elements having an equivalence relationship are associated with each other.

[3-1-5-4-6-5. Generation of hierarchical concept relationship information]
The inter-element relationship determination unit 2058 uses the important term relationship information to determine, for example, a conceptual relationship between the elements represented by each of the plurality of important terms extracted by the important term extraction unit 2051 from the invention document based on the result of the hierarchical concept relationship determination process. In the case where there is a hierarchical relationship, it is possible to include hierarchical conceptual relationship information that expresses the relationship. Expressions of relationships between elements in the hierarchical conceptual relationship information include, for example, expressions that associate important words and phrases related to multiple elements that have a conceptual hierarchical relationship, and important words and phrases related to superordinate concepts (for example, "elastic body"). ) may be expressed in any way, such as in a hierarchical structure with one or more important words related to the corresponding subordinate concept (e.g. "spring" and "rubber") at the bottom. .

[3-1-5-5. Claim citation relationship detection section]
The claim citation relationship detection unit 2057 executes a citation relationship analysis process that analyzes the citation relationship of each clause of the claim in the invention document acquired by the invention document acquisition unit 2021, and generates a result of this citation relationship analysis process. Maintain complaint citation related information.

The citation relationship analysis process by the claim citation relationship detection unit 2057 analyzes the heading description of each claim in the invention document (for example, the "Claim" line between the opening corner brackets and the ending corner brackets). This is achieved by searching for a description that corresponds to a predetermined cited item number description pattern from the written content of the portion other than the description). The predetermined quotation destination item number description pattern may be included in the setting information acquired by the setting information acquisition unit 201.

An example of a description that corresponds to a predetermined cited reference item number description pattern is, for example, a cited reference item number character string containing a number is placed after the character string "Claim" representing a claim section, and the cited reference number is An example is a description in which the character string "description" is placed after the previous item number character string. When the claim citation relationship detection unit 2057 detects a description that corresponds to a predetermined quotation destination item number description pattern in the citation relationship analysis process, for example, the claim citation relationship detection unit 2057 uses a number in the citation destination item number character string as the citation destination item number. Extract. The claim citation relationship detection unit 2057 detects, for example, if a plurality of numbers in the cited item number character string in the description corresponding to the detected predetermined cited item number description pattern are conjunctions (for example, "or") expressing parallelism or selection. , or if they are arranged separated by delimiters (for example, ",", comma characters ",", etc.), each of the plurality of numbers is extracted as the cited item number. Further, the claim citation relationship detection unit 2057 detects, for example, that the cited item number character string in the description corresponding to the detected predetermined cited cited item number description pattern is a range expression character (for example, " ~'') is placed and includes a character string with a second number following the range expression character, all characters greater than or equal to the first number and less than or equal to the second number are included. Extract each integer as a quotation destination item number. Specifically, the claim citation relationship detection unit 2057 calculates the integers ``1'', ``2'', ``5'', ``based on the cited statement ``claim 1, 2, 5, or 11'' in a certain claim section. 11'' is extracted as the cited item number, and the integers ``1'', ``2'', and ``3'' are extracted based on the cited statement ``stated in any one of claims 1 to 3'' in the claim section. is extracted as the cited item number.

The claim citation relationship detection unit 2057 obtains claim citation relationship information that includes, for each term of a claim in an invention document, information that associates that term with a set of cited item numbers extracted for that term in the citation relationship analysis process. , may be updated and retained.

The claim citation relationship detection unit 2057 also performs citation relationship analysis processing on the claims of the revised claim information when the claim information generation unit 205a or the like generates revised claim information representing the revised content of the claims of the invention document. The revised claim citation relationship information generated as a result can be retained.

[3-1-5-6. Warning information generation section]
The warning information generation unit 2059 executes a warning information generation process to generate warning information included in the patent application related information output by the output unit 208 of the invention document analysis system 20.

The warning information generation unit 2059 performs the warning information generation process by searching for inappropriate descriptions included in the invention document acquired by the invention document acquisition unit 2021, and generates one or more points indicating indications regarding the detected inappropriate descriptions. This is achieved by generating warning information including information. The inappropriate description searched by the warning information generation unit 2059 may be a description that may be inappropriate. FIG. 21 shows an example of items of warning information generated by the warning information generation unit 2059 and entries corresponding to the items.

If there is no indication information as the content of the warning information, the warning information generation unit 2059 generates warning information indicating that there is no warning. Note that the warning information generating unit 2059 may omit generating the warning information when there is no indication information as the content of the warning information.

[3-1-5-6-1. Points out based on the standard provisions of invention documents for translation]
The warning information generation process by the warning information generation unit 2059 includes, for example, checking whether the invention document satisfies a plurality of conditions defined in the invention document standard regulations for translation acquired by the invention document standard regulations for translation acquisition unit 204. This is achieved by including, for example, indication information in the warning information that points out when the conditions are not met.

As a specific example, for each condition item (see FIG. 14) specified in the invention document standard regulations for translation, the warning information generation unit 2059 generates information in the invention document that corresponds to the pattern indicated by the pattern information corresponding to the condition item. When searching for an inappropriate description and detecting the inappropriate description, for example, information indicating the position in the invention document where the inappropriate description was detected, and the name of the condition item corresponding to the condition item. The warning information includes indication information indicating the above. Information indicating the position where an inappropriate description is detected in the invention document may include, for example, the line number of the inappropriate description among the line numbers assigned to each line of the invention document in line order, and the line number of each line of the claim in the invention document. It may also include the number of the section containing the inappropriate statement, the paragraph number of the paragraph containing the inappropriate statement among each paragraph of the specification in the invention document, etc. For example, the warning information generation unit 2059 uses a plug-in module for performing processing to determine whether the invention document conforms to each of a plurality of items of the standard provisions for invention documents for translation. Warning information generation processing may also be performed.

As an example of the warning information generation process, the warning information generation unit 2059 may, for example, satisfy the condition that each sentence of the specification includes a subject in the specification of the invention document acquired by the invention document acquisition unit 2021. If it is detected that a description that does not have a subject is included, indication information indicating that there is no subject and the content of the detected description is included in the warning information. As a specific example, the warning information generation unit 2059 detects an error related to the condition item “lack of subject in each sentence of the specification” shown in FIG. 14 from the specification of the invention document acquired by the invention document acquisition unit 2021. A description of a sentence that corresponds to a pattern indicated by pattern information for extracting an appropriate description is searched, and when a description of a sentence that corresponds to the pattern is detected, indication information is included in warning information. Note that there may be cases in which technical experts and non-technical experts can assume an implicit subject even if there is no subject, but this pointing information can be useful for calling attention.

For example, when the warning information generation unit 2059 detects that the invention document acquired by the invention document acquisition unit 2021 includes a description that does not satisfy the condition that the invention document does not include a predetermined inappropriate description, The warning information includes indication information indicating that a potentially inappropriate description is included and the content of the detected description. As a specific example, the warning information generation unit 2059 generates the conditions for “character strings that are estimated to be desirably removed” in the invention document standard regulations for translation (see FIG. 14) from the invention document acquired by the invention document acquisition unit 2021. Search for descriptions of character strings that correspond to the pattern indicated by the pattern information for extracting inappropriate descriptions related to items (for example, "only", "preferably", etc.), and search for descriptions of character strings that correspond to that pattern. When detected, the indication information is included in the warning information. For example, character strings such as "only" and "preferably" are not necessarily impossible to write, but they may be written inadvertently, so this pointing information can be useful for alerting people. .

For example, the warning information generation unit 2059 satisfies the condition that the invention document acquired by the invention document acquisition unit 2021 does not include inappropriate predetermined words or phrases for the important words extracted by the important phrase extraction unit 2051. If it is detected that a description that is not included is included, the system will send indication information indicating that a description of important words with inappropriate predetermined words has been added and the content of the detected description. Include in warning information. As a specific example, the warning information generation unit 2059 generates “character strings that may be preferably not added to important words in claims” shown in FIG. 14 from the claims of the invention document acquired by the invention document acquisition unit 2021. To extract inappropriate descriptions related to condition items, search for descriptions of character strings that correspond to the pattern indicated by the pattern information (for example, "desired", "predetermined", etc.), and search for character strings that correspond to the pattern. If a description of the above is detected, the indication information is included in the warning information. For example, the character string "predetermined" does not necessarily cause a problem, but it may be written inadvertently, so this pointing information can be useful for calling attention.

For example, the warning information generation unit 2059 prevents the invention document acquired by the invention document acquisition unit 2021 from including a description that would cause a disadvantage in terms of rights interpretation when the invention document is used to obtain rights in the United States. If it is detected that there is a statement that does not satisfy the condition that the condition is not met, information indicating that the statement that causes disadvantage in terms of rights interpretation is included and the content of the detected statement. be included in the warning information. As a specific example, the warning information generation unit 2059 generates the “character string in which it may be desirable to change to another expression in a U.S. application” shown in FIG. 14 from the invention document acquired by the invention document acquisition unit 2021. Search for descriptions of character strings that correspond to the pattern indicated by the pattern information for extracting inappropriate descriptions related to condition items (for example, "What are the characteristics of the present invention?"), and search for descriptions of character strings that correspond to the pattern. When detected, the indication information is included in the warning information. For example, the phrase "features of the invention" included in a U.S. application based on an invention document refers to features of the claimed subject matter (e.g., features of the invention as claimed in each and every section of the claim). Since there is a possibility that it could be interpreted as referring to the above), this indication information can be useful for alerting people.

For example, the warning information generation unit 2059 may be configured to provide information on the invention document acquired by the invention document acquisition unit 2021 in a manner that is not understood in a foreign country when the invention document is used to acquire rights in a foreign country (that is, a country other than Japan). If it is detected that the document contains descriptions that do not meet the condition of not including words or phrases that may be difficult to translate, the document may contain descriptions that may be difficult to understand in foreign countries. The warning information includes indication information indicating the fact and the content of the detected description. As a specific example, the warning information generation unit 2059 detects inappropriate descriptions related to the condition item “Use of unique Japanese terminology that is not well known in foreign countries” shown in FIG. 14 from the invention document acquired by the invention document acquisition unit 2021. Search for a description of a character string that corresponds to the pattern indicated by the pattern information for extracting the information (for example, "Lost thing", "Ikigai", etc.), and if a description of a character string that corresponds to that pattern is detected, the pointed out information will be displayed. Include in warning information. In the pattern indicated by the pattern information for extracting inappropriate descriptions related to the condition item "Use of unique Japanese terminology that is not well known in foreign countries", the symbol "○" has a good meaning in Japan, and the symbol "○" has a bad meaning in Japan. A symbol having a meaning unique to Japan, such as the symbol "

For example, the warning information generation unit 2059 generates, in the specification of the invention document acquired by the invention document acquisition unit 2021, important words and phrases with an importance level higher than a predetermined level (for example, importance value If it is detected that there is a statement that does not satisfy the condition that the key word or phrase (with 1 or more) is stated in the specification of the invention document, the key word or phrase will be explained in the specification. The warning information includes indication information indicating that there is a possibility that the warning information has not been used, and the important words and phrases. The importance level of the important phrase is determined by the importance level determination unit 2055. For example, the predetermined degree may be fixedly determined in advance, or may be determined in the configuration information acquired by the configuration information acquisition unit 201. As a specific example, the warning information generation unit 2059 extracts from the invention document acquired by the invention document acquisition unit 2021 the information shown in FIG. When a description is detected that corresponds to the pattern indicated by the pattern information for extracting inappropriate descriptions related to the condition item ``Not described in If it is detected that something has not been done), include the indication information in the warning information.

When an inappropriate description is detected in an invention document, the warning information generation unit 2059 determines the reason why one condition item specified in the invention document standard regulations for translation, which was the basis for the detection, is inappropriate. When the indication information includes a condition item explanation message such as an explanation to the user, the condition item explanation message may be included in the indication information. For example, the warning information generation unit 2059 may include, in the warning information, version information (for example, version number) of the invention document standards for translation used as a basis for detecting inappropriate descriptions.

For example, the warning information generation unit 2059 generates one or more conditions narrowed down by the setting information acquired by the setting information acquisition unit 201 or the selection operation by the user, from among a plurality of conditions defined in the standard regulations for invention documents for translation. An invention in which an evaluation value of the conformity (e.g. conformance rate, etc.) of an invention document to the invention document standard provisions for translation is determined based on the results of an inspection to see whether the conditions are met, and the evaluation value is indicated. Document evaluation information may be included in the warning information.

The warning information generation process by the warning information generation unit 2059 may be performed, for example, instead of inspecting inappropriate descriptions in invention documents based on the standard provisions for invention documents for translation, or based on the standard provisions for invention documents for translation. In addition to this, it may also be carried out based on various conditions set separately from the standard provisions for invention documents for translation. In this case, in the warning information generation process, the warning information generation unit 2059 inspects whether the description in the invention document satisfies various conditions set separately from the standard provisions for invention documents for translation, and When one or more conditions are not met, for example, indication information pointing out that the condition is not met, and information indicating the position in the invention document where the statement where the condition is not met is detected. may be included in the warning information.

[3-1-5-6-2. Indications based on various conditions established separately from the standard provisions for invention documents for translation]
Hereinafter, an example will be described in which the warning information generation unit 2059 points out a description in an invention document that does not satisfy the conditions, based on various conditions set separately from the standard provisions for invention documents for translation, in the warning information generation process. . Note that information related to one or more of the various conditions shown here may be included in the condition items of the standard provisions for invention documents for translation. An invention document that includes a description that does not satisfy one or more of these various conditions used in the warning information generation process does not necessarily have to be inappropriate; , may be defined for the purpose of calling attention to the user in view of the possibility that it may be inappropriate in substance or form.

For example, the warning information generation unit 2059 determines whether each heading (for example, a description with an opening corner bracket and an ending corner bracket) is appropriate in the specification of the invention document acquired by the invention document acquisition unit 2021. If it is detected that a description that does not satisfy the condition is included, the warning information includes information indicating the position of the description and indication information indicating that the condition is not met. . A list of character strings for appropriate headings in the specification of an invention document can be predefined based on, for example, laws and regulations, so the warning information generation unit 2059 can determine the specification of an invention document by referring to the list of character strings. It is possible to detect descriptions related to inappropriate headings in books. The list of character strings for the appropriate heading is, for example, a list of character strings such as "Title of Invention" and "Technical Field." Note that since a serial number may be added to the end of the character string of some headings, the warning information generation unit 2059 further adds characters to which a serial number can be added to the end of the word in order to detect descriptions related to inappropriate headings. A list of columns (for example, "Patent Documents", "Examples", etc.) may be used.

For example, if the warning information generation unit 2059 detects that the invention document acquired by the invention document acquisition unit 2021 includes a description that does not satisfy the condition of not including blank lines, the warning information generation unit 2059 generates the description. The warning information includes information indicating the location of the condition and indication information indicating that the condition is not met. A blank line is, for example, a line containing only a line feed code. For example, a line that does not contain any characters other than half-width or full-width blank spaces may be regarded as a blank line. By specifying a description that is considered to be a blank line in the setting information, the warning information generation unit 2059 refers to the specification regarding the description that is considered to be a blank line in the setting information acquired by the setting information acquisition unit 201. It may also be possible to recognize blank lines.

For example, when the warning information generation unit 2059 detects that the invention document acquired by the invention document acquisition unit 2021 includes a description that does not satisfy the condition that a character string that is a registered trademark is not included, The warning information includes information indicating the position of the description and indication information indicating that the condition is not met. Since the list of character strings of registered trademarks can be defined, the warning information generation unit 2059 can identify descriptions that include character strings that are registered trademarks in invention documents by referring to the list of character strings of registered trademarks. Can be detected. The warning information generation unit 2059 also collects character strings of multiple registered trademarks via the network by accessing a predetermined website, database, etc. that provides information on multiple registered trademarks on the Internet. Then, a list of character strings of registered trademarks may be generated and used for the warning information generation process. Note that an example of a description that does not satisfy the condition of not including a character string that is a registered trademark is a description of a character string that is a registered trademark. In addition, the warning information generation unit 2059 may include a character string clearly indicating that it is a registered trademark, such as "(registered trademark)" immediately after the character string that is a registered trademark, among the descriptions of character strings that are registered trademarks. Only the description of a character string that is a registered trademark that is not included may be treated as a description that does not satisfy the condition of not including a character string that is a registered trademark.

For example, the warning information generation unit 2059 detects that the invention document acquired by the invention document acquisition unit 2021 includes a description that does not satisfy the condition that the same element name is written immediately before the same code. If this is detected, the warning information includes indication information indicating the same code and element names of a plurality of different expressions to which the code is added. Note that, for example, when the warning information generation unit 2059 detects a description in which the same code is attached to a plurality of element names with different expressions in an invention document, the warning information generation unit 2059 generates a single element name between each element name. The indication information may be expressed in such a way that it is possible to identify whether or not the character string of another element name includes the character string at the end. As a specific example, the specification of the invention document includes the statement "The first device 101 and the second device 102 are..." and further includes the statement "The device 102 is... and the third device is..." 101 is...", the indication information includes, for example, "same code different expression information → 101: 1st device, 3rd device" and "same code different expression information → 102: Device (second device).

For example, the warning information generation unit 2059 generates a description in the specification of the invention document acquired by the invention document acquisition unit 2021 that does not satisfy the condition that element names for which a code is specified are written with a code added. If it is detected that a symbol is included, the element name and its code are included, and a statement pointing out that there is a mixture of descriptions with a code added to the element name and descriptions without a code. information in the warning information. As a specific example, if the specification of the invention document contains both the description "liquid crystal display 301" and the description "liquid crystal display" without "301" added, the warning information generation unit 2059: The warning information includes pointing information pointing out that in the specification of the invention document, the description "liquid crystal display 301" with a code added to the element name and the description "liquid crystal display" with no code added are mixed. let

For example, the warning information generation unit 2059 adds the codes used in the drawing document acquired by the drawing document acquisition unit 2022 to the element names in the specification of the invention document acquired by the invention document acquisition unit 2021. If it is detected that a description that does not meet the conditions for being described is included, indication information indicating that the code is not used in addition to the element name in the specification is provided. be included in the warning information. Note that the code in a drawing document is a searchable PDF file, for example, when the drawing document is the content of a searchable PDF file, or when a PDF file containing the drawing document as content can be converted into a searchable PDF file. It can be specified by searching for a character string that matches the code pattern indicated by the code string pattern information in the condition information of the setting information.

For example, the warning information generation unit 2059 does not write, in the invention document acquired by the invention document acquisition unit 2021, a code that is not used in the drawing document acquired by the drawing document acquisition unit 2022 by adding it to the element name. If it is detected that a description that does not meet the above conditions is included, the existence of a description in which a code that is not used in the drawing document is added to the element name and the position of the description in the invention document will be sent. Include indication information indicating this in the warning information. For example, if the drawing document is the content of a searchable PDF file, or if a PDF file containing the drawing document as content can be converted into a searchable PDF file, the warning information generation unit 2059 may By searching each code added to the element name from a searchable PDF file, for example, codes that are not used in the drawing document can be identified.

For example, the warning information generation unit 2059 generates a single line in the specification of the invention document acquired by the invention document acquisition unit 2021 in a line other than the heading line (for example, “title of invention”, “technical field”, etc.). If it is detected that there is a statement that does not satisfy the condition that the period character "." is placed at the end of the description in the text, the description of the invention document will be written to the effect that there is no period character ".". The warning information includes indication information indicating the position of the line in the warning information. In addition, the contents of the line start with a type of opening bracket (e.g., opening angle bracket, opening parenthesis, etc.), or start with a blank character followed by a type of opening bracket, and end with the corresponding type of closing bracket. The rows may be treated similarly to the heading rows here. Descriptions on lines other than the heading line do not necessarily have to end with the period character ``.'', but if you forget to add the period character ``.'' where it should be, or the sentence is This indication information can be useful for calling attention to cases where there is a break in the middle, etc.

For example, the warning information generation unit 2059 generates claims that do not satisfy the condition that one or more period characters “.” are written in each claim section of the invention document acquired by the invention document acquisition unit 2021. If it is detected that a paragraph is included, the warning information includes indication information indicating that the period character "." is not included and the paragraph number of the paragraph in the claims of the invention document.

For example, if there is a parenthetical description between an opening parenthesis and an ending parenthesis in the specification of an invention document acquired by the invention document acquisition unit 2021, the warning information generation unit 2059 generates the parenthetical description. and the description of the character string immediately before the parenthesized description is clear (in other words, the meaning of the parenthetical description is clear). When detected, the warning information includes indication information indicating that the meaning of the parenthetical statement is not clear and the position of the parenthetical statement in the specification of the invention document. Specifically, the warning information generation unit 2059 searches, for example, for parenthetical descriptions that follow character strings other than blank characters in the specification of the invention document, and detects parenthetical descriptions whose significance is clear. If it does not correspond to a predetermined proper parenthesized character string pattern representing the description, it is determined that the condition that the meaning of the parenthesized description is clear is not satisfied. The predetermined proper parenthesized character string pattern is, for example, a statement in parentheses that is the English notation of the technical term that is the immediately preceding word, a statement that is an example of the immediately preceding word, or a synonym or addition of the immediately preceding word. In the case of an explanatory statement, it is determined to correspond to this predetermined appropriate character string pattern in parentheses. The setting information includes proper parenthesized character string pattern information indicating a predetermined proper parenthesized character string pattern, and the warning information generation unit 2059 generates a proper parenthesized character string pattern for the setting information acquired by the setting information acquisition unit 201. An appropriate character string pattern within parentheses may be identified and used by referring to the character string pattern information within parentheses. Examples of descriptions that fall under the specified proper parenthesized character string pattern include half-width or full-width alphabetic characters, blank characters, slashes "/", hyphens "-", ":", etc. used in English expressions. An example of this is a description that contains only characters that are In addition, another example of a description that corresponds to the predetermined appropriate parenthesized character string pattern is a character string such as "for example," "that is," "that is," "so to speak," or "in other words" in parentheses. Examples include descriptions that begin with a character string "hereinafter referred to as..." in parentheses.

For example, the warning information generation unit 2059 detects, in the specification of the invention document acquired by the invention document acquisition unit 2021, a description that does not satisfy the condition that words with the same meaning are consistently written in the same notation. If it is detected that words with the same meaning are inconsistently written, and the content of each description related to the detected inconsistent notations, the warning information shall include. let For example, the warning information generation unit 2059 may define a predetermined inconsistent notation list that associates the words with different notations for various words that may be expressed inconsistently, and Based on the unified notation list, it is possible to search for words with the same meaning mixed in different notations in the specification of the invention document. The setting information includes a predetermined inconsistent notation list, and the warning information generation unit 2059 refers to the predetermined inconsistent notation list in the setting information acquired by the setting information acquisition unit 201 to generate an invention. It may also be possible to search for words with the same meaning mixed in different notations in the specification of a document. The predetermined inconsistent spelling list is, for example, a list listing a plurality of spellings that have the same meaning and the same pronunciation for various words. Examples of the predetermined list of inconsistent expressions include, for example, associating "that is" with "that is", associating "of course" with "of course", and associating "or" with "or". Here is a list. Note that although it is not necessarily necessary to unify the notation of words with the same meaning in invention documents, this pointing information can be useful, for example, as a warning to those who wish to improve the formality of invention documents.

For example, when any section of the claim of the invention document acquired by the invention document acquisition section 2021 includes a statement that cites multiple sections, the warning information generation section 2059 includes each section of the cited section in the statement. If it is detected that there is a statement that does not satisfy the condition that the terms be expressed in a format that cites them in an alternative manner, a statement indicating that multiple sections are not cited in an alternative manner and the The warning information includes indication information indicating the number of the detected section and the content of the description related to the citation. Examples of statements that do not meet this condition include "or", "any", etc., such as "the control device according to claims 1 and 2", "the control device according to claims 1 to 3", etc. This is a statement that cites multiple items without using alternative words. Note that there may be cases where it is not necessary to alternatively cite multiple cited sections in each section of the claim of the invention document, but this pointing information is used as a cautionary note. Can be useful.

For example, the warning information generation unit 2059 generates a reference character string (for example, a reference character indicated by the reference character string pattern information in the condition information of the setting information) in any section of the claim of the invention document acquired by the invention document acquisition unit 2021. If a word or phrase follows a string that corresponds to the column pattern, the referenced word that is the same as that word appears in the part that precedes the reference string in that section or in the section that that section cites. If it is detected that there is a description of a reference character string that does not meet the condition that it must be described in Warning information includes indication information indicating the item number of the section and the content of the description related to the reference character string. In addition, in the above conditions, "paragraphs cited by that paragraph" include paragraphs that are further cited by the paragraph that the paragraph cites, paragraphs that are further cited by the paragraph that is cited by the paragraph, etc. This applies to the citation section based on one or more chains of citations. The warning information generation unit 2059 can refer to the claim citation relationship information held by the claim citation relationship detection unit 2057 in order to determine the item cited by the claim item. For example, if a word or phrase is written after a reference character string in any section of a claim of an invention document, and the referenced word or phrase that is the same as that word or phrase precedes the reference character string in that section. If the section is not listed in the cited section, and there are multiple sections cited by that section, one or more chains of citations within or from each of the multiple sections cited. If the referenced phrase is not described in the quotation destination section based on obtain. Note that the indication information should include information on the cited item number as one or more paths consisting of each item in the chain of citations for which there was no previous description of the referenced phrase. It's okay. As a specific example, claim 1 states "sensor", claim 2 and claim 3, which do not cite claim 1, do not state "sensor", and claim 3 is In a case where claim 2 is cited, when "the sensor" is stated in claim 4 which selectively cites claims 1 and 3, but "sensor" is not stated in advance. For example, the warning information generation unit 2059 generates a message indicating that there is no description of the preceding referenced phrase, “4” indicating the detected term, and “the sensor” as the description related to the reference character string. , "2-3-4", which is the information of the cited item number as a chain path in which there was no description of the preceding referenced phrase, is included in the pointing information.

For example, the warning information generation unit 2059 generates a condition that, among the claims of the invention document acquired by the invention document acquisition unit 2021, words indicating a subordinate concept are not used for an independent claim that is a term that does not cite other terms. If it is detected that a statement that does not satisfy the following (in other words, a statement that uses a word or phrase indicating a subordinate concept) is included, the item number of the independent clause, that word, and the word or phrase indicating a subordinate concept are used. The warning information includes indication information indicating that the For example, the warning information generation unit 2059 uses a predetermined list of subordinate concept words that lists words indicating subordinate concepts to search for descriptions using words indicating subordinate concepts from the independent claims of the claims of the invention document. The setting information includes a predetermined list of subordinate concept words and phrases, and the warning information generation unit 2059 refers to the predetermined list of subordinate concept words in the setting information acquired by the setting information acquisition unit 201 to generate the invention. It is also possible to search for descriptions using words indicating subordinate concepts from the independent clauses of the claims of the document. Note that in the predetermined subordinate concept word/phrase list, a word/phrase indicating a superordinate concept of the word/phrase may be included in association with a word/phrase indicating a subordinate concept. In this case, when the warning information generation unit 2059 detects a description using a word or phrase indicating a subordinate concept in an independent claim of a claim of an invention document, the warning information generation unit 2059 may include, for example, a correction of the word or phrase indicating a subordinate concept in the indication information. As candidates, a predetermined list of lower-level concept words may be referred to, and words indicating a higher-level concept corresponding to the lower-level concept may be included. As an example of a predetermined subordinate concept word list, the subordinate concept terms ``display'', ``send'', ``print'', etc. are associated with the broader concept term ``output'', and the subordinate concept terms ``keyboard'', ``mouse'', etc. ”, “touch pad”, etc., and a list that associates the phrase “input device” as a general concept.

Note that when the analysis unit 205 of the invention document analysis system 20 analyzes an invention document, as a result of the hierarchical concept relationship determination processing performed by the inter-element relationship determination unit 2058, it is determined that there are Information indicating conceptual hierarchical relationships can be obtained. Therefore, based on the information indicating the conceptual hierarchical relationship between a plurality of important words obtained by the invention document analysis system 20 for each of one or more invention documents, the words of the lower concept and the words of the broader concept are used. It becomes possible to generate or update a predetermined list of subordinate concept words and phrases that are associated with each other. Using the predetermined subordinate concept phrase list generated or updated based on one or more invention documents acquired by the invention document acquisition unit 2021 in this way, the warning information generation unit 2059 It may be possible to search for whether an independent claim of a claim of an acquired invention document includes a description using a phrase indicating a subordinate concept. The warning information generation unit 2059 uses the invention document to be searched for a description using words indicating a subordinate concept in an independent claim of a claim of the invention document as a basis for generating a predetermined list of subordinate concept terms to be used for the search. It's okay. Note that a subordinate concept is only a relative one, and it is not necessarily inappropriate to use words that indicate a subordinate concept in the independent claim of an invention document; However, the warning information generating unit 2059 uses an appropriately defined predetermined list of subordinate concepts, so that the indication information related to the words and phrases indicating subordinate concepts in the warning information are not described in an unnecessarily limited manner. This may be useful as a warning to avoid including it in the independent claim of the claim of the invention document.

For example, the warning information generation unit 2059 extracts important words extracted by the important word extraction unit 2051 from the claims of the invention document acquired by the invention document acquisition unit 2021, regardless of their importance, if they are described in the specification. If it is detected that the description in the specification of the invention document does not satisfy the above conditions, the warning information includes the important word and phrase and indication information indicating that the important word is not stated in the description. It may also be possible to do so. For example, instead of detecting that the description in the specification of the invention document does not satisfy this condition, the warning information generation unit 2059 converts this condition into "Details for Carrying Out the Invention" in the specification of the invention document. , and that the description of the content of "Example" or "Example n" (n is a natural number) is not satisfied. As a specific example, the warning information generation unit 2059 generates a description of each important term extracted by the important term extracting unit 2051 from the claims of the invention document in the specification of the invention document or the “Description of Carrying Out the Invention” of the specification. ”, “Example”, etc., and if the important phrase cannot be detected, it is assumed that this condition is not satisfied, and indication information indicating the important phrase etc. is included in the warning information.

For example, the warning information generation unit 2059 detects that the elements represented by each of the plurality of important terms extracted by the important term extraction unit 2051 from the invention document acquired by the invention document acquisition unit 2021 have a structural inclusion relationship. In the case determined by the inter-element relationship determination unit 2058, if one of the elements in a compositional inclusive relationship is an element of a thing, the other element is also an element of a thing, and one of the elements is an element of a method. If it is detected that the other element does not satisfy the condition that the other element is also a method element, each important phrase indicating the elements and the elements are consistent in the classification of things and methods. The warning information includes indication information indicating that the warning information will not be used. The warning information generation unit 2059 determines whether the element expressed by a certain important phrase is a method element, for example, based on whether the important phrase has a word listed in a predetermined method ending list as a ending. The predetermined method ending list is, for example, a list listing "method", "step", "procedure", etc. In the determination process related to this indication information, the warning information generation unit 2059 may treat the element as an object element if it determines that the element expressed by a certain important word is not a method element. . The setting information includes a predetermined method ending list, and the warning information generation unit 2059 refers to the predetermined method ending list in the setting information acquired by the setting information acquisition unit 201 to identify important words. It may also be determined whether the represented element is a method element.

For example, the warning information generation unit 2059 generates a description that does not satisfy the condition that the invention document acquired by the invention document acquisition unit 2021 does not contain characters with a character code other than the character code that can be used for patent applications in Japan. When it is detected that the text contains an inappropriate character, the warning information includes indication information indicating that the character is inappropriate in the description and that the character is not represented by a usable character code. The character code that can be used for patent applications in Japan is the Shift_JIS code based on JIS-X0208-1997 "Kanji code system for information exchange". Note that if a predetermined control code (for example, "0x07" for bell control, etc.) is included in the inappropriate characters in the description that does not meet the conditions, the warning information generation unit 2059 generates the predetermined control code. Control by the control code may be invalidated by converting the code to a predetermined character (for example, "●", etc.), and the inappropriate character after conversion may be included in the indication information. For example, when the character code of an inappropriate character corresponds to a circled number, the warning information generation unit 2059 may explicitly include in the indication information that the circled number cannot be used.

For example, the warning information generation unit 2059 generates a statement in the claim of the invention document acquired by the invention document acquisition unit 2021 that does not satisfy the condition that the item number of the claim section is a continuous number of natural numbers in line order. When it is detected that the section includes the claim, the warning information includes the section number of the claim section in the statement and pointing information indicating that the section number of the section is inappropriate.

For example, the warning information generation unit 2059 applies a condition to the claim of the invention document acquired by the invention document acquisition unit 2021 that the section number of the cited section in the claim section is smaller than the section number of the cited section. If it is detected that a statement that does not satisfy the following is included, the section number of the section from which the claim is cited in that statement and the section number of the section to which that section is cited are indicated to the effect that the section number is inappropriate. The indication information is included in the warning information.

The warning information generation unit 2059 may, for example, insert double-byte blank characters or opening parentheses (for example, opening corner brackets, opening parentheses, etc.) at the beginning of lines other than comment lines in the invention document acquired by the invention document acquisition unit 2021. When it is detected that a line that does not satisfy the condition of being a character is included, the system sends indication information indicating the contents of the line and that the character at the beginning of the line may be inappropriate. Include in warning information. The comment line is specified, for example, based on the comment specifying information of the setting information acquired by the setting information acquisition unit 201. Note that although it is not necessary that there be a full-width space character or an opening parenthesis character at the beginning of a line in an invention document, this pointing information can be useful, for example, as a warning to someone who wants to improve the formality of an invention document. In addition, the warning information generation unit 2059 may, for example, if the invention document acquired by the invention document acquisition unit 2021 includes a period character “.” at the end of a line other than a comment line, the beginning of the line is a full-width space character. When it is detected that a line that does not satisfy the condition that the above condition is met is detected, warning information is sent to indicate the contents of the line and that the characters at the beginning of the line may be inappropriate. It may also be included in

For example, the warning information generation unit 2059 detects that the invention document acquired by the invention document acquisition unit 2021 includes a description that does not satisfy the condition that comma characters and full stop characters are not written consecutively. Sometimes, the warning information includes indication information indicating the contents of the description and the fact that the description is inappropriate. As a specific example, the warning information generation unit 2059 searches the invention document acquired by the invention document acquisition unit 2021 for a description in which two or more of either a comma character or a full stop character is consecutive, and searches for such a description. When detected, the indication information is included in the warning information. As a specific example, when the warning information generation unit 2059 detects a description in which one comma character and one full stop character are consecutive in an invention document, the warning information is included in the warning information. Can be mentioned.

For example, when the invention document acquired by the invention document acquisition unit 2021 includes an abstract, the warning information generation unit 2059 determines whether the number of characters in the abstract of the abstract does not exceed a predetermined upper limit number of characters. When it is detected that the summary contains a description that is not specified, the warning information includes indication information indicating that the number of characters in the summary exceeds a predetermined upper limit number of characters. This predetermined upper limit number of characters is, for example, 400 characters. As a specific example, the warning information generation unit 2059 counts the number of characters written from after the heading “Summary” to before the heading “Selection drawing” of the abstract in the invention document acquired by the invention document acquisition unit 2021. Then, if the counted number of characters exceeds a predetermined upper limit number of characters, the indication information is included in the warning information.

For example, when a plurality of paragraph numbers are described in the specification of an invention document acquired by the invention document acquisition unit 2021, the warning information generation unit 2059 generates a message that each paragraph number is expressed as a four-digit Arabic numeral. If it is detected that it contains a description of a paragraph number that does not satisfy the condition that it is a consecutive number starting from "0001", the detected paragraph number and the description of the paragraph number are Indication information indicating that the warning information is inappropriate is included in the warning information. As a specific example, the warning information generation unit 2059 may, for example, write numbers between the opening bracket and the closing bracket in the specification of the invention document acquired by the invention document acquisition unit 2021 in order from the beginning. When the paragraph number detected is not expressed as a 4-digit Arabic numeral, and when multiple paragraph numbers are not written in consecutive numbers starting from "0001". In this case, the indication information is included in the warning information. Note that, for example, if the number of paragraph numbers detected from the specification of the invention document acquired by the invention document acquisition unit 2021 exceeds 9999, the warning information generation unit 2059 generates a 4-digit number for each of the multiple paragraph numbers. Instead of the condition that the numbers are expressed in Arabic numerals and are consecutive numbers starting from "0001", multiple paragraph numbers are each expressed in 5-digit Arabic numerals and starting from "00001". Using the condition that the starting serial number is used, when a description that includes a description of a paragraph number that does not satisfy the condition is detected, indication information is included in the warning information.

For example, in the invention document acquired by the invention document acquisition unit 2021, the warning information generation unit 2059 continuously displays each of the formula headings, chemical formula headings, and table headings for each document such as claims, specifications, etc. When it is detected that a heading that does not meet the condition of being a numbered description is included, the detected heading and the corresponding heading among mathematical formulas, chemical formulas, and tables are inappropriately described. The warning information includes indication information indicating that. Here, the header of the formula is a description in which a number is written between the character "number" following the opening corner bracket and the closing corner bracket. The header of a chemical formula is a description in which a number is written between the opening corner bracket, the character "C" following it, and the closing corner bracket. In addition, the table heading is a description in which a number is written between the characters "table" following the opening corner bracket and the closing corner bracket. The numbers included in each heading of a mathematical formula, chemical formula, or table may be a number consisting of one or more Arabic numerals, or may be supplemented with alphanumeric characters, the minus sign "-", or the period character. It may be a number to which a branch number expressed by ".", opening parenthesis, and closing parenthesis, singly or in combination, is added. An example of a heading for multiple mathematical formulas with consecutive numbers is to write the mathematical formulas in the order of "Math 1", "Math 2", "Math 3-1", "Math 3-2", etc. for each document. Examples include headings. As a specific example, the warning information generation unit 2059 may, for example, sequentially from the beginning of each document such as a specification or a claim of an invention document acquired by the invention document acquisition unit 2021, write a “number” following the opening corner bracket. Search for a description with a number between the closing corner bracket as a formula heading, and if the multiple formula headings detected for each document are not descriptions with consecutive numbers added, Include the indication information in the warning information.

For example, the warning information generation unit 2059 determines that the invention document acquired by the invention document acquisition unit 2021 does not satisfy the condition that appropriate drawing numbers are added to descriptions of drawing headings and descriptions regarding drawings. When it is detected that a description of a heading of a drawing or a description regarding a drawing is included, the detected description and indication information indicating that the description is inappropriate are included in the warning information. Here, the heading of the drawing is a description in a format in which a drawing number exists between the characters "Figure" following the opening corner bracket and the closing corner bracket. A description regarding a drawing is a description in which a drawing number exists between the character "Diagram" and a predetermined character string (for example, a character string "Reference", a character string of a particle, etc.). This predetermined character string may be fixedly determined in advance, for example, or may be determined in the configuration information acquired by the configuration information acquisition unit 201. A proper drawing number following the letters ``Diagram'' must consist of one or more Arabic numerals, or be preceded by an alphanumeric character, a minus sign ``-'', a period character ``.'', or a leading dot. This is a number to which a branch number expressed as a single or combination of parentheses and closing parentheses is added. Drawing numbers other than appropriate drawing numbers are treated as inappropriate drawing numbers. As a specific example, the warning information generation unit 2059 searches the invention document acquired by the invention document acquisition unit 2021 for descriptions of drawing headings and drawings, and detects descriptions of drawing headings or drawings. If the description includes an inappropriate drawing number, indication information is included in the warning information.

For example, the warning information generation unit 2059 generates a heading “Patent document n”, “Non-patent document n”, or “Example n” to which a number “n” is added to the invention document acquired by the invention document acquisition unit 2021. ”, if it is detected that these headings contain descriptions that do not meet the condition that appropriate numbers are added to these headings, then the The warning information includes indication information indicating that the description and the number in the description of the heading are inappropriate. Appropriate numbers are written in Arabic numerals representing natural numbers, and are written in consecutive numbers without branch numbers in the order of description in each of the headings "Patent Document n," "Non-Patent Document N," and "Example N." It is. As a specific example, the warning information generation unit 2059 generates, for example, “patent document n”, “non-patent document n” to which the number “n” is added, from the invention document acquired by the invention document acquisition unit 2021, or The description of "Example n" is searched, and if the detected description includes an inappropriate number, indication information is included in the warning information.

For example, the warning information generation unit 2059 detects the possibility that the invention document acquired by the invention document acquisition unit 2021 does not satisfy the condition that there are no typos in the important words extracted from the invention document by the important phrase extraction unit 2051. When it is detected that a description of a certain important word or phrase is included, the warning information includes the description of the detected important word or phrase and indication information indicating that there is a possibility that the important word or phrase is written incorrectly. As a specific example, the warning information generation unit 2059 may detect, for example, a case where two important terms extracted by the important term extraction unit 2051 from an invention document acquired by the invention document acquisition unit 2021 differ from each other by only one character. However, there is a possibility that there is a typo in either of the two important words, so in order to alert people, information indicating the two important words and the fact that there is a possibility of a typo in either of them is posted. Include in warning information. An example of two important phrases that differ by one letter from each other is that if one character of one of the two important phrases is deleted, one of the two important phrases and the other one of the two important phrases are deleted. These are the two important words and phrases that are the same as the important words. Another example of two important phrases that differ from each other by one character is to delete the nth (n is a natural number) character from the beginning of one of the two important phrases, and Two important words or phrases where, if the nth character from the beginning of the other important word is deleted, one of the important words after deletion is the same as the other important word after deletion. It is. The warning information generation unit 2059 may include, in the indication information, the number of descriptions of each of the two important terms in the invention document, the number of descriptions in the specification of the invention document, and the like. As a specific example, when an invention document includes two important words, "infrared sensor" and "infrared sensor," the warning information generation unit 2059 may include which of the two important words in the indication information. Here are some examples that indicate that there is a possibility of a typographical error. Another specific example is when an invention document contains two important terms: "lighting control program storage medium" and "lighting control program storage medium," or when an invention document contains the words "lighting control device". Here is an example in which the warning information generation unit 2059 indicates that there is a possibility of a typo in either of the two important words in the indication information when the two important words "certification control device" are included. , can be mentioned. Here, the warning information generation unit 2059 generates indication information indicating that there is a possibility of a typographical error when two important phrases extracted from the invention document are two important phrases that differ from each other by only one character. An example of what is included in warning information is shown. Modifying this example, the warning information generation unit 2059 determines whether two important words extracted from an invention document have a number of characters greater than a predetermined first number of characters (for example, five characters). When the important phrases are two important phrases that differ from each other by a predetermined second number of characters (for example, two characters), the warning information may include indication information indicating that there is a possibility of a typographical error.

The warning information generation unit 2059 may, for example, format the invention document acquired by the invention document acquisition unit 2021 in such a way that the descriptions of important terms expressing the same concept extracted by the important term extraction unit 2051 from the invention document are unified in form. When it is detected that there are multiple important words and phrases that do not satisfy the condition that the expression is Indication information indicating that there is a formal inconsistency is included in the warning information. In this case, the warning information generation unit 2059 includes, in the indication information, the number of statements in the invention document of each of the plurality of important words whose descriptions are inconsistent in form, the number of statements in the specification of the invention document, etc. It may also be possible to do so. As a specific example, the warning information generation unit 2059 may detect two important words extracted by the important word extraction unit 2051 from the invention document acquired by the invention document acquisition unit 2021, and the only difference is whether the two important words are half-width or full-width. If the two important words contain one or more characters expressing the same meaning, the notation of the two important words is not uniform in form, so the warning information includes pointing information to call attention to it. As a specific example, if an invention document contains two important phrases: "first sensor" that includes the full-width character "1" and "first sensor" that includes the half-width character "1". An example is an example in which the warning information generation unit 2059 indicates in the indication information that the notations of the two important words are formally inconsistent.

For example, the warning information generation unit 2059 generates information so as to match the relationship between the code and element name described in the “code explanation” column in the specification of the invention document acquired by the invention document acquisition unit 2021. If it is detected that a description that does not satisfy the condition that the code added to the element name is included, the detected element name and code are displayed in the "Code explanation" column. The warning information includes indication information indicating that the code corresponding to the element name does not match the code. As a specific example, the specification of the invention document includes the statement "first device 102", and furthermore, in the "explanation of symbols" column, "101 first device" and "102 second device" are included. , etc., the warning information generation unit 2059 sends a message to the indication information indicating that the code for "first device 102" in the specification does not match "101" in the "explanation of code". An example of this includes information such as:

For example, when the warning information generation unit 2059 includes indication information indicating an indication based on a predetermined condition on the invention document acquired by the invention document acquisition unit 2021, the warning information generating unit 2059 generates a description of the reason for the indication etc. Explanation that provides an explanation to the user, an explanation to the user about the disadvantage caused by the statement that has been pointed out, or an explanation to the user of how to appropriately correct the statement that has been pointed out. It may also contain a message. The explanation of the reason in this explanation message may include, for example, information such as legal provisions and judicial precedents that are the basis for the indication, or information on reference materials related to the indication. For example, this explanatory message may be fixedly defined for each condition that is the basis for pointing out the indicated information, or may be defined in the setting information acquired by the setting information acquisition unit 201. However, the warning information generation unit 2059 may generate the warning information according to a predetermined algorithm according to the conditions.

[3-1-5-6-3. Points regarding charges, etc.]
For example, if the number of claims of the invention document acquired by the invention document acquisition unit 2021 is a patent application based on the invention document filed in a prescribed region such as the United States or Europe, the warning information generation unit 2059 may be configured to When the fees required for patent applications, etc. exceed a predetermined number, which is a boundary at which the fees will increase relatively rapidly, the warning information includes pointing information to that effect. This predetermined number may be defined by the setting information acquired by the setting information acquisition unit 201, for example. The warning information generation unit 2059 may include information indicating a predetermined area where the application may be applied in the indication information. Further, the warning information generation unit 2059 may include indication information in the warning information for each of a plurality of predetermined areas.

Further, the warning information generation unit 2059 determines whether or not to include various types of indication information related to each predetermined region in the warning information, according to the designation in the setting information or the designation by the user of the invention document analysis system 20. , the indication information may be generated according to the determination. As a specific example, the warning information generation unit 2059 generates various types of indication information related to the United States only when “United States” is specified in the setting information, etc., and uses the generated indication information. It may also be included in the warning information.

[3-1-5-6-4. Information related to the indicated information included in the warning information]
The warning information generation unit 2059 generates a predetermined indication based on the above-mentioned conditions etc. when including one or more indication information in the warning information as part of the patent application related information output by the output unit 208. Based on level standards, the pointing level may be determined for each piece of pointing information according to the content pointed out by the pointing information, the conditions on which the pointing information is based, and the like. An example of a standard for a predetermined pointing level is a standard for determining a pointing level such that the more important the content pointed out by the corresponding pointing information is, the higher the pointing level is. Another example of the criteria for the predetermined level of indication is that the invention document that is the target of the indication information and the translation result document that is expected to be generated when the invention document is translated appropriately are based on the indication information. A criterion for determining the indication level is such that the higher the possibility of a mismatch in the units of corresponding words and phrases, the higher the indication level. An example of a relatively high possibility that the invention document and the translation result document will be inconsistent on a word-by-phrase basis is, for example, an example of indication information pointing out that there is a clerical error in the invention document. An example of this is, for example, indication information that points out the omission of a code to be added to an element in an invention document. Yet another example of the criteria for the predetermined level of indication is that the more likely it is that the description of the invention document that is the subject of the indication information will significantly deteriorate the quality of the document to enable appropriate translation, the higher the level of indication. There are criteria for determining the level to be high. An example of a statement in the invention document that is the subject of the pointed information that is likely to significantly reduce the quality of the document to enable appropriate translation is, for example, a referenced phrase in one of the claims of the invention document. Examples include cases in which a statement of (that is, a word or phrase that follows a reference character string) is not mentioned in the section that precedes the statement in the section, or in the section to which the section is cited, or in the section to which the section is chained.

The warning information generation unit 2059 may indicate the corresponding pointing level in the pointing information according to the judgment result of the pointing level, or may include the pointing information in the warning information in descending order of pointing level. good. For example, the warning information generation unit 2059 generates a case in which an issue is pointed out based on a condition corresponding to each condition item of the invention document standard provisions for translation acquired by the invention document standard provisions for translation acquisition unit 204; The level of indication can be determined in one or both of the cases where the indication is made based on conditions other than the standard provisions of invention documents for use.

For example, the warning information generating unit 2059 generates information based on the number of pointing information to be pointed out for the invention document acquired by the invention document acquiring unit 2021, or based on the number of pointing information and the pointing level of each pointing information. An evaluation value may be calculated using a predetermined arithmetic expression based on the above, and invention document evaluation information including the calculated evaluation value may be included in the warning information. The predetermined arithmetic expression is, for example, an expression in which the evaluation value is the sum of the pointing level values of each piece of pointing information in the warning information. In this case, the warning information generation unit 2059 generates a message based on the number of pointing information pointed out with respect to the claim statement of the invention document acquired by the invention document acquiring unit 2021, or based on the number of pointed out information. An evaluation value is calculated using a predetermined calculation formula based on the indication level of each indication information, and the invention document evaluation information including the calculated evaluation value related to the statement of the claim is included in the warning information. Good too. In addition, the warning information generation unit 2059 generates information based on, for example, the number of pointed out information regarding the description of the specification of the invention document acquired by the invention document acquisition unit 2021, or based on the number of pointed out information. An evaluation value is calculated using a predetermined calculation formula based on the indication level of each indication information, and invention document evaluation information including the calculated evaluation value related to the description of the specification is included in the warning information. Good too. For example, the warning information generation unit 2059 generates information based on the number of pointed out information for each invention document acquired by the invention document acquisition unit 2021 based on conditions corresponding to each condition item of the invention document standard provisions for translation. Or, based on the number of pointed out information and the pointed out level for each pointed out information, a predetermined calculation formula is used to determine the index, for example, to indicate the degree of compliance with the standard provisions for invention documents for translation. An evaluation value may be calculated, and invention document evaluation information including the calculated evaluation value may be included in the warning information.

The warning information generation unit 2059 uses any of the setting information among the plurality of setting information defined by the technical field, the department to which the inventor belongs, etc., acquired by the setting information acquisition unit 201, for example. When the warning information includes indication information pointed out for the invention document acquired by the invention document acquisition unit 2021 based on the conditions determined by the invention document acquisition unit 2021, the group to which the used setting information belongs is identified. The information may be included in the corresponding indication information.

For example, the warning information generation unit 2059 generates, in all or part of the indication information, line information (for example, line number) of the description part of the invention document that is the target of the corresponding indication, the claim number of the part, and , at least one of the paragraph numbers in the specification of that part. When the warning information generation unit 2059 includes the line number of the description part of the invention document that is the target of the indication in each indication information in the warning information, the output unit 208 generates patent application related information including the warning information. It is useful to output patent application related information by including a line number added to each line of the invention document.

[3-1-5-6-5. Generation of warning information part in display document]
In the invention document analysis system 20, when the output unit 208 outputs patent application related information including warning information as a display document (for example, an HTML document compliant with HTML5), the warning information generation unit 2059 outputs patent application related information including warning information as a display document (for example, an HTML document compliant with HTML5). Generates warning information included in patent application related information.

The warning information portion of the patent application related information as a display document generated by the warning information generation unit 2059 includes, for example, one or more display elements. For example, a browser may interpret a document for display and display each display element.

One or more display elements included in the warning information portion can display, for example, all or part of any indication information. The display elements included in the warning information portion may have a link attribute. The display elements included in the warning information section are created by the warning information generation unit 2059, for example, to link to the relevant description location of the description indicated in the indication information (for example, a description that is potentially inappropriate) (in other words, the browser Display control information may be added to cause the display element to function as a GUI element (to cause the display to transition to the description location indicated by the indication information in response to an operation on the display element). The relevant description location to be linked to is, for example, the relevant part (for example, the relevant line) of the invention document included in the patent application related information as a display document, or the claim information included in the patent application related information as a display document. This is the relevant part (for example, the relevant section), or the relevant part (for example, the relevant paragraph) in the specification information included in the patent application related information as a display document. The display control information is, for example, a script that is associated with an operation type (e.g., mouse click operation) in a tag added to a display element (e.g., an HTML tag compliant with HTML5) to transition the display to the corresponding description location. This is information that describes functions etc. In warning information as a display document, a specific example of a display element that displays one piece of pointing information pointing out inappropriate descriptions in a complaint is as follows: In an HTML document, the warning information and the complaint information are each <article>. When the warning information is written sandwiched between the <article> tag and the <article> tag, the content of the warning information is written sandwiched between the <span> tag and the </span> tag, for example, For example, the <span> tag may include the name of a function in JavaScript (registered trademark) that is activated in response to a mouse click operation or the like. This function is written, for example, in a script file specified in the HTML file of an HTML document that is loaded into the browser. This is a function to search for and display a description. The indication information that can be displayed in the display element in the warning information is the line number of the description part of the invention document that is the target of the corresponding indication, the claim number of that part, or the paragraph number of the specification of that part. , the warning information generation unit 2059 activates the line number, claim item number, or paragraph number in response to the operation when generating display control information added to the display element. It may also be used as an argument to a script function. By using this argument, the function can display inappropriate lines in the invention document, display inappropriate claim sections in claim information, display inappropriate paragraphs in specification information, etc. can be achieved relatively easily. Note that the invention document, claim information, or specification information in the patent application related information as a display document may be the one that corrects inappropriate descriptions corresponding to the indications in the indication information in the warning information.

For example, a display element that has a link attribute with indication information as its display content may display an <a> tag (that is, an anchor tag) for a link that specifies the relevant description location of the inappropriate description pointed out in the indication information as display control information. It may be included as

In the warning information part of the patent application related information as a display document, the relevant description location that is the link destination of the display element related to the indicated information is, for example, in the revised claim information included in the patent application related information as a display document. It may be the relevant part (for example, the relevant section), or the relevant part (for example, the relevant paragraph) in the revised specification information included in the patent application related information as a display document. The revised claim information in the patent application related information as a display document corresponds to the claims in the invention document, and may include modified inappropriate descriptions corresponding to the points pointed out in the warning information. . In addition, the revised specification information in the patent application related information as a display document is equivalent to the specification of the invention document, and inappropriate descriptions have been corrected in response to the points pointed out in the warning information. May contain content.

When generating warning information in a display document, the warning information generation unit 2059 determines an indication level corresponding to one or more indication information to be included in the warning information, and displays the indication information. Display control information for displaying the indication information in a display mode according to the determined indication level may be added to the display element for displaying the indication information. For example, when one or more points of warning information are displayed by the browser, the warning information generation unit 2059 may determine whether or not the point level of the pointed information exceeds a predetermined threshold for each point information. The display control information may be generated so as to selectively add and display the characters "Error" or "Warning" accordingly. As another example, when one or more points of warning information are displayed by the browser, the warning information generation unit 2059 displays each point of information in a color corresponding to the point level of the pointed information. The display control information may be generated as follows. As a specific example of making this color different, the warning information generation unit 2059 adds a class name according to the indication level to the <span> tag added before the indication information included in the HTML document, and For example, the font color, font background color, etc., as display attributes of a class are defined in a CSS file that is read into a browser as display attributes of the class so that they are different for each class name corresponding to each indication level.

When generating warning information in a display document, the warning information generation unit 2059 displays one or more pieces of indication information to be included in the warning information in a display element for displaying the indication information according to an operation. Display control information may be added that functions as a GUI for correcting the description pointed out by the pointing information and making the correction result part of the corrected claim information or the corrected specification information. . As a specific example, if two important words and phrases extracted from an invention document are two important words and phrases that differ from each other by only one character, the warning information generation unit 2059 may generate a typographical error in the warning information in the display document. Include a display element for pointed information indicating that there is a possibility that Display control information may be added to provide a function as a GUI to be reflected on the screen. As another specific example, the warning information generation unit 2059 may detect, for example, that two important words extracted from an invention document contain one or more characters that express the same meaning, but differ only in whether they are half-width or full-width. , the warning information in the display document includes a display element of indication information indicating that the notation is inconsistent, etc., and the display element selects one of the two important words, Display control information may be added that functions as a GUI that reflects the selection results in the revised claim information or revised specification information.

[3-1-5-7. Complaint information generation section]
The claim information generation unit 205a executes claim information generation processing to generate claim information included in the patent application related information output by the output unit 208 of the invention document analysis system 20. Claim information is information representing claims in an invention document.

In order to deal with the case where the invention document to be analyzed by the invention document analysis system 20 includes a useful claim but does not include a useful specification, the claim information generation unit 205a generates Japanese patent information. In order to support the efficient preparation of application documents or Japanese manuscripts for translation for patent applications outside Japan, in addition to claim information, various information contained in patent application related information ( For example, claims-based specification information, etc.) may be generated.

The claim information generation unit 205a can execute a claim-based drawing information generation process that generates claim-based drawing information included in the patent application related information output by the output unit 208 of the invention document analysis system 20, for example. Claim-based drawing information is information related to drawings that is generated based on claims in invention documents and can be used as a basis for generating drawings for patent applications.

The claim information generation unit 205a can execute a claim-based specification information generation process that generates claim-based specification information included in the patent application related information outputted by the invention document analysis system 20 by the output unit 208, for example. Claim-based specification information is information related to a specification that is generated based on the claims in an invention document and can be used as a basis for generating a specification for a patent application.

The claim information generation unit 205a may execute a claim-based summary information generation process that generates claim-based summary information included in the patent application related information output by the output unit 208 of the invention document analysis system 20, for example. Claim-based abstract information is information about an abstract that is generated based on the claims in an invention document and can be used as a basis for generating an abstract for a patent application.

For example, the claim information generation unit 205a performs a modified claim information generation process that generates modified claim information in which claims of an invention document that are included in the patent application related information output by the invention document analysis system 20 by the output unit 208 are modified. can be executed.

If the invention document acquired by the invention document acquisition unit 2021 does not include a claim, the claim information generation unit 205a performs claim information generation processing, claim-based drawing information generation processing, claim-based specification information generation processing, and claim-based Execution of summary information generation processing and revised claim information generation processing is omitted.

[3-1-5-7-1. Generation of complaint information]
The claim information generation unit 205a realizes claim information generation processing by generating claim information representing claims included in the invention document acquired by the invention document acquisition unit 2021. The claim information may be a partially modified version of the claim included in the invention document acquired by the invention document acquisition unit 2021. An example of this modification is when a claim includes a predetermined control code such as "0x07" for bell control (for example, a code within the range of "0x00" to "0x1f"). An example of this is converting it into a predetermined character such as "●".

In the invention document analysis system 20, when the output unit 208 outputs patent application related information including claim information as a display document (for example, an HTML document), the claim information generation unit 205a outputs patent application related information including claim information as a display document. Generate complaint information included in related information. FIG. 22 shows an example of claim information in patent application related information as a display document displayed on a browser. In addition, in FIG. 22, some contents of the complaint information are omitted for convenience.

The claim information portion of the patent application related information as a display document generated by the claim information generation unit 205a includes, for example, one or more display elements corresponding to the contents of each line of each section of a claim, If there is one or more important words extracted by the important word extraction unit 2051 in each line, it includes one or more display elements corresponding to each important word, and in each line of each term, there are corner brackets. If a heading such as "Claim n" (n is a natural number) or a comment is included, one or more display elements corresponding to the heading or comment are included. As a specific example, when complaint information is written between an <article> tag and a </article> tag in an HTML document, the complaint information generation unit 205a creates each line of each section of the complaint under the <article> tag. , write the display element corresponding to that line by sandwiching the contents of that line between the <span> tag and </span> tag, and write it under the <span> tag of the display element corresponding to that one line. , for each important phrase in the line, the display element corresponding to the important phrase is described by sandwiching the important phrase between <span> tag and </span> tag. For example, a browser may interpret a document for display and display each display element. The display element corresponding to each line of each item included in the claim information portion makes the contents of the corresponding line displayable by a browser, for example.

The claim information generation unit 205a controls the display mode of the heading, comment, or important word/phrase included in the claim information portion of the display document, for example, on the display element corresponding to the heading, comment, or important word/phrase. (For example, display control information for highlighting, displaying in color, display transition, etc.) is added.

An example of highlighting important words and phrases is, for example, displaying important words and phrases in bold. An example of highlighting important words with relatively high importance is, for example, displaying important words with relatively high importance in a font with a larger font size than important words with low importance. .

A specific example of generating display control information to be added to each of a plurality of important words and phrases in order to display each of the different important words in a different display mode (for example, displaying them in different colors) is as follows. When the HTML file of the HTML document has an important word or phrase sandwiched between <span> and </span> tags and described as a display element, the complaint information generation unit 205a generates a <span> tag related to the display element. For example, identification information (for example, an identification code including an identification symbol and a serial number) defined in the important word management information to distinguish the important word from other important words is given as the class name. In this example, the CSS file of the HTML document contains information that defines the color as the display format for the contents of the <span> tag for each class name, which is identification information for important words (in other words, information indicating the color added to important words). ), different important words and phrases are displayed in different colors. To give a specific example, important words and phrases that are different from each other in complaint information can be set to have a certain width (for example, the width of one character expressed as "1em" using the unit "em") using the border-left property. ) and a left border line painted in mutually different colors is displayed in a display mode (see FIG. 22). Note that, for example, the left border line may be expressed as a solid line, and for this purpose, the style of the border line should be set to solid. As a specific example, the class names corresponding to the display elements of a plurality of important words with different constituent character strings in a claim are defined as kwd1, kwd2, kwd3, etc., and the class with the class name kwd1 is , set "solid 1em #f00;" to the border-left property, "#666;" to the color property, "bold;" to the font-weight property, and set "solid 1em #f00;" to the border-left property in the class with class name kwd2. 1em #f0c;", the color property is set to "#666;", the font-weight property is set to "bold;", and in the class with the class name kwd3, the border-left property is set to "solid 1em #f66;", and the color property is set to "bold;". “#666;” and “bold;” are set in the font-weight property. In FIG. 22, the color (that is, display color) on the left border line with the width of one character is expressed by a pattern, and a plurality of different colors are expressed by a plurality of different patterns. Note that the method of expressing different important words by changing the color of the left border added to each important word is better than the method of expressing the important words by using different font colors or font background colors. This has the advantage that the readability of the important words themselves does not deteriorate regardless of whether the background color is dark (for example, black) or bright (for example, white). Also, since the left border line is not a character, it will be ignored when copying and pasting display contents that include important words, so copy and paste the display contents into a text editor, etc., and remove important words. It does not get in the way when editing included descriptions, etc. Note that, as illustrated in FIG. 22, an important word or phrase displayed in a display mode with a left border, such as the width of one character painted in the display color, may be Although it is possible to add a border line (that is, at least one of the right border line, top border line, and bottom border line), it is also useful to display it concisely to improve the visibility of important words. Yes, and for this purpose, the right border, top border, and bottom border may not be added to the important word. Note that adding any boundary line with zero width to an important term is the same as not adding that boundary line. Note that a monospaced font may be preferentially used as a font for displaying some or all of the characters of patent application related information (for example, claim information and specification information) as a display document; Then, the width of the left border line of the important word may be set to one character (for example, 1em is specified). As a result, if a monospaced font is available, the width of each character is aligned with the rectangle painted with the display color added as the left border of the important word, so the visibility of each character, each important word, etc. can be improved.

For each important term, the claim information generation unit 205a selects class names to be added to the tags constituting the display element of the important term in descending order of the first appearance line number of the important term (for example, the first appearance line number in a claim of an invention document). When using identification information that is an identification code that includes a specified serial number, information that determines the display color (for example, the color of the left border of the display element) for each class name related to the display element of important words in the CSS file, etc. For example, the difference in display color between multiple classes that have the same identification symbol in the identification code as a class name and whose serial number parts are consecutive is relatively different than between multiple classes whose serial number parts are not consecutive. The display color may be specified so that the size is sufficiently large. In this example, the serial numbers corresponding to important words may be determined in any order between important words with the same first appearance line number. Another example of information that determines the display color for each class name (for example, the color of the left border of the display element) for the display element of important words is, for example, the identification symbol in the identification code as the class name. All of the differences in the display color added to each important phrase in each set of two important phrases among the plurality of important phrases in the first group that are the same and whose serial number part is less than or equal to the first predetermined number (for example, 100). Two important words out of a plurality of important words of a mixed group in which a plurality of important words of the second group and a plurality of important words of the first group are mixed with a plurality of important words of the second group whose minimum value in the set has a serial number larger than a predetermined number. The display color may be specified so that the difference in the display color added to each important term in each set is greater than the minimum value for all the sets. Since the minimum value of the difference in display color for each important phrase with a relatively small first appearance line number (for example, each of the 100 kinds of important phrases in the first group) is relatively large, each important phrase can be It can be identified relatively quickly and correctly. This difference in display color can be, for example, a difference in hue components (for example, a difference in position when expressed on a hue wheel such as a Munsell ring), or a difference in each component of RGB values (that is, each of the red component, green component, and blue component). Differences etc. As a result, for example, when claim information in a display document is displayed by a browser, the display elements for each important term in each section near the beginning of the claim will be displayed in a different color than the different important terms written nearby. They will be displayed sufficiently different and can be easily identified. Furthermore, when claim information in a display document is displayed by a browser, the same important words in each section of the claim will be displayed in the same display color, making it easy to quickly recognize that they are the same important words. It could be possible. As a simple example of the relationship between the serial numbers determined in ascending order of the first line number of the important words and the display colors of the important words, the display colors of the four types of important words with serial numbers 1 to 4 are shown on the color wheel. The hues of 0°, 90°, 180°, and 270° are defined as colors on the color wheel, and the display colors of the four important words with subsequent serial numbers from 5 to 8 are determined as hues of 10°, 100°, and 270° on the color wheel, respectively. 190° and 280° colors, and the display colors of four types of important words with subsequent serial numbers from 9 to 12 are determined as hues of 20°, 110°, 200°, and 290° on the color wheel, respectively. The display colors of the four types of important words with serial numbers 13 to 16 are set as hues 30°, 120°, 210°, and 300° on the color wheel, and thereafter, the display colors of each of the 32 types of important words are determined in the same way. An example is given in which the In this example, among the four types of important words in the first group whose serial numbers are 4 or less, the minimum value of the difference in the display color added to each important word in each of the two sets of important words among all the sets. (that is, a hue difference of 90°) is the 8 types of important words of the mixed group, which is a mixture of the 4 types of important words of the second group whose serial number is greater than 4 and 8 or less, and the 4 types of important words of the first group. Among them, the display color for each important word is set so that the difference in the display color added to each important word in each of the two sets of important words is greater than the minimum value (i.e., hue difference of 10°) in all the sets. has been specified. In this example, the hue increment is set to 10 degrees for each group of four types of important words, but if the increment is set to 5 degrees, more types of important words can be expressed by adding different display colors. . Adding different display colors to important words with serial numbers up to a certain range based on display control information for each important word (for example, information on color properties for each class for each class name corresponding to the important word) It is possible to uniformly represent important words and phrases whose serial numbers exceed a certain range by adding the same display color.

The complaint information generation unit 205a adds an identification code, etc. for identifying important terms with different contents (i.e., character strings), for example, as a class name, to the <span> tag related to the display element of each important term in the complaint information. In addition to adding an id attribute. This id attribute includes, for example, an identification symbol that identifies the claim information part of each part of the patent application-related information, an identification code that identifies important words, and a serial number within important words that have the same content. For example, an identifier that is a combination of multiple important words and phrases with the same content written in claim information in the order of their description (in other words, an identifier that is set to identify one display element in an HTML document). be.

As another example, if there is a structural inclusion relationship between the elements represented by each of the plurality of important words, the display elements of each of the plurality of important words have similar hues. It is also possible to add display control information that determines the display color (for example, the color of the left border line of the display element) to be displayed. As a result, when complaint information in a display document is displayed by a browser, important words representing each of multiple elements that have a structural inclusion relationship will be displayed in somewhat similar display colors. It may be possible to quickly recognize important words that are related to each other.

The complaint information generation unit 205a generates a display element corresponding to the important word/phrase, for example, as a GUI for performing a display transition related to the important word/phrase (for example, displaying information with respect to the important word/phrase) in response to an operation on the browser. Display control information is added in order to add a function (that is, to make the display element an operation-compatible GUI element).

As a specific example of generating display control information that adds a GUI function corresponding to an operation to important words, the complaint information generation unit 205a adds a <span> tag and a </span> tag to an HTML file of an HTML document. When important words are written across the <span> tag, a function in JavaScript (registered trademark) that is associated with various operations (for example, onmouseover attribute, onclick attribute, etc.) and activated in response to the operation. An example is to describe the function name, etc. This function may be written, for example, in a script file specified in an HTML file that is loaded into the browser.

To give a specific example of a display element of one important term in the complaint information in the display document generated by the complaint information generation unit 205a, when the complaint information includes multiple important terms "equipment control unit", The display element corresponding to the fifth “equipment control unit” is, for example, “<span class=”c003” id=”claiminfo-c003-005” onclick=’wordProc(”claiminfo-c003-005”)’ title= "...(Information added display content)...">Device control unit</span>", etc. In this example, the identification code of the important phrase "equipment control unit" is "c003", and the identification code of the claim information "claiminfo" is used as the value of the id attribute of the <span> tag of the display element, and the identification of the important phrase A character string including the code "c003" and the number "005" in the order of description is defined. Furthermore, in this example, when a mouse click operation is performed on a display element, the wordProc function is called with the value of the id attribute as an argument. As a result, the wordProc function can, for example, specify the display element from the argument and refer to the string of important words related to the display element. Extract the word identification code "c003", the order number "005", etc., and specify that the id attribute of the first display element of the same important word is "claiminfo-c003-001". It is now possible to realize the process of displaying the first display element, for example, the process of replacing the display content of the display element of all important words "equipment control unit" in complaint information with the character string etc. input by the user. It becomes possible to realize the following.

For example, the complaint information generation unit 205a is configured such that a script function that is activated when a mouse-over operation is performed on a display element of an important term displayed in a browser displays information related to the important term (that is, an important phrase). It may also be determined that additional information related to the word or phrase is displayed. The information related to the important word/phrase may be, for example, all or all of the important word/phrase management information (see FIG. 18) held in the analysis unit 205 and the attribute information of the important word/phrase (see FIG. 19), which is one of the contents. This is some information. FIG. 23 shows an example of information related to important words and phrases displayed in response to a mouse-over operation on a display element of important words in claim information of patent application related information displayed on the browser. In the example of FIG. 23, when the mouse pointer moves over the important phrase "Part B," the importance of the important phrase and the important phrase are defined as information related to the important phrase "Part B" pointed by the mouse pointer. Add the item number of the claim section, the item number of the claim section that refers to the important word, the code associated with the important word in the invention document, etc. near the important word "Part B". is displayed.

For example, the complaint information generation unit 205a is configured such that a script function that is activated when a mouse click operation is performed on a display element of an important phrase displayed on a browser performs an important phrase update process that changes the important phrase, and an important phrase update process that changes the important phrase. Code insertion processing that inserts the code after the important word when no code is written after the key word and there is a code corresponding to the key word, and the key word from claim information, specification information, or other related invention documents. processing to enumerate lines containing important terms, processing to search for lines containing important terms from claim information, specification information, etc., search sites on networks such as the Internet (i.e., searches that can be accessed via a network) processing to search for web pages containing the important term, definitions of the important term, synonyms of the important term, related images of the important term, etc., or searching for the important term on translation sites, etc. on networks such as the Internet. It may be determined that a process of translating words or the like is performed. Search sites on the Internet, such as the Google (registered trademark) search site, can, for example, provide URL parameters to search for the definition of a search term, search for related images related to the term, or search for the term. There are some services that allow searching for such translated words and phrases. A search site that searches for character strings, images, and other information from accessible web pages and the like may be realized by a server on a LAN.

For example, the complaint information generation unit 205a may be configured such that a script function that is activated when a mouse click operation is performed on a display element of an important term displayed in a browser displays a menu of options, and selects an important phrase according to the selection operation. Word update processing, code insertion processing, important word inclusion enumeration processing, processing to search for lines containing important words from claim information, specification information, etc., web sites containing important words on search sites on networks such as the Internet, etc. Processing to search for a page or the like, processing to translate important words or phrases on a translation site on a network such as the Internet, etc. may be determined to be selectively executed. In order to display a menu that includes options for executing a process of searching for web pages, etc. that include the important term on a search site on a network such as the Internet, the complaint information generation unit 205a, for example, displays the menu to be generated. Include an <a> tag (that is, an anchor tag) for a link in the element, and set the href attribute of the <a> tag to the important word that can be obtained from the innerHTML property of the object that displays the important word that was clicked with the mouse. The function includes processing for setting a description following the URL of a search site on a network such as the Internet, using a character string containing the contents of the URL as a URL parameter. In its function, the complaint information generation unit 205a sets a URL parameter corresponding to the function in the href attribute following the URL of a search site that provides various functions such as image search and translation for the menu to be displayed. By writing each <a> tag, it is possible to include options for various functions such as image search and translation in the menu.

The complaint information generation unit 205a causes the processing results (for example, search results, translation results, etc.) by this function to be displayed in a range that does not hide the display element of the important word displayed on the browser and clicked with the mouse. , this function may be defined in advance. An example of a process of searching for a line containing one important term α from specification information within a JavaScript (registered trademark) function is to search for a line containing one important term α from specification information in an HTML document. The display element for each important term (e.g. <span> tag) includes an identification symbol to identify it as a specification information part of patent application related information, an identification code to identify each important term, and a serial number within the same important term. (In other words, an identifier defined to identify one display element in an HTML document) is added as an id attribute. For example, add the first number of the serial number, etc., and send the formed identifier as an argument to document. An example of specifying the display element of the important term α by executing the getElementById method, and displaying the line containing the specified display element in the browser display area by executing the scrollIntoView method on the specified display element, for example. can be mentioned. In addition, as an example of a specific method for realizing the enumeration process that includes important words, the contents of each line in which the important words are described in the complaint information are combined with the above-mentioned identification symbol, identification code, and serial number. Similarly, the contents of each line in which the important term is written in the statement information are collected by changing the serial number of the identifier, and each line that is the collection result is listed and the contents of each line are An example of this is emphasizing and displaying important words. Another example of a specific method for implementing the important word inclusion description enumeration process is to extract the related information based on other invention documents related to the invention document from which important words are extracted by the important word extraction unit 2051. The content of each line in which the important word is written in other invention documents is collected by searching for the important word, and each line that is the collection result is listed and the important word in each line is highlighted and displayed. Examples include: In this case, the claim information generation unit 205a adds identification information of other related invention documents (for example, the file name of the invention document, the application number of the patent application, the publication number, the patent number, the title of the invention, etc.) to the display content in this case. ) may be included. Other invention documents related to the invention document from which important words are extracted by the important word extraction unit 2051 are, for example, one or more invention documents specified by the user (for example, specification in the setting information). or information for identifying the patent document as prior art document information following the heading "Patent document n" (n is a natural number) in the specification of the invention document from which the important words were extracted. (For example, a publication number) may be the patent document. The invention document analysis system 20 can acquire the content of a patent document from a website via the Internet based on the publication number. In addition, for important words and phrases related to the operation target GUI element, which is a display element that has undergone an operation such as a mouse click operation, the important word inclusion description enumeration process searches, enumerates, and displays each line in which the important word is written. , may be partially modified, for example, in a sentence in which the important word is written as the subject (for example, following the important word or the symbol immediately after it, the case particle ``wa'' or the case particle compound ``wa'') It may be a process of searching for a sentence containing "" and listing and displaying each sentence collected as a search result. In addition, for example, the key word/phrase inclusive description enumeration process that enumerates the descriptions related to the key word/phrase that is the target of the operation may be performed by adding the line containing the key word/phrase to claim information, specification information, or another invention document (that is, the generation of claim information, etc.). Invention documents that are different from the original invention document) are listed and displayed, the important words and phrases in each listed and displayed line are highlighted, and the displayed line has instructions for operations on that line. Accordingly, it may be a process of adding display control information to make the line function as a GUI for displaying links to multiple lines including the line in claim information, specification information, or other invention documents. good. As a result, by further operating on one of the lines listed as including the operated important word/phrase, the user can further display the claim information surrounding that one line, Multiple lines in specification information or another invention document can be easily confirmed. Note that the important word inclusion description enumeration process that enumerates the descriptions related to the important words that are the target of the operation, for example, enumerates and displays the lines containing the important words from the claim information and specification information, and The important words and phrases in each displayed line are highlighted, and the displayed line is used as a GUI to display links to multiple lines including that line in claim information or specification information, depending on the operation on that line. It may also be a process of adding display control information for making that row function. Link display can transition display content to a link destination (for example, other information that is not currently displayed) in response to an operation on a display element (for example, a mouse click operation with the mouse pointer positioned on the display element). It is a display. Note that for each of the plurality of important words and phrases in the claim information in the display document, the lines in which the important words are written in the claim information and specification information can be specified in advance, so the important words and phrases can be included according to the operations for the important words. When the description enumeration process is executed, instead of searching for the line in which the important term is written in the claim information and specification information, all or part of the information corresponding to the search result is generated in advance. The execution speed of the important term inclusion description enumeration process may be increased by utilizing the generated information.

FIG. 24 shows an example of a pop-up screen (for example, a window displayed by pop-up) displayed in response to a mouse click operation on a display element of an important phrase in claim information of patent application-related information displayed on a browser. The pop-up screen illustrated in FIG. 24 is displayed in response to a mouse click operation on the important term "Part B", and includes a plurality of push buttons that are GUI elements for executing various searches regarding the important term "Part B". It also includes the execution result of the important word inclusion description enumeration process for the important word "Part B". A push button can be generated using, for example, the <button> tag, and the URL of the search site, URL parameters, etc., such as "https://...", are set in the href property of the location object as the onclick attribute in the <button> tag. By doing so, it may be possible to execute a search in response to a button press operation. In the execution result of the important word inclusion description enumeration process illustrated in FIG. 24, the important words "control program" and "microprocessor" that are not described in the claims of the invention document but are included in paragraph "0042" of the specification Each of the characters is highlighted by adding a character background color, which is expressed by enclosing it in a rectangular frame in the figure, without adding a left border line. As an example of a method of operating the invention document analysis system 20, for example, a user (for example, an inventor or a patent attorney) uses the invention document analysis system 20 by referring to the execution result of the important word inclusion description enumeration process corresponding to the important words. Depending on the need for future use, the important word/phrase setting dictionary may be edited so that the important word/phrase setting dictionary includes the important word/phrase and a sentence that appropriately explains the important word/phrase.

For example, the complaint information generation unit 205a temporarily or constantly divides the display screen of the browser into a plurality of display areas, and divides the display area of the complaint information and the display area of the processing result by the function activated in response to the operation. They may be displayed in parallel (see FIG. 17).

The claim information generation unit 205a searches for a character string having the same content as any of the important words extracted by the important word extraction unit 2051 from the contents of each line of each claim section of the invention document. may identify display elements of one or more important words in that line. For example, if a line in a certain claim section contains only the statement "When processing input data," the claim information generation unit 205a uses the important word extraction unit 2051 to select Search for the important word "processing" that has already been extracted using a method that excludes verbs from the extraction target from that line, and treat that "processing" as a display element of the important word as the content of that one line in complaint information. You can also use it as Conversely, since "processing" cannot be extracted from that line by excluding verbs from the extraction target, the complaint information generation unit 205a treats "processing" as a display element of important words as the content of that line. It may be said that there is no such thing.

For example, the claim information generation unit 205a generates a message for highlighting in a display element corresponding to a heading such as "Claim n" (n is a natural number) enclosed in corner brackets, which is included in the claim information part. Alternatively, display control information may be added in order to add a function as a GUI for performing display transition to a display element corresponding to the immediately preceding or following heading, for example, in response to an operation on the browser. As a specific example of generating display control information that adds the function of a GUI for display transition to the immediately following heading to a certain heading "Claim n" (n is a natural number), the claim information generating unit 205a generates an HTML document. If the heading "Claim n" (n is a natural number) is written between the <span> tag and </span> tag in the HTML file of For example, a function name, etc. of a function in JavaScript (registered trademark) that is activated in response to an operation is written in association with a function (attributes, etc.). This function may be written in a script file specified in the HTML file, for example. For example, the claim information generation unit 205a performs a process of searching and displaying a display element with the heading "Claim m" (m is a number 1 greater than a natural number n) enclosed in corner brackets from the content of the claim information. This function can be defined to be executed.

For example, the complaint information generation unit 205a converts the reference character string (for example, "said", "above", "that", "that", etc.) included in the complaint information part into a <span> tag and a </span> tag. As a display element sandwiched between tags, display for highlighting the reference character string (for example, displaying it in bold, underlining, displaying in a predetermined color such as red, etc.) on the display element. Control information may also be added. The display element related to the string with the <span> tag can be highlighted using, for example, the text-decoration property, font-weight property, border-left property, color property, background-color property, etc. as display attributes in CSS. This can be achieved using display control information that adjusts the value. For example, in complaint information in an HTML file, different class names are defined for various display elements such as reference character strings, comments, and headings, and in the CSS file specified in the HTML file, reference character strings, comments, By setting different property values for highlighting for each class name that corresponds to various display elements such as headings, it is possible to It becomes possible to change the display mode.

For example, when the inappropriate description indicated by the pointing information in the warning information generated by the warning information generating unit 2059 is included in the complaint information part, the complaint information generation unit 205a converts the inappropriate description into For example, as a display element sandwiched between a <span> tag and a </span> tag, display control information for highlighting inappropriate descriptions may be added to the display element.

The claim information generation unit 205a, which generates claim information based on the claims in the invention document that has been acquired by the invention document acquisition unit 2021 and that has been analyzed by the analysis unit 205, generates a drawing document (for example, a PDF file) corresponding to the invention document. can be acquired by the drawing document acquisition unit 2022, when the claim information includes an important phrase that is the same as a signed element name included in the invention document, the important phrase can be acquired from the drawing according to the operation. Display control information for making a display element with a GUI function for displaying a portion of a document that includes the code (for example, a page that includes the code) may be added to the display element of the important term. This display control information defines, for example, the function name of a script function that is activated with a code as an argument in response to an operation, or the function itself. For example, if the drawing document is the content of a searchable PDF file, or if a PDF file containing the drawing document as content can be converted to a searchable PDF file, the code can be extracted from the searchable PDF file. By searching, it is determined to be specific. The claim information generation unit 205a includes a function for searching a code from a drawing document, for example, a process of dividing a PDF file of a drawing document into pages, a process of searching for the code from each divided PDF file divided into pages, It may also include a description for converting the divided PDF file whose code has been detected as a search result into an image, displaying the converted image, etc. In each of these processes, for example, Python (registered Various existing libraries (for example, "PyPDF2", "pdfminer", etc. provided by third parties) may be used to manipulate PDFs available in JavaScript (trademark) or JavaScript (registered trademark). In order to realize the functions of the invention document analysis system 20, an invention document analysis processing program including a portion written in Python (registered trademark), for example, is run on a web server (for example, a local host web server on the same terminal device as the browser). It can function as a server), start a browser, and provide the browser with an HTML document related to patent application-related information for display. The part of the program written in Python (registered trademark) works in conjunction with a script written in JavaScript (registered trademark), which is a component of an HTML document that is interpreted, displayed, and executed by the browser. It is possible to share processing (for example, processing of PDF files, etc.). For example, the Python (registered trademark) library "Eel" provided by a third party makes it possible to link the Python (registered trademark) program part and the script part written in JavaScript (registered trademark) in an HTML document. It has the function of

For example, in response to an operation on an important word corresponding to a code, the claim information generation unit 205a displays a function related to display control information for displaying a portion of the drawing document that includes the code in the browser and performs the operation. Alternatively, it may be determined that the portion of the drawing document that includes the key word is displayed within a range where the display element of the key word or phrase is not hidden. For example, the complaint information generation unit 205a divides the display screen of the browser into a plurality of display areas, and displays the complaint information display area and the display area of a part of the drawing document etc. displayed in parallel in response to an operation. You may. For example, the claim information generation unit 205a uses display control information added to a display element of an important term corresponding to a code to cause a function to be activated in response to an operation on the important term to be activated in a drawing document in response to a further operation. It may be specified to perform a process of switching and displaying multiple parts including the code (for example, two or more pages including the code), or reduce all parts including the code in the drawing document as necessary. It may also be determined that a process is performed to display all the information at once. For example, the claim information generation unit 205a generates a function related to display control information for displaying a portion of the drawing document that includes the code in response to an operation on an important phrase corresponding to the code. The code may be highlighted (for example, if the drawing document is the content of a searchable PDF file, a background color may be added to the code as a marker).

In addition, in a display document, when display control information for realizing a GUI function for performing processing in response to an operation is added to a display element, the operation is a mouse-over operation on the display element, The operation is not limited to mouse click operations, etc. For example, in a function that is activated in response to a mouse click operation on a display element, a menu, pop-up screen, etc. may be displayed to prompt further operations, and the options shown in the menu, pop-up screen, etc. It is also possible to execute processing according to the selection operation.

The complaint information generation unit 205a generates full-width or half-width blank characters, tabs, line feeds, The character codes for return, form feed, or vertical tab may be removed, and the line feed at the end of the line may be represented by the HTML tag <br> for line feed.

[3-1-5-7-2. Generation of claim-based drawing information]
The claim information generation unit 205a realizes claim-based drawing information generation processing by generating claim-based drawing information based on the claims included in the invention document acquired by the invention document acquisition unit 2021. The claim-based drawing information is information indicating the drawing number of each drawing, important words and phrases to be written in each drawing, relationships between important words and phrases, and if a code corresponds to the important word, a code, etc. The claim-based drawing information does not need to be complete enough to be used as it is as a drawing for a patent application. The claim information generation unit 205a can generate drawing type information indicating the drawing type of each drawing in association with the drawing number of each drawing in the claim-based drawing information. The drawing types include, for example, a block diagram, a flowchart, a diagram showing the recorded contents of a recording medium, a diagram showing means, functions, etc. implemented by a program.

If the invention document acquired by the invention document acquisition unit 2021 includes a specification, the claim information generation unit 205a determines that there is a good possibility that drawing information (for example, a drawing document) has already been generated. Since this is assumed, the execution of the claim-based drawing information generation process may be omitted.

[3-1-5-7-2-1. Determination of sign]
When generating claim-based drawing information, the claim information generation unit 205a extracts codes from among the important words extracted by the important word extraction unit 2051 from the claims of the invention document acquired by the invention document acquisition unit 2021. For an important phrase that satisfies a predetermined condition to be determined, a code determination process is executed to determine a code to be associated with an element expressed by the important phrase.

The claim information generation unit 205a creates claim-based drawing information by associating the important words and phrases whose codes have been determined by the code determination process with the codes. The claim information generation unit 205a may omit the code determination process when the invention document acquired by the invention document acquisition unit 2021 includes a specification.

Among the important words extracted by the important word extraction unit 2051 from the claims of the invention document, the predetermined conditions related to the important words whose codes are to be determined in the code determination process are, for example, the category type ( For example, the condition is related to the category type (category type classified as one of programs, data, objects, and methods). For example, an important phrase classified into the program or data category type does not satisfy a predetermined condition for which a code is to be determined, and the complaint information generation unit 205a does not determine a code corresponding to the important phrase. This is because the important words and phrases classified into the program or data category type can usually be understood by themselves. The category type of an important word or phrase, program, data, or method, can be determined by, for example, whether the important word or phrase includes a character string predetermined for each category at the end of the word. An important term that does not fall under any of these can be identified as a category type of object.

For example, if the elements represented by a plurality of important words identified as a method or product category type have a structural inclusion relationship, the plurality of important words satisfy a predetermined condition to be determined as a sign, and a claim is made. The information generation unit 205a determines the code corresponding to each of the plurality of important words. A configuration diagram showing the elements represented by each of the plurality of important words or phrases usually makes it possible to understand each element. Note that in the configuration diagram, it is useful that each element is labeled with a code. Whether or not elements expressed by multiple important words have a compositional inclusion relationship is determined by determining whether the elements represented by the important words and other important words are one attribute of each important word, which is specified by the attribute specifying unit 2056. This can be achieved by referring to the "compositional inclusion relationship with the represented element".

The claim information generation unit 205a determines a code based on the structural inclusion relationship of elements expressed by each important phrase that satisfies a predetermined condition for which the code is to be determined.

As a specific example, an element that is structurally included in the element expressed by any other important word or phrase, but is not structurally included in the element expressed by any other important word or phrase, is For important words and phrases, a code expressed by a code character that can be used as a code is determined. Then, the claim information generation unit 205a adds a code character to the end of the code for the first important word when the element represented by the first important word includes the element represented by the second important word. is determined as the code for the second important term. In this example, when the complaint information generation unit 205a determines the code corresponding to the first important phrase to be, for example, “10”, it determines the code corresponding to the second important phrase to be, for example, “101”.

As another specific example, an element that is structurally included in the element expressed by any other important word or phrase, but is not structurally included in the element expressed by any other important word or phrase, is For one important term, a code indicated by a string of Arabic numerals is tentatively determined. Then, the claim information generation unit 205a generates a response to the first important phrase when the element represented by the first important phrase includes the element represented by each of n important phrases (n is a natural number). The code is finally determined by adding zeros "0" equal to the number of k digits (k is a natural number) required to represent n to the end of the temporarily determined code. In other words, in a code determination process for determining a sign for an important phrase, the claim information generation unit 205a determines a sign for an important phrase that represents an element that structurally includes an element represented by another important phrase. can determine its sign depending on the number of elements (for example, n) that are represented by other important words and phrases that are structurally included in the element represented by the important word. The claim information generation unit 205a adds a character string corresponding to 1 to n expressed in k-digit Arabic numerals to the code tentatively determined for the first important phrase as a code for each of the n important phrases. Determine the sign formed by In this example, the claim information generation unit 205a tentatively determines the code corresponding to the first important word to be "1", and the element represented by the first important word is, for example, the element represented by each of the 15 important words. If the structure includes 15 words, add 0's for the number of 2 digits necessary to represent the 15 words to the temporarily determined code 1 corresponding to the first important word. Finally, the code "100" corresponding to the first important word/phrase is determined. Subsequently, the claim information generation unit 205a determines each code of "101" to "115" corresponding to the 15 important words representing the 15 elements included in the element represented by the first important word. do. In the above example, the first key word phrase that represents an element that structurally includes the element expressed by any other key word or phrase, but is not structurally included in the element expressed by any other key word or phrase. There may be more than one important word/phrase applicable to .

The claim information generation unit 205a may also execute the code determination process when the invention document acquired by the invention document acquisition unit 2021 includes a specification. If a code already corresponds to an important phrase that satisfies a predetermined condition for which a code is to be determined, the claim information generation unit 205a determines a code by using the corresponding code. For example, when determining a code in a case where a code already corresponds to an important term, the claim information generation unit 205a replaces the code written in the "code explanation" column of the specification of the invention document with the code that already corresponds to the important term. The code may be determined by preferentially adopting the code described in other parts. Note that in the "Explanation of Code" column of the specification of the invention document, if multiple codes are written in association with the same important term, the claim information generation unit 205a selects the first code listed. Alternatively, the code may be determined by selectively adopting one of the codes that comes first by sorting, such as natural ordering. The invention document analysis system 20 uses, for example, the standard library "sort" for sorting in Python (registered trademark), or "natsort", a library provided by a third party for natural order sorting. It is also possible to use this to perform sorting. Whether or not a code already corresponds to an important term can be determined based on the “code” as an attribute of the important term specified by the attribute specifying unit 2056. The claim information generation unit 205a determines that a code already corresponds to one or more of the important words and phrases that satisfy a predetermined condition for which a code is to be determined, and that one or more of the important words that satisfy the predetermined condition are , if the codes do not yet correspond, the claim information generation unit 205a determines a code corresponding to the important phrase for which the code does not yet correspond, so that the code does not overlap with the already corresponding code. or in order to make it possible to identify that the code is an automatically determined code, a predetermined character string (e.g. "auto", "at" ” etc.) may be added.

[3-1-5-7-2-2. Determination of drawing contents]
The claim information generation unit 205a determines drawing content as a claim-based drawing information generation process.

Specifically, as a claim-based drawing information generation process, the claim information generation unit 205a, for example, generates a plurality of important words extracted by the important word extraction unit 2051 from the claims of the invention document acquired by the invention document acquisition unit 2021. When there is a structural inclusion relationship between the represented elements, a configuration diagram is generated based on the inclusion relationship. An important phrase representing each element that has a structural inclusion relationship is an attribute identified by the attribute specifying unit 2056 based on the determination result of the inter-element relationship determining unit 2058. This can be determined by the item ``Constitutional inclusion relationship with ``.

For example, in a configuration diagram, an important word or phrase representing the included element is enclosed, and the important word or phrase representing the included element is placed outside a frame line (e.g., a rectangular frame) in which the symbols corresponding to the important word are connected with leader lines. is placed to surround the important phrase on the containing side and the frame line of the element on the included side, and the code corresponding to the important phrase on the containing side (for example, the code determined in the code determination process) is drawn with a leader line. It is a diagram depicting a connected frame line. If the code corresponding to an important phrase included in the configuration diagram is not determined, the code and leader line are omitted. When the inclusion relationship has a nested structure, frames related to a plurality of important words and phrases are nested. The configuration diagram may be, for example, an unfinished drawing composed of a plurality of character strings arranged at relatively appropriate positions, omitting the above-mentioned frame lines and leader lines.
The claim information generation unit 205a can generate the above-described configuration diagram as a drawing in the claim-based drawing information that is included in the patent application related information and to be displayed in the display document.

To give a specific example of a configuration diagram using important words a, b, c, d, and e that represent each element related to a thing, the element represented by the important word a is the important word b. If the important word c and the important word d each contain elements expressed in parallel, and the element expressed by the important word d includes the element expressed by the important word e, then the important word a is placed in a frame in the drawing. In the rectangular frame described in , the rectangular frames in which important word b, c, and d are written without enclosing important word a are arranged in parallel so that they do not surround each other. A rectangular frame in which the important word d is written in the frame is placed, and a rectangular frame in which the important word e is written in the frame without surrounding the important word d is placed, and each rectangular frame has a , if there is a code defined corresponding to the important phrase written in the frame, that code is written connected to the rectangular frame by a leader line.

In addition, if the "first line number" as an attribute of the important term is specified to represent the number of the first line stated in the claim of the invention document, Each of the elements may be listed in order from the top to the bottom of the drawing in descending order of "first appearance line number" as an attribute of the important word representing the element. As a result, for example, when a claim section is described in a component enumerated type, the claim information generation unit 205a generates a configuration diagram corresponding to the enumeration order, so that the generated configuration diagram can be understood. can become easy.

In cases where it is assumed that each section of a claim relating to a method in an invention document is stated in an enumerated format with the components arranged in order of operation for ease of understanding, the claim When there is an inclusive relationship between elements expressed by a plurality of important terms related to a method, the claim information generation unit 205a may generate a flowchart instead of a configuration diagram. As a specific example, if the important term "equipment control method" has an inclusive relationship that includes the important term "sensing procedure", the important term "operation pattern determination procedure", and the important term "device operation control procedure", If the "first appearance line number" as an attribute of important terms is "sensing procedure", "operation pattern determination procedure", and "equipment operation control procedure" in descending order, "equipment control processing" is included in the drawing. Add the character string "Start" and write a rectangular frame with rounded corners written inside the frame, place a rectangular frame with "Sensing procedure" written inside the frame below the rounded rectangular frame, and below that Place a rectangular frame with "Operation pattern determination procedure" written in the frame, place a rectangular frame with "Equipment operation control procedure" written in the frame below it, and below that, place "Operation control processing". A flowchart is drawn by arranging rectangular frames with rounded corners in which the character string "end" is written, and connecting each frame with line segments from top to bottom.

As a claim-based drawing information generation process, the claim information generation unit 205a, for example, generates information between elements expressed by a plurality of important terms extracted by the important term extraction unit 2051 from the claims of the invention document acquired by the invention document acquisition unit 2021. If there is a relationship between a recording medium and the contents of the recording medium, a diagram representing the recorded contents of the recording medium is generated based on the relationship. Important words and phrases representing each element in the relationship between a recording medium and the contents of the recording medium are identified by the attribute identifying unit 2056 based on the determination result of the inter-element relationship determining unit 2058. It can be determined by the item "Contents". Even if there is a relationship between a recording medium and the content of the recording medium between the elements represented by the plurality of important phrases extracted from the claims of the invention document by the important phrase extraction section 2051, the claim information generation section 205a When the contents of the recording medium are only elements classified into one program, generation of a configuration diagram can be omitted.

For example, a diagram showing the recorded contents of a recording medium may be drawn by enclosing important words representing the elements of the recording medium, and inside a frame line (for example, a rectangular frame) in which the codes corresponding to the important words are connected by leader lines. It is a diagram depicting one or more important words and phrases representing data, information, etc. corresponding to the recorded content of an element, each surrounded by a minimum frame line (for example, a rectangular frame with rounded corners). Note that the drawing representing the recorded content of the recording medium may be an unfinished drawing in which the above-mentioned frame lines and leader lines are omitted.

As a claim-based drawing information generation process, the claim information generation unit 205a, for example, generates information between elements expressed by a plurality of important terms extracted by the important term extraction unit 2051 from the claims of the invention document acquired by the invention document acquisition unit 2021. When there is a relationship between a program and its implementation contents, a diagram representing the means, functions, etc. realized by the program is generated based on that relationship. Important words and phrases representing each element in the relationship between a program and its realization content are determined by the “realization content of the element represented by the important words” of the attribute identified by the attribute identification unit 2056 based on the determination result of the inter-element relationship determination unit 2058. Can be distinguished by item.

[3-1-5-7-2-3. Determination of drawing number]
The claim information generation unit 205a determines a drawing number for each determined drawing content as a claim-based drawing information generation process.

As a specific example, the claim information generation unit 205a generates, for each drawing such as a determined configuration diagram, the key word with the smallest "first claim number" as an attribute among the important words representing elements included in the drawing content. The "first appearing claim number" is specified as the "first appearing claim number" of the drawing. Then, the claim information generation unit 205a determines the drawing number of each drawing in the descending order of the "initial claim number" specified for each drawing so that it becomes a consecutive number starting with 1. If the "first claim number" of multiple drawings is the same, the claim information generation unit 205a may, for example, compare the drawings with the same "first claim number" based on a predetermined drawing priority order. Determine the order and determine the drawing number. Examples of the predetermined drawing priority order include configuration diagrams related to objects, configuration diagrams related to data, diagrams representing recorded contents of recording media, configuration diagrams related to methods, and diagrams representing means and functions realized by programs. The order of the figures is listed.

[3-1-5-7-2-4. Generation of claim-based drawing information part in display document]
In the invention document analysis system 20, when the output unit 208 outputs patent application related information including claim-based drawing information as a display document (for example, an HTML document), the claim information generation unit 205a outputs patent application related information including claim-based drawing information as a display document (for example, an HTML document). Generate claim-based drawing information included in patent application related information.

The claim-based drawing information portion of the patent application related information as a display document generated by the claim information generation unit 205a includes, for example, the drawing number of each drawing, important words and phrases to be written in each drawing, and relationships between important words. If a code corresponds to an important word or phrase, it includes one or more display elements corresponding to the code, etc. As a specific example, when the claim information generation unit 205a writes claim-based drawing information in an HTML document by sandwiching it between an <article> tag and a </article> tag, the claim information generation unit 205a may write a drawing number under the <article> tag. For each drawing in descending order, describe a display element that indicates the drawing number of that drawing by sandwiching the drawing number between <span> and </span> tags, and then describe each display element that corresponds to that drawing. However, the codes are written between, for example, a <span> tag and a </span> tag that include a class name for identifying each code. For example, a browser may interpret a document for display and display each display element. The claim information generation unit 205a controls the display mode of the display element (e.g., highlighted display, colored display, etc.) on the display element corresponding to the drawing number, code, or important phrase included in the claim-based drawing information part. It is possible to add display control information for

For example, when generating claim-based drawing information in patent application related information as a display document as an HTML document compliant with HTML5, the claim information generation unit 205a generates a <canvas> file with a predetermined id attribute added to the HTML file. Including the tag, use the function etc. to obtain the canvas object using the getElementById method that specifies the value of its id attribute, obtain the drawing context using the canvas object's getContext method, and use the drawing context's fillText method, strokeRect method, and quadraticCurveTo method. By defining processes for drawing character strings, rectangular frames, leader lines, etc., it is possible to draw a configuration diagram.

[3-1-5-7-3. Generation of claim-based statement information]
The claim information generation unit 205a realizes claim-based specification information generation processing by generating claim-based specification information based on the claims included in the invention document acquired by the invention document acquisition unit 2021. The claim-based specification information represents the content of an unfinished specification, and does not need to be complete to the extent that it can be used as a specification for a patent application. Claim-based specification information may be used to efficiently create an appropriate specification in cases where the specification in an invention document is incomplete.

Claim-based specification information includes, for example, headings similar to the specifications of Japanese patent applications (e.g., "Title of the Invention," "Problem to be Solved by the Invention," "Mode for Carrying Out the Invention," etc.) ), and the content that follows each heading.

The claim information generating unit 205a uses the important terms extracted by the important term extracting unit 2051 from the claims included in the invention document acquired by the invention document acquiring unit 2021 to generate information on each claim in the invention document (for example, each claim). ) from which all claimable origin terms are extracted. A claimed starting claim is a claim that does not cite any other claim, or a claim that cites another claim but has a different claimed subject matter than the cited claim.

The claim information generation unit 205a specifies, for each extracted claimable starting point, an important phrase that represents the claimable object of the claimed starting point. The important words and phrases extracted by the important words and phrases extracting unit 2051 have one attribute specified by the attribute specifying unit 2056 as “whether or not it represents an element that is the subject of a claim,” so the complaint information generating unit 205a With reference to this attribute, etc., important words and phrases representing the claim subject of the claim subject starting point are identified. If a plurality of identical key words exist as elements of the set of key words of all the specified claim targets, the claim information generation unit 205a removes all but one key word from the plurality of key words. . Then, the claim information generation unit 205a selects the content to be written in the "Title of the Invention" column of the claim-based specification information from the remaining set of important terms to be claimed (that is, a set of important terms to be claimed that are different from each other). Determine the invention name string. For example, when there is only one important phrase as an element of a set of important phrases of mutually different claims, the invention name character string is a character string of that one important phrase. For example, if there are multiple key words as elements of a set of key words that are the subject of a claim that are different from each other, the invention title character string should be written in a predetermined format by listing each of the key words in a predetermined order. This is the placed string. An example of the predetermined order related to this enumeration is the natural number of the first appearing claim number specified by the attribute specifying unit 2056 as an attribute of the important phrase among each important phrase in a set of mutually different important phrases to be claimed. Among these, the smallest order is listed. According to an example of this predetermined order of enumeration, an invention name character string is determined in which the important words and phrases of the claimed subject are listed in accordance with the order in which they are stated in the claims. An example of a predetermined format for arrangement is a format in which important words other than the last are connected with a comma character "," and the last important words are connected with "and". As a specific example, if the key terms in a set of key terms for mutually different claims, arranged in order of first claim number, are "lighting control device," "lighting control method," and "lighting control program." In the claim-based specification information, the invention name character string arranged in a predetermined format to follow the "title of the invention" is, for example, "lighting control device, lighting control method, and lighting control program."

The claim information generation unit 205a uses, for example, a set of important terms of mutually different claim subjects determined based on important terms expressing the claimed subject for each claim originating claim, and generates an ``invention'' in the claim-based specification information. Determine the content to be written in the column “Problems the company is trying to solve.” The claim information generating unit 205a uses each important phrase in a set of mutually different important phrases to be claimed and a predetermined character string specified corresponding to each important phrase to generate a predetermined format in a predetermined format. generates one or more sentences connected with each other, and determines the generated sentence as the written content that follows the heading "Problem to be Solved by the Invention". The content of this predetermined character string is, for example, an incomplete message for writing support, which is indicated by the incomplete message information included in the setting information acquired by the setting information acquisition unit 201. Note that the claim information generation unit 205a determines whether the elements expressed by each important term in a set of mutually different important terms to be claimed belong to a category such as a product, a method, a program, a program recording medium, a data structure, a data recording medium, etc. Depending on which category it belongs to, a predetermined character string predetermined for each category may be specified, or a predetermined character string may be specified uniformly. As a specific example, when the key terms in a set of key terms for different claims are "lighting control device," "lighting control method," and "lighting control program," the claim information generation unit 205a generates a claim. For example, the description following "Problem to be solved by the invention" in the base specification information can be changed to "Prior art: ■ {Reasons, etc.}, ■ {Problems/effects: Reduction of ~/Improvement of ~, etc.}" The present disclosure provides a lighting control device that can be useful for {Issue/effect: reduction of ~/improvement of ~, etc.}.The present disclosure also provides a lighting control device that can be useful for ``Provides a lighting control method and a lighting control program.'' Here, the part shown in the format example in which a character string is sandwiched between a solid block and an opening curly brace "■{" and an ending curly brace "}" in the content of claim-based specification information related to an incomplete specification. The unfinished description part shown in the unfinished message for writing support indicates that it is an unfinished part, that is, a part that should be edited as appropriate, in order to be used as a specification for a patent application.

The claim information generation unit 205a uses, for example, a set of important terms of the claim subject determined based on the important terms expressing the claim subject for each claim originating item, and uses the set of important words and phrases of the claim subject that are determined based on the important terms that express the claim subject for each claim subject starting point. Determine the content to be entered in the column “Means of The claim information generation unit 205a uses at least one of the item number of each claim subject starting point, each important term of the claim subject, and a predetermined character string specified corresponding to each important term One or more sentences connected in a predetermined format are generated, and the generated sentence is determined as the written content that follows the heading "Means for Solving the Problem." The content of this predetermined character string is, for example, an incomplete message for writing support, which is indicated by the incomplete message information included in the setting information acquired by the setting information acquisition unit 201. Note that the claim information generation unit 205a determines to which category an element represented by each important term in the set of important terms to be billed belongs to, for example, a product, a method, a program, a program recording medium, a data structure, a data recording medium, etc. Depending on the category, a predetermined character string may be specified in advance for each category, or a predetermined character string may be specified uniformly. As a specific example, when the key terms in a set of key terms for different claims are "lighting control device," "lighting control method," and "lighting control program," the claim information generation unit 205a For example, the description following "Means for Solving the Problems" in the book is changed to "The lighting control device according to the present disclosure is defined as ■Claim 1. Also, the lighting control method according to the present disclosure is defined as ■Claim 2. , the lighting control program in the present disclosure is defined as ``Claim 3''. This written content may be divided into multiple lines. In this written content, the part shown in the format example of claim n to which a solid block "■Claim n" is added is the content of claim n (n is a natural number) or is predetermined in the written content. This indicates that the content can be replaced by a predetermined conversion process (for example, conversion process of deleting line breaks, period characters, etc.). Note that the claim information generation unit 205a converts the "■Claim n" part of this written content into the written content of claim n, or into content obtained by performing a predetermined conversion process on the written content. , may be replaced.

The claim information generation unit 205a can determine the content to be written in the "Brief description of drawing" column of the claim-based specification information, for example, based on the claim-based drawing information etc. generated in the claim-based drawing information generation process. As a specific example, first, the claim information generation unit 205a identifies main important words and phrases for each drawing corresponding to the drawing type of the drawing, based on the claim base drawing information and the drawing type information. For example, in a drawing whose drawing type is a block diagram or a flowchart, the key word or phrase representing the element that includes the largest number of elements in its structure becomes the main key word or phrase of the drawing. Furthermore, in a drawing whose drawing type is a diagram showing the recorded contents of a recording medium, the important words and phrases representing the elements that are the recording medium are the main important words and phrases of the drawing. In addition, in a drawing whose drawing type is a diagram representing means, functions, etc. realized by a program, the important words and phrases representing elements that are programs are the main important words and phrases of the drawing. Then, the claim information generation unit 205a generates claim-based specification information by specifying a character string in a predetermined format that includes the drawing number, the identified main important words, and the drawing type for each drawing. Determine the content to be written in the "Brief description of drawings" column. As an example, if the main important phrase of the drawing with drawing number 1 is "control system" and the drawing type is a configuration diagram, the complaint information generation unit 205a generates a column of "brief description of drawing". In "Figure 1," which is written in one line between opening and closing corner brackets, there is a sentence that reads, "Figure 1 is a configuration diagram showing an example of the configuration of a control system." The following string.

For example, the claim information generation unit 205a generates "invention" information in the claim-based specification information based on the item number of each claim section of the invention document, important phrases representing the claimed subject matter of that section, and a predetermined character string. It is possible to determine a part of the contents of the column "Detailed Description of the Invention" (that is, the contents of the description following the heading of "Detailed Description of the Invention"). The content of this predetermined character string is, for example, an incomplete message for writing support, which is indicated by the incomplete message information included in the setting information acquired by the setting information acquisition unit 201. Note that the claim information generation unit 205a determines whether the elements expressed by the key terms of the claimed subject matter in each claim of the invention document are, for example, a product, a method, a program, a program recording medium, structured data, a data recording medium, etc. Depending on which category it belongs to, you may specify a predetermined character string for each category (for example, for a program, a string that refers to installation on a computer, etc.), or you can uniformly specify a predetermined string. You may also specify a character string. As a specific example, the key phrases of claims 1 to 4 of the claims of the invention document are, in order, "lighting control device," "lighting control device," "lighting control method," and "lighting control program," In the case where claim 2 cites claim 1, the claim information generation unit 205a may include, for example, the contents to be written in a part (for example, the first part, etc.) of the "Detailed Description of the Invention" column. "The lighting control device according to the present disclosure is: ■Claim 1. Accordingly, ■ {reasons, etc.}, ■ {effects: can / can increase the possibility of / can be useful for /, etc.} .Here, ■Claim 2 may be used.This allows for ■{reasons, etc.}, ■{effects: ~can be done/~can be done/can increase the possibility of ~/can be useful for~, etc.}. , the lighting control method in the present disclosure is ■Claim 3. Thereby, ■ {effects}. Also, the lighting control program in the present disclosure is ■Claim 4. This lighting control program is applied to a device equipped with a microprocessor. By installing the lighting control program, the equipment, etc. can realize ■ {Program implementation contents, etc.: lighting control method, lighting control function, etc.}. Therefore, by this lighting control program, ■ {effects} can be realized. Note that the above-described lighting control device and related technology may be embodied in various forms as a whole or in part, such as a device, a device, an integrated circuit, a system, a method, a computer program, a computer-readable recording medium, etc. Determine “obtain.” This written content may be divided into multiple lines. In this content, the part shown in the format example of claim n to which a solid block “■Claim n” is added is added to the content of claim n (n is a natural number), or is previously added to the content of claim n (n is a natural number) It must be a part that can be replaced with the content that has been subjected to prescribed conversion processing (for example, conversion processing that deletes line breaks, period characters, and citations that cite claim sections). shows. In addition, the claim information generation unit 205a converts the "■Claim n" part of this written content into the written content of claim n, or the content obtained by performing the above-described predetermined conversion process on the written content. It is also possible to replace it with . The claim information generation unit 205a generates a corresponding code immediately after the important term whose corresponding code has been determined by the code determination process among the important terms extracted by the important term extracting unit 2051 in the written content after the replacement. You may also insert . As a specific example of the above-mentioned predetermined conversion process, when the contents of claim 5 include the statement ``of an automobile equipped with the lighting control device according to claim 1...'', the statement may be changed to ``the above-mentioned...''. An example of this is converting into an expression that does not use a "claim" on the premise that the content of the cited claim is stated above, such as "for an automobile equipped with a lighting control device...".

For example, the claim information generation unit 205a generates the "information for carrying out the invention" of the claim-based specification information based on the claim-based drawing information and drawing type information generated in the claim-based drawing information generation process, and a predetermined character string. Part of the content to be written in the "Format" column can be determined. The content of this predetermined character string is, for example, an incomplete message for writing support, which is indicated by the incomplete message information included in the setting information acquired by the setting information acquisition unit 201. As a specific example, the claim information generation unit 205a generates each drawing in descending order of drawing number based on the claim base drawing information and the drawing type in order to sequentially write it in a part of the "Detailed Description of the Invention" column. The following statements related to drawings shall be made. That is, the claim information generation unit 205a writes a sentence including the drawing number and drawing type of each drawing in order of decreasing drawing number in a part of the "Detailed Description of the Invention" column, Following that sentence, for each element based on the relationship between the elements included in the drawing, key words representing that element and other key words if necessary to explain the key word; Write a sentence including an unfinished message for writing support to explain the important phrase. Note that if the important word/phrase setting dictionary acquired by the important word/phrase setting dictionary acquisition unit 203 includes a sentence that explains the important word in association with the important word/phrase, the complaint information generating unit 205a generates a complaint base. In the "Detailed Description of the Invention" column of the specification information, instead of or in addition to the incomplete message for description support to explain the important words, A sentence corresponding to the important word in the word/phrase setting dictionary and explaining the important word may be written. Other important words and phrases that are necessary to explain the important words and unfinished messages for writing support to explain the important words are, for example, "element expressed by the important word and element expressed by the other important word". It can be determined according to the attributes of the important words and phrases, such as "the structural inclusion relationship of the important words", "the recorded content of the element represented by the important words", and the "realization content of the elements represented by the important words", and the important words. For example, the claim information generation unit 205a determines, for each of the plurality of important words and phrases extracted from the claims of the invention document by the important words and phrases extraction unit 2051, the composition of the element represented by the important word and the element represented by other important words. A claim is made based on an unfinished message for description support to explain the key word, which is determined based on the inclusion relationship, the recorded content of the element represented by the key word, or the realization content of the element represented by the key word. It is possible to determine a part of the description content following the heading of "Detailed Description of the Invention" in the base specification information. For example, the claim information generation unit 205a determines, for each of the plurality of important terms extracted by the important term extraction unit 2051 from the claims of the invention document, that the element represented by the important term structurally includes the element represented by other important terms. Or, when it is structurally included in the element expressed by another important word, a code is determined for that important word based on the structural inclusion relationship of that important word and other important words. Sign determination processing may be performed. The claim information generation unit 205a generates important words and phrases in code determination processing from among important words and phrases included in at least a part of the written contents following the heading of “Detailed Description of the Invention” in the claim-based specification information. For important words and phrases whose codes have been determined, that code is inserted immediately after the important words and phrases. As a specific example, in the claim-based drawing information, "Figure 1" is a "configuration diagram" of the main important phrase "control system", and "control system" is a "human sensor" and "control content determining device". "Lighting control signal transmitting device" is included in the configuration, and the code corresponding to "control system" is "10", the code corresponding to "human sensor" is "11", and "control content determining device". In the case where the corresponding code is determined to be “12” and the code corresponding to “Lighting control signal transmitter” is determined to be “13”, the claim information generation unit 205a determines that “Practicing the invention” in the claim-based specification information 1 is a block diagram showing an example of the configuration of the control system 10. Attribute/shape/size/material/aspect/input/output/generation-related information, etc.)}.The control system 10 includes a human sensor 11, a control content determining device 12, and a lighting control signal transmitting device 13. The sensor 11 is: ■{Description of the human sensor 11}.The control content determining device 12 is: {Description of the control content determining device 12}.The lighting control signal transmitting device 13 is: {Description of the lighting control signal transmitting device 13} }.''. For example, after writing a description related to "FIG. 1" in a part of the "Detailed Description of the Invention" column in the claim-based specification information, the claim information generation unit 205a creates a description following the description. A description related to FIG. 2 will be provided. Note that, since the same element may be included in multiple drawings indicated by the claim-based drawing information, the claim information generation unit 205a generates "for carrying out the invention" in the claim-based specification information in order to prevent redundant descriptions. In the ``format'' column, sentences with the same content may not be written multiple times. For example, the claim information generation unit 205a selects, among all the important words and phrases extracted by the important phrase extraction unit 2051 from the claims of the invention document acquired by the invention document acquisition unit 2021, the “for carrying out the invention” in the claim-based specification information. If there are important words or phrases that are not used in the explanations, etc. related to each of the drawings mentioned above in the "Format" column, include an incomplete message to support the description to explain the important words or phrases that are not used in the explanations, etc. The text may be added to the "Detailed Description of the Invention" column. As a result, the "Detailed Description of the Invention" column in the claim-based specification information can include explanations of all important words and phrases stated in the claims of the invention document.
In addition, as a modification of the claim information generation unit 205a, the claim information generation unit 205a can generate inclusions between important words expressed in the claims of the invention document, for example, without using the claim base drawing information and the drawing type information. A sentence indicating the inclusive relationship between the important words based on the relationship may be determined as part of the description following the heading of "Detailed Description" in the claim-based specification information. Further, as another example of the claim information generation unit 205a, the claim information generation unit 205a generates an element representing each of a plurality of important terms and phrases extracted by the important term extraction unit 2051 from the claims of the invention document. To explain the important word/phrase, which is determined based on the structural inclusion relationship with the element expressed by other important word/phrase, the recorded content of the element expressed by the key word/phrase, or the content of realization of the element expressed by the key word/phrase. A part of the written content following the heading "Detailed Description of the Invention" in the claim-based specification information may be determined based on the unfinished message for writing support, etc. For example, if the important phrase "control system" extracted from the complaint includes the important phrases "human sensor," the important phrase "control content determining device," and the important phrase "lighting control signal transmitting device." In this example, the complaint information generation unit 205a generates a message that says, as part of the description following the heading “Detailed Description of the Invention,” “The control system includes, for example, a human sensor and a control content determining device. It may include a lighting control signal transmitting device, or it may not include, for example, a human sensor, a control content determining device, or a lighting control signal transmitting device. Please modify this sentence as appropriate based on your judgment as to whether it is an essential element or not, and whether it is an element that should be included in this example embodiment).'' It is also possible to write a character string with an incomplete message for writing support, which does not necessarily indicate a relationship and is used after being modified as appropriate. Further, for each important word/phrase in this character string, a corresponding code may be added immediately after the important word/phrase. As a result, inventors, etc. or patent attorneys can easily create specifications that include embodiments that can accommodate future changes to independent claims, etc., based on claim-based specification information. there is a possibility.

The claim information generation unit 205a can include a fixed phrase in the complaint-based specification information, for example, according to the fixed phrase information in the setting information acquired by the setting information acquisition unit 201. Further, the claim information generation unit 205a can insert a paragraph number into each line in the claim-based specification information, for example.

In the invention document analysis system 20, when the output unit 208 outputs patent application related information including claim-based specification information as a display document (for example, an HTML document), the claim information generation unit 205a outputs the patent application related information including claim-based specification information as a display document. Generates claim-based specification information included in patent application related information. FIG. 25 shows an example of claim-based specification information in patent application related information as a display document displayed on a browser. FIG. 25 is an example of claim-based specification information generated in response to the claim information illustrated in FIG. 22. In FIG. 25, a part of the claim-based specification information is omitted for convenience. For example, a solid block in the claim-based specification information illustrated in FIG. 25, an incomplete message (for example, an important word The display element of "■{Description of Part A 11}" (incomplete message for writing support) related to "Part A" is an operation-compatible GUI element, and is acquired by the invention document acquisition unit 2021 in response to a mouse click operation, etc. Display control information is added so that an important word-inclusive description enumeration process for enumerating the descriptions of each line including important words related to the incomplete message for description support is executed from another invention document related to the invention document that has been created. It is also good to be present. This other invention document is, for example, one or more invention documents specified by a user's designation (for example, designation in setting information). This operation-compatible GUI element for executing the important word inclusion description enumeration process related to the incomplete message for description support allows users (for example, inventors, patent attorneys, etc.) to It may be easier to revise the sentence to include appropriate content. For example, in the important word inclusion description enumeration process related to the incomplete message for writing support "■ {Description of A part 11}" related to the important word "Part A", for example, the case particle " A statement followed by the case particle compound "towa" or the case particle compound "towa" may be searched for in another invention document related to the invention document. As a result, the important word/phrase inclusion description enumeration process can enumerate lines including sentences that describe the important word/phrase "Part A" as the subject.

The claim-based specification information portion of the patent application-related information as a display document generated by the claim information generation unit 205a is, for example, a claim-based specification information portion related to an incomplete specification generated in the above-described claim-based specification information generation process. Contains one or more display elements corresponding to the contents of each line in the specification information, and if there is one or more important words extracted by the important word extraction unit 2051 in each line of each item, each important word is including one or more corresponding display elements. As a specific example, when claim-based specification information is written between an <article> tag and a </article> tag in an HTML document, the claim information generation unit 205a writes the claim-based specification information under the <article> tag. For each line of claim-based specification information related to an incomplete specification generated in the document information generation process, display elements corresponding to that line are displayed using <span> tags and </span> tags to express the contents of that line. For each important term in that line, write the display element corresponding to that important term under the <span> tag of the display element corresponding to that one line, using the <span> tag and the </span> tag. It is written by putting the important words between tags. For example, a browser may interpret a document for display and display each display element. Each display element corresponding to each line included in the claim-based specification information portion allows, for example, to display the contents of the corresponding line.

The claim information generation unit 205a generates, for example, a display element corresponding to a heading, an important phrase, an incomplete message for writing support, or a code included in the claim-based specification information portion of the display document. Display control information for controlling aspects (for example, highlighting, colored display, display transition, etc.) is added. This display control information can be realized, for example, in the same manner as the display control information added to the display element in the claim information in the display document described above. Therefore, the display element of the important word/phrase written in the claim-based specification information part of the display document may, for example, display the description part containing the important word/phrase in the claim information of the display document in accordance with the operation. can function as an operation-compatible GUI element. For example, the claim information generation unit 205a adds display control information for controlling the display mode of the display element to the display element of the important word included in the claim-based specification information in the display document, and adds display control information to the display element for controlling the display mode of the display element. In order to make the display element of the important phrase function as a GUI for displaying (for example, displaying a link) the written part of the claim information in the display document that includes the important phrase according to the operation of the display element. belongs to. By displaying claim-based specification information on a browser, a user (for example, an inventor or a patent attorney) may be able to efficiently create an appropriate specification. Further, the user can create an invention document including claims and the specification by including the created specification in the invention document etc. that served as the basis for creating the claim-based specification information. Then, the user can cause the invention document analysis system 20 to analyze the invention document including the claims and the specification again, including warning information, specification information, etc. output as the analysis result of the specification of the invention document. Patent application related information will be available.

An example of display control information added to a display element corresponding to a code included in a claim-based specification information portion of a display document will be described below.

When a code is included in the claim-based specification information, the claim information generation unit 205a displays the code in a portion including the same code as the code in the claim-based drawing information in the display document according to the operation. Display control information for making the display element have the GUI function may be added to the display element of that code. This display control information defines, for example, the function name of a script function that is activated with a sign as an argument in response to an operation, or the function itself. It is specified that a part in the drawing information is specified by searching for its code. For example, the claim information generation unit 205a generates a function related to display control information for displaying a portion containing the same code as the code in the claim-based drawing information in response to an operation on a code in the claim-based specification information. It may be determined that a portion of the claim-based drawing information that includes the same code as the code in the claim-based drawing information is displayed within a range that does not hide the display element of the code that is displayed and operated on the browser. For example, the claim information generation unit 205a divides the display screen of the browser into a plurality of display areas, and divides the display screen of the browser into a plurality of display areas, and divides the display area of the claim-based specification information and the claim information that is displayed according to the operation related to the code in the claim-based specification information. A display area such as a portion of the base drawing information may be displayed in parallel. For example, the claim information generation unit 205a uses display control information added to a display element for a code to cause a function that is activated in response to an operation on the code to be activated in response to a further operation to generate a function that is the same as the code in the claim base drawing information. It may be specified to perform a process of switching and displaying multiple parts including a code (for example, two or more drawings including the code), or all parts including the same code as the code in claim-based drawing information. It may be determined that a process is performed to reduce the size and display the images all at once as necessary. For example, the claim information generation unit 205a generates a function related to display control information for displaying a portion containing the same code as the code in the claim-based drawing information in response to an operation on a code in the claim-based specification information. The same code as that in the claim-based drawing information may be highlighted.

Hereinafter, an example of display control information added to a display element corresponding to an incomplete message for description support included in a claim-based specification information portion of a display document will be described.

The claim information generation unit 205a generates a display having a GUI function for changing the contents of the incomplete message for writing support in accordance with an operation when the claim-based statement information includes an incomplete message for writing support. A display element configured by adding display control information for an element (that is, an operation-compatible GUI element) to an incomplete message for writing support (for example, a <span> tag, etc.) to the incomplete message for writing support. ) may be added. Addition of display control information is realized, for example, by writing information indicating an operation and information indicating a function activated by the operation in a <span> tag added to the incomplete message for writing support. This display control information is information for starting a script that executes the above-mentioned important word inclusion description enumeration process in response to an operation when an important word is included in the incomplete message for writing support, or is information for starting a script that executes the above-mentioned important word inclusion description enumeration process, or as shown in the example below. This is information for starting a script that executes a process, etc. This display control information defines, for example, the function name or the function itself of a script function that is started with an incomplete message for writing support as an argument in response to an operation. Display an input field with an incomplete message for support as the default value, reflect the editing operation on the input field, and use the string of the editing operation result as the display content of the display element instead of the original incomplete message for support. It is determined that In addition, the claim information generation unit 205a may generate, for example, an incomplete message for description support following the subject "X device 1 is" included in the claim-based specification information, such as "■{Description of X device 1 ( Function/effect/operating principle/attribute/shape/size/material/aspect/input/output/generation-related information, etc. /Input/Output/Generation related information'' Display control for converting the unfinished message for writing support into a display element with a GUI function to enable selection of options when the message lists options such as The information may be added to the display element of the incomplete message for writing support. Then, the complaint information generation unit 205a converts the function of the script related to the display control information into the content of the incomplete message for writing support when a selection operation is performed on the option of the displayed incomplete message for writing support. , another predetermined unfinished message suitable for the selected option (for example, function, action, operation, principle, attribute, shape, size, material, mode, input/output, or generation-related information) For each option, it may be determined that the message is changed to a description support incomplete message that includes one example sentence corresponding to the option from among the incomplete example sentences suitable for explaining the option. For example, if "input/output" is selected on the GUI of the unfinished message for writing support that includes the options in this example, the unfinished message for writing support that includes that option will be set to "input/output". Suitable for writing support such as "■{(What) is input from (where), and X device 1 receives that input and outputs (what) as the processing result of X device 1 to (where)}" Changed to unfinished message.

[3-1-5-7-4. Generation of claim-based summary information]
The claim information generation unit 205a realizes claim-based summary information generation processing by generating claim-based summary information based on the claims included in the invention document acquired by the invention document acquisition unit 2021. Claim-based summary information represents the contents of an incomplete summary. Claim-based abstract information does not need to be complete enough to be used as an abstract for a patent application.

The claim-based abstract information includes, for example, headings similar to the abstracts of patent applications in Japan (for example, "Abstract", "Selected Figures", etc.), and includes descriptions following each of the headings.

The claim information generating unit 205a generates a “ Decide what to write in the “Summary” column. The content of this predetermined character string is, for example, an incomplete message for writing support, which is indicated by the incomplete message information included in the setting information acquired by the setting information acquisition unit 201. As a specific example, the complaint information generation unit 205a may write, for example, "■ {Technical Issues/ Effect: Provides a control system useful for reducing/improving...'', etc., and is the subject of the first claim of the claim. After the heading ``Solution'', write a sentence containing important words and phrases, and add the content of the first paragraph of the claim included in the invention document, line break deletion processing, conversion such as processing that makes the subject of the claim the subject, etc. The result of the processing is written, and a code corresponding to the important word (for example, the content of the "code" as an attribute of the important word) is inserted immediately after each important word in the description. As a result, the description following the heading "Solution Means" becomes, for example, "The control system 10 includes a human sensor 11, a control content determining device 12, and a lighting control signal transmitting device 13." . For example, when the number of characters in the "Summary" column exceeds a predetermined upper limit (for example, 400), the claim information generation unit 205a may include a message to that effect in the claim base summary information.

The claim information generation unit 205a stores, for example, the drawing number of a figure representing a characteristic matter related to one claimed subject of a claim in the column of “selected drawing” in the claim-based abstract information in a normal invention document. "Figure 1" is included based on the assumption that there is a relatively high possibility that the Note that there are cases where some effect is aimed at by changing the order of the figures used in the invention document, and therefore it is not necessarily necessary to set the drawing number of a figure related to one claimed subject matter of a claim to 1 in the invention document.

As another example, the claim information generation unit 205a may include, for example, in the column of "selected drawings" in the claim base summary information, important words and phrases included in the first section of the claim following the characters "drawing". The drawing number of the drawing that includes the most elements expressed by the important words extracted by the extraction unit 2051 is described.

The claim information generation unit 205a adds the Japanese version of the abstract for the US application ( In other words, it is also possible to include descriptions corresponding to the Japanese manuscript for translation of the abstract. For example, the claim information generation unit 205a adds line break deletion processing to the content of the first paragraph of the claim included in the invention document, and sets the subject of the claim as the subject, as description content equivalent to the Japanese version of the abstract for US applications. The result of conversion processing such as processing is used. For example, the claim information generation unit 205a may cause the translation unit 2054 to machine translate the description content corresponding to a Japanese version abstract for a US application, and the number of words as a result exceeds a predetermined upper limit (for example, 150 words). In some cases, a message indicating that it is desirable to reduce the number of characters may be included in the claim base summary information.

The claim information generation unit 205a includes, in addition to the incomplete abstract of the Japanese patent application, the incomplete abstract of the Japanese international application in the claim-based abstract information generated by the claim-based abstract information generation process. It is also possible to include written content corresponding to . For example, the claim information generation unit 205a adds line break deletion processing to the content of the first paragraph of the claim included in the invention document, and adds the subject of the claim to the content corresponding to the incomplete abstract of an international application in Japanese. Immediately after each important word, the code corresponding to the important word (for example, the content of the "code" as an attribute of the important word, or the code determination process) ) is inserted between opening and closing parentheses.

In the invention document analysis system 20, when the output unit 208 outputs patent application related information including claim-based abstract information as a display document (for example, an HTML document), the claim information generation unit 205a outputs the patent application related information including claim-based abstract information as a display document. Generates claim-based abstract information included in patent application related information.

The claim-based summary information part of the patent application related information as a display document generated by the claim information generation unit 205a is, for example, the content of each line of the claim-based summary information generated in the above-described claim-based summary information generation process. If there are one or more important words extracted by the important word extraction unit 2051 in each line, one or more display elements corresponding to each important word are included. include. As a specific example, when the claim information generation unit 205a writes the claim base summary information between the <article> tag and the </article> tag in the HTML document, the claim information generation unit 205a writes the claim base summary information under the <article> tag. For each line that is the content of the complaint-based summary information generated in the book information generation process, the display element corresponding to that line is described by sandwiching the content of that line between <span> and </span> tags. Then, under the <span> tag of the display element corresponding to that one line, for each important term in that line, the display element corresponding to that important term is added using the <span> tag and the </span> tag. Write by inserting important words. If the content of each line of claim-based summary information includes a heading, an incomplete message for writing support, or a code, the display element for that line will be the heading, incomplete message for writing support, or a code. , or a display element corresponding to the code. For example, a browser may interpret a document for display and display each display element. Each display element corresponding to each line included in the claim-based summary information portion allows, for example, to display the contents of the corresponding line.

The claim information generation unit 205a generates, for example, a display element corresponding to a heading, an important phrase, an incomplete message for writing support, or a code included in the claim base summary information portion of the display document. Display control information for controlling aspects (for example, highlighted display, colored display, etc.) is added.

[3-1-5-7-5. Generation of revised complaint information]
The claim information generation unit 205a realizes the modified claim information generation process by generating modified claim information by modifying the claims included in the invention document acquired by the invention document acquisition unit 2021. The modified version of the claim information is obtained by modifying a part of the claim information generated by the claim information generation process. For example, the user can use the content of the revised claim information as a claim of an invention document and cause the invention document analysis system 20 to analyze the invention document again, thereby generating new warning information including invention document evaluation information. It becomes possible to obtain information related to patent applications.

[3-1-5-7-5-1. Correction of inappropriate descriptions and deletion of unnecessary descriptions]
The claim information generation unit 205a determines a correction method in advance when a statement in a claim of an invention document acquired by the invention document acquisition unit 2021 is a target for generating indication information in the warning information generation unit 2059. When the claim information is corrected, the description is corrected according to the correction method to generate corrected version claim information. A specific example of generating revised claim information by modifying the claim description will be described below.

The claim information generation unit 205a may, for example, convert the description in the claims of the invention document acquired by the invention document acquisition unit 2021 into the invention document standard regulations for translation acquired by the invention document standard regulations for translation acquisition unit 204 (see FIG. 14). ), if the pattern indicated by the pattern information of a condition item to which modification information has been added corresponds to the pattern indicated, the statement in the claim is modified based on the modification information to generate modified claim information representing the modified claim. generate.

For example, the claim information generation unit 205a generates the number of each claim of the invention document acquired by the invention document acquisition unit 2021 (for example, the number written in the “n” part of the heading “Claim n” in the first line of each item). (numbers represented by strings of Arabic numerals) are not consecutive numbers along the line order, the numbering of each item of the claim should be revised so that the numbers are consecutive starting from 1 along the line order, and , if there is a section cited in other sections that cites another section in any claim section, the cited section number in the section cited in other sections should be corrected to correspond to the number of each section after the amendment. By doing so, revised claim information is generated.

For example, the claim information generation unit 205a generates the number of each claim of the invention document acquired by the invention document acquisition unit 2021 (for example, the number written in the “n” part of the heading “Claim n” in the first line of each item). If the number represented by a character string containing a mixture of half-width characters and full-width characters, the number of each item of the claim is modified to include one type of character (for example, a double-width character), and If any claim section contains a statement in which the referenced item number in the other-claim cited part that cites another claim is not one of the types of characters, By correcting the cited item number to one of the types of characters (for example, double-byte characters), the revised claim information is generated.

For example, when the claim of the invention document acquired by the invention document acquisition unit 2021 includes a description of a mathematical formula represented by a heading such as “number n” written in corner brackets, the claim information generation unit 205a , when the numbers of the formula are not consecutive numbers in the line order even after considering the branch numbers, the revised claim information is generated by modifying the numbers so that they are consecutive numbers in the line order. The claim information generation unit 205a can modify, for example, a chemical formula or a table in the same way as modifying the formula number.

For example, the claim information generation unit 205a generates a comment line or a comment that can be specified based on the comment identification information of the setting information acquired by the setting information acquisition unit 201 in the claim of the invention document acquired by the invention document acquisition unit 2021. If the comment line or comment is included, the modified claim information is generated by deleting the comment line or comment. Note that whether to execute or suppress the comment line in the complaint or the function of deleting the comment in the revised complaint information may be specified in the setting information acquired by the setting information acquisition unit 201 or the like.

For example, the claim information generation unit 205a determines whether the claim of the invention document acquired by the invention document acquisition unit 2021 includes a description of consecutive comma characters “,” or a description of consecutive period characters “.” In this case, modified claim information is generated by modifying the comma character "," to one character or the period character "." to one character.

For example, if a full-width or half-width space character or tab is included at the end of a line of a claim in the invention document acquired by the invention document acquisition unit 2021, the claim information generation unit 205a deletes the space character or tab. By doing so, revised claim information is generated. In addition, the claim information generation unit 205a generates modified claim information by correcting the line feed code at the end of the line of the claim to a predetermined code for the end of the line if it is not a predetermined code for the end of the line. It may also be generated. This predetermined code for the end of a line is, for example, a code formed by arranging 0x0d and 0x0a in this order. The predetermined code for the end of a line may be specified by setting information acquired by the setting information acquisition unit 201. For example, the claim information generation unit 205a may modify the line feed code to a predetermined code for the end of a line only when the patent application related information is output as a text file or the like instead of as a display document. It may also be done.

For example, if the claims of the invention document acquired by the invention document acquisition unit 2021 include half-width characters, the claim information generation unit 205a corrects the claims by changing the half-width characters to corresponding full-width characters. Generate version claim information. In this case, the claim information generation unit 205a may decide not to change to full-width characters when the claims of the invention document include half-width English characters sandwiched between opening parentheses and closing parentheses. good. Further, in the case where the numbers of each section of the claims of the invention document are all uniformly half-width characters, the claim information generation unit 205a may not change the number of each section to full-width characters. If a claim of an invention document includes half-width characters, the setting information acquisition unit 201 acquires the type of half-width characters (for example, alphabetic characters, katakana characters, symbols, etc.), character code range, etc. that are to be changed to full-width characters. The claim information generation unit 205a may change half-width characters to full-width characters in a claim to generate revised claim information in accordance with the designation. Conversely, if the claims of the invention document include full-width characters, the setting information acquired by the setting information acquisition unit 201, etc. In accordance with the specification, the claim information generation unit 205a may change full-width characters in the claim to half-width characters to generate revised claim information.

For example, when one or more half-width blank characters are included at the beginning of each line of a claim of the invention document acquired by the invention document acquisition section 2021, the claim information generation unit 205a generates the one or more half-width blank characters. By changing a blank character to a single full-width blank character, a modified version of claim information is generated.

For example, when two or more half-width blank characters are consecutively included in a part other than the beginning of each line of a claim of an invention document acquired by the invention document acquisition unit 2021, the claim information generation unit 205a generates a blank character. Modified version claim information is generated by changing two or more half-width blank characters to one half-width blank character. This is useful, for example, to correct a description in which multiple blank characters are mistakenly inserted between English words. For example, assuming that an English sentence with multiple spaces intentionally added at the end of each sentence needs to be written in a claim in an invention document, the user can specify in the settings information that this correction will not be made. may be selectable, and this modification may be suppressed according to setting information.

[3-1-5-7-5-2. Modifications to conform to specified format]
If the setting information acquired by the setting information acquisition unit 201 includes information specifying the format of claims in the revised claim information, the claim information generating unit 205a formats the claims of the invention document in accordance with the specified format. By modifying the information so that

An example of information specified regarding the claim format in the revised claim information included in the setting information is the specification of whether or not to add a corresponding code to a word or phrase in a claim, if there is a corresponding code. The following information is included. When there is a code corresponding to a word or phrase, if there is a code determined by the code determination process for that word or phrase, in the invention document (for example, claim, specification, etc.) acquired by the invention document acquisition unit 2021. If there is a description with a code added to the word or phrase, or if there is a description of the code corresponding to the word in the "Explanation of code" column of the specification of the invention document.

When the setting information acquired by the setting information acquisition unit 201 includes information specifying that a code corresponding to a word in a complaint is to be added to the word, the complaint information generating unit 205a , among the important words and phrases extracted by the important word extraction unit 2051 in the claims of the invention document, for the written part where no code is added to the important word with the corresponding code, the written part starts immediately after the written part of the important word. By inserting the description of the code between the parentheses and the closing parentheses, the revised claim information is generated. Note that if there are multiple codes corresponding to the important phrase, the claim information generation unit 205a selects one of the codes that comes first by sorting the multiple codes, such as in natural order, and selects the first code. The same code may be used to insert after the important word. Further, if the setting information acquired by the setting information acquisition unit 201 includes information specifying that the code not to be added to the word/phrase in a complaint when there is a corresponding code, the complaint information generating unit 201 205a, for a description part in which a code is added to an important word with a corresponding code among the important words and phrases extracted by the important word phrase extraction unit 2051 in the claim, a description related to the code in the description part (for example, By deleting the code between the opening parentheses and the closing parentheses, the revised claim information is generated.

In the setting information acquired by the setting information acquisition unit 201, information specified regarding the claim format in the revised claim information includes, for example, the description of the claimed subject or the specification of the cited section to the claimed subject in each claim section. In the description, the preceding written part "contains" or "comprises" the first part of any one of "and", "and", or an abbreviation, the second part of a line break or an abbreviation, In the case where the third part of either "comprising" or "containing" and the fourth part of line break or omission are arranged in this order: , the claim information generation unit 205a formats the claim of the invention document according to the specification. By making the following corrections, revised claim information is generated.

[3-1-5-7-5-3. Addition of claims section]
For example, the claim information generation unit 205a determines that in the claim of the invention document acquired by the invention document acquisition unit 2021, there is an independent claim that describes a claimed subject belonging to the product category in the constituent feature enumeration type and belongs to the method category. If there is no clause stating the claimed subject matter, by performing a predetermined character string conversion process on the independent claim stating the claimed subject matter belonging to the product category, it will be made similar to the independent claim to some extent. The revised claim information is generated by generating a section describing the claimed subject matter belonging to the method category, which represents the contents of the invention document, and adding the generated section after the final section of the claim of the invention document. Note that the claim information generation unit 205a can determine whether a claim section of an invention document is an independent claim based on, for example, claim citation relationship information held by the claim citation relationship detection unit 2057. The claim information generation unit 205a may use any method to determine whether the statement in the claim section of the invention document is of the constituent feature enumeration type. For example, if the character string "and" or "and," is present at the end of any line in a claim section of an invention document, the claim information generation unit 205a determines that the statement in that section is a component enumerated type. It can be presumptively determined that

For example, the claim information generation unit 205a detects that in the claim of the invention document acquired by the invention document acquisition unit 2021, there is an independent claim stating a claimed subject belonging to the method category in the constituent feature enumeration type, and the program category or program If there is no clause stating the claimed subject matter belonging to the recording medium category, by performing a predetermined character string conversion process on the independent claim stating the claimed subject matter belonging to the method category, Alternatively, by performing the process of generating a clause that cites the independent claim, a clause that describes the claimed subject matter belonging to the program category or the program recording medium category that expresses the same contents to some extent as the independent claim can be generated. Modified claim information is generated by generating a section that describes the claimed subject matter to which it belongs, and adding the generated section after the final section of the claim of the invention document.

In the setting information acquired by the setting information acquisition unit 201, a claim section is added to the revised claim information in which the claim subject belongs to any of the above-mentioned method category, program category, and program recording medium category. The claim information generation unit 205a may include information specifying whether the claim information is correct or not, and the claim information generation unit 205a may execute or suppress the addition of a claim section to the revised claim information in accordance with the specification.

When adding a claim term to the revised claim information, the claim information generation unit 205a, for example, sends a claim to the claim citation relationship detection unit 2057 that represents the citation relationship between each claim term in the claim information. In addition to the citation relationship information, modified claim citation relationship information representing the citation relationship between each clause of a claim in the revised claim information is generated and retained.

[3-1-5-7-5-4. Addition of claims corresponding to each region where the application can be filed]
The claim information generation unit 205a performs the above-described modified claim information generation process to correct the inappropriate descriptions mentioned above, delete unnecessary descriptions, and modify the claims included in the invention document to conform to the specified format. After generating revised claim information representing claims that have been revised by adding claim sections, etc., based on the revised claim description, information that corresponds to each region where the application can be filed (for example, the United States, Europe, etc.) is generated. Claims may be generated and each generated claim may be further included in the revised claim information.

For example, the claim information generation unit 205a includes, in the setting information acquired by the setting information acquisition unit 201, correction information that defines a correction method to be used when generating revised claim information for the area where the application is planned. In this case, a claim modified according to the modified information may be added to the modified version of the claim information.

Further, for example, if the setting information acquired by the setting information acquisition unit 201 includes information specifying a region where the application is to be filed where a claim is to be generated, the claim information generation unit 205a may be configured to correspond to the specified region. It is also possible to generate a new claim and add it to the revised claim information. This added claim can be used as a claim in a Japanese manuscript for translation for a patent application to the applicant. In addition, for example, the setting information acquired by the setting information acquisition unit 201 may include information specifying the content of claim modification that is useful for filing an application in a region where the claim should be generated. The generation unit 205a may generate a claim corresponding to the revised content of the specified claim and add it to the revised claim information. Examples of information specifying the content of modification of a claim include information specifying that one term in a claim that cites multiple terms be changed to a single term that cites only one of the multiple terms; Information specifying that a term citing multiple terms is to be changed into the same number of terms, each citing only one different one of the multiple terms, and multiple terms cited in the claim. specifies that if any of the cited terms in a single term cited multiple terms, that one term is to be changed to a single term that cites only one of the multiple terms. Examples include information. In response to the specification of these setting information, the complaint information generation unit 205a adds the specification in the setting information to each claim item (that is, claim set) that the complaint information generation unit 205a adds to the revised complaint information. A corresponding heading may be added. Examples of headings for claim sets added to the revised claim information include "Claims for the United States," "Claims for Europe," "Claims for Eliminating Multiple Dependencies," and "Claims for Eliminating Multiple Dependencies."

Hereinafter, a specific example of how the claim information generation unit 205a generates claims corresponding to each region where an application can be filed will be described.

If the setting information acquired by the setting information acquisition unit 201 includes information specifying the United States as a possible region for filing an application, or if a claim cites multiple terms, one of the multiple terms In the case where the claim information generating unit 205a includes information specifying that a single section is to be changed to a single section that only cites a section, the claim information generation unit 205a generates, for example, revised claim information representing a revised claim (that is, a claim in a state in which the above-mentioned modification has been made). When a claim in (information) has one term that cites multiple terms, each term of the claim is changed by changing that term to a term that cites only the one with the lowest item number among the multiple terms. , further appended to the modified claim information representing the original modified claim. Further, when the setting information acquired by the setting information acquisition unit 201 includes information specifying the United States as a region where the application can be filed, the claim information generating unit 205a determines that the number of claim terms is a predetermined number (for example, 20). Within the scope, a term citing multiple terms in a claim, for example, a term citing only the one with the lowest item number among the multiple terms, and a term next to the item number among the multiple terms. Each term of the claim that has been changed into two terms, including the smaller one cited term, may be further added to the revised claim information representing the original revised claim. Note that in a US patent application, although it is not necessarily inappropriate to state a section in which multiple sections are cited in a claim, such a change may be useful in some cases.

The configuration information acquired by the configuration information acquisition unit 201 indicates that when any of the cited terms of a single term that cites multiple terms in a claim cites multiple terms, that one term is included in the multiple terms. In the case where the claim information generating unit 205a includes information specifying that only one of the claims be changed to one term citing the revised claim, the claim information generating unit 205a may, for example, add the following information to the claim in the revised claim information representing the revised claim: When any of the quotation destinations of a term that cites multiple terms cites multiple terms, that one term is changed to the term that cites the term with the lowest item number among the multiple terms. If the term with the smallest item number cites multiple terms, the term with the smallest item number is added to the term that cites the term with the smallest item number among the multiple terms cited. , and each section of the modified claim is further added to the modified claim information representing the original modified claim.

In a case where the setting information acquired by the setting information acquisition unit 201 includes information specifying the United States as a potential filing destination region, the claim information generating unit 205a may, for example, A claim includes a section stating the claimed subject matter belonging to the program category, and a clause stating the claimed subject matter belonging to the program recording medium category that includes a program statement that has a higher correlation with the program statement in that section than a prescribed standard. is separately included, each section of the claim after deleting the section describing the claimed subject matter belonging to the program category is further added to the modified claim information representing the original modified claim.

In a case where the setting information acquired by the setting information acquisition unit 201 includes information specifying the United States as a potential filing destination region, the claim information generating unit 205a may, for example, A claim includes a section stating the claimed subject matter belonging to the program category, and a clause stating the claimed subject matter belonging to the program recording medium category that includes a program statement that has a higher correlation with the program statement in that section than a prescribed standard. is not separately included, by applying a predetermined character string conversion process to the contents of the section describing the claimed subject matter belonging to the category of the program, that section will be changed to the section that describes the claim subject belonging to the category of the program. The recording medium that non-temporarily records the claim subject is changed to the claim subject, and each clause of the claim after the change is further added to the revised claim information representing the original revised claim.

When the setting information acquired by the setting information acquisition unit 201 includes information specifying Europe as a region where the application can be filed, the claim information generation unit 205a may, for example, Among the important words and phrases extracted by the important words and phrases extraction unit 2051 in the claim, for the written part where no code is added to the important word with the corresponding code, an opening parenthesis and an ending parenthesis are placed immediately after the written part of the important word. Insert the statement with the reference numeral between , and add each clause of the inserted claim to the modified claim information representing the original modified claim. The code corresponding to the important term is, for example, a code determined by the code determination process, or a code added to the important term in the invention document (for example, claim, specification, etc.) acquired by the invention document acquisition unit 2021. This is the code where it is described, or the code where there is a description of the code corresponding to the important phrase in the "Explanation of code" column of the specification of the invention document. Note that if there are multiple codes corresponding to the important phrase, the claim information generation unit 205a selects one of the codes that comes first by sorting the multiple codes in natural order, for example, and selects the first code. The same code may be used to insert after the important word.

When the claim information generation unit 205a adds each claim section (that is, a claim set) corresponding to each region where the application can be filed to the revised claim information, for example, the claim information generation unit 205a instructs the claim citation relationship detection unit 2057 to In addition to claim citation relationship information that represents the citation relationship between each claim term in the claim information, modified claim citation relationship information that represents the citation relationship between each claim term for each claim set in the revised claim information. Generate and retain information.

[3-1-5-7-5-5. Generation of revised claim information part in display document]
In the invention document analysis system 20, when the output unit 208 outputs patent application related information including revised claim information as a display document (for example, an HTML document), the claim information generation unit 205a outputs the patent application related information including revised claim information as a display document. Generate revised claim information included in patent application related information.

The modified claim information portion of the patent application related information as a display document generated by the claim information generation unit 205a includes, for example, one or more pieces of information corresponding to the contents of each line of each section of the modified claim in the modified claim information. If there is one or more important words extracted by the important word extraction unit 2051 in each row of each term, one or more display elements corresponding to each important word are included, and each of the When each row of a section includes a heading such as "Claim n" (n is a natural number), one or more display elements corresponding to the heading are included. As a specific example, when the revised claim information is written between the <article> tag and the </article> tag in an HTML document, the claim information generation unit 205a creates the revised claim information under the <article> tag. For each line of each clause of the amended claim as Under the <span> tag of the corresponding display element, for each important word in that line, describe the display element corresponding to that important word by sandwiching the important word between the <span> and </span> tags. do. For example, a browser may interpret a document for display and display each display element. Each display element corresponding to each line of each section included in the revised claim information portion allows, for example, to display the contents of the corresponding line.

To give a specific example of a display element of one important term in the revised claim information in the display document generated by the claim information generation unit 205a, multiple important terms "equipment control unit" are described in the revised claim information. For example, the display element corresponding to the fourth "equipment control unit" in the case of )' title="...(Information added display content)...">Device control section</span>", etc. In this example, the identification code of the important term "equipment control section" is "c003", and the identification code of the revised claim information "revclaiminfo" and the important term A character string including the identification code "c003" and the number "004" in the order of description is defined. Further, in this example, when a mouse click operation is performed on a display element, the wordProc function is called with the value of the id attribute as an argument. As a result, the wordProc function can, for example, specify the display element from the argument and refer to the character string of the important word related to the display element. Extract the word identification code "c003", the order number "004", etc., and specify that the id attribute of the first display element of the same important word is "revclaiminfo-c003-001". It is now possible to realize the process of displaying the first display element. For example, the display content of the display element for all important words "equipment control unit" in the revised complaint information can be replaced with the character string, etc. input by the user. It becomes possible to realize processing such as

When each display element of the revised claim information part in the display document is displayed by a browser, the user can copy and paste the displayed content (for example, multi-line content) to create and edit the document. It is configured so that it can be incorporated into an application program (for example, a text editor, etc.) having a function and used as a claim for a patent application or a Japanese manuscript for translation of claims for a patent application. For example, by operating an application program having a document creation and editing function, the user can freely edit part or all of the content of the imported revised claim information portion or output it to a file. In addition, through operations such as copying and pasting, editing, and outputting, the user can transfer part or all of the contents of the revised claim information section or the editing results to a patent application application program (e.g., Internet application software). can be imported into. In addition, the user can obtain the invention document including the edited version of the revised claim information, etc. by the invention document acquisition unit 2021 of the invention document analysis system 20, and analyze it again. Application-related information is available.

The claim information generation unit 205a controls the display mode (for example, for the heading or important word/phrase) on the display element corresponding to the heading or important word/phrase included in the revised claim information portion of the display document. Display control information for highlighting (highlighting, colored display, display transition, etc.) is added. The display control information is the same as, for example, the display control information added to the display element of the claim information portion of the patent application related information as the display document.

The claim information generation unit 205a uses a modified version of the claim information portion of the display document as a display element, which is a description part modified from the claim of the invention document, and displays the display element in a display mode that emphasizes the display element by adding an underline or the like. For example, display control information for adjusting text-decoration property, border-bottom property, etc. as display attributes may be added to the display element so as to display it. In the revised claim information in the display document, for the description part that has been corrected by deleting it from the claims in the invention document, the border of the display attribute in the display element composed of <span></span> tags that do not contain any character strings. It is also possible to add display control information such as adjusting properties or the like to display a border line with a different width. In the revised version of the claim information, the description part modified from the claim of the invention document is not necessarily modified appropriately, so the modified description part should be displayed in a display mode that emphasizes it. This makes it relatively easy for the user to refer to the warning information and complaint information, check whether the modifications in the revised complaint information are appropriate, and then utilize the contents of the revised complaint information.

[3-1-5-8. Statement information generation section]
The specification information generation unit 205b executes a specification information generation process to generate specification information included in the patent application related information outputted by the output unit 208 of the invention document analysis system 20. The specification information is information regarding the specification in the invention document.

The specification information generation unit 205b may execute, for example, a specification heading list generation process that generates a specification heading list included in the patent application related information output by the output unit 208 of the invention document analysis system 20. The specification heading list is a list listing various headings in the specification in the invention document.

For example, the specification information generation unit 205b generates revised specification information that is a revised specification of an invention document included in the patent application related information output by the invention document analysis system 20 by the output unit 208. document information generation processing.

If the invention document acquired by the invention document acquisition unit 2021 does not include a specification, the specification information generation unit 205b performs a specification information generation process, a specification heading list generation process, and revised specification information. Skip the generation process.

The specification information generation unit 205b can execute, for example, a code list generation process that generates a code list included in the patent application related information output by the output unit 208 of the invention document analysis system 20.

[3-1-5-8-1. Generation of statement information]
The specification information generation unit 205b realizes the specification information generation process by generating specification information representing the specification included in the invention document acquired by the invention document acquisition unit 2021. The specification information may be a part of the specification included in the invention document acquired by the invention document acquisition unit 2021. An example of this modification is when the statement includes a predetermined control code such as "0x07" for bell control, converting the predetermined control code to a predetermined character such as "●". .

In the invention document analysis system 20, when the output unit 208 outputs patent application related information including specification information as a display document (for example, an HTML document), the specification information generation unit 205b outputs patent application related information including specification information as a display document. Generate specification information included in patent application related information.

The specification information portion of the patent application related information as a display document generated by the specification information generation unit 205b includes, for example, one or more display elements corresponding to the contents of each line of the specification, and includes, for example, one or more display elements corresponding to the contents of each line of the specification. If there is one or more important words extracted by the important word extraction unit 2051, it contains one or more display elements corresponding to each important word, and in each line, "Name of the invention", "Invention Headings such as ``Form for Implementation'', paragraph numbers, etc. (for example, descriptions with opening and closing corner brackets added), or if comments are included, the headings, or Contains one or more display elements corresponding to comments. As a specific example, when the specification information is written between an <article> tag and a </article> tag in an HTML document, the specification information generation unit 205b writes each line of the specification under the <article> tag. , write the display element corresponding to that line by sandwiching the contents of that line between the <span> tag and </span> tag, and write it under the <span> tag of the display element corresponding to that one line. , for each important term in that line, describe the display element corresponding to that important term by sandwiching the important term between <span> tag and </span> tag, and add the display element to the important term in that line. Regarding a code, a display element corresponding to the code is described by sandwiching the code between a <span> tag and a </span> tag. For example, a browser may interpret a document for display and display each display element. Each display element corresponding to each line included in the specification information portion allows, for example, to display the contents of the corresponding line.

The specification information generation unit 205b corresponds to headings, comments, important words, drawing numbers (for example, "Figure 1", "Figure 2", etc.), or codes included in the specification information part of the display document. For example, display control information for controlling the display mode (e.g., highlighting, colored display, display transition, etc.) of the heading, comment, important phrase, drawing number description, or code is added to the display element. . The specification information generation unit 205b extracts, for example, an extraction range of important words from among the important words and codes included in the specification information portion according to the important word extraction range information in the setting information acquired by the setting information acquisition unit 201. Display control information may be added only to important words and symbols described in (for example, "Detailed Description" and "Examples").

As a specific example of generating display control information to be added to a certain important phrase in order to display different important phrases in mutually different display modes (for example, displaying them in different colors, etc.), the statement information generation unit 205b , if the important word is written in the HTML file of the HTML document between the <span> tag and </span> tag, the <span> tag identifies the important word from other important words. An example of such identification information is a class name. The specification information generation unit 205b defines a class name as identification information of an important term so that it can be distinguished whether or not the important term is stated in a claim of an invention document. Note that the specification information generation unit 205b determines whether or not an important term is stated in a claim of an invention document using the value of “first appearance claim number” as an attribute of the important term specified by the attribute specifying unit 2056. This can be determined by whether or not the value is other than 0. To give a specific example of the class name of each important term, the class names of a plurality of mutually different important terms described in the claim are "c001", "c002", "c003", etc. The class names of a plurality of mutually different important words that are not stated in the claims are "s001", "s002", "s003", etc., respectively. In such an example, the CSS file of the HTML document contains information that determines the color used to display the contents of the <span> tag for each class name, which is used as identification information for important words (in other words, information that is added to important words). For example, mutually different important words that are also stated in the claims can be distinguished from the colors added to each important word in the claim information part. Important words and phrases that are displayed in the same color and are not stated in the claims and are different from each other are displayed in different colors depending on the manner in which a different color from the important words that are also stated in the claims is added. To give a specific example, mutually different important words and phrases that are also stated in the claims in the specification information have a certain degree of width, for example, due to the border-left property, similar to the display mode in the claim information. However, mutually different important words that are not stated in the claims and are displayed in a display mode with a left border line painted in mutually different colors (that is, display colors) can be defined using the background-color property, for example. The characters are displayed in a display mode with mutually different character background colors. As a result, when the document is displayed on a browser, it becomes easy to understand whether the important words and phrases in the specification information in the display document are the important words and phrases described in the claims. Note that, for example, in a case where an important word/phrase described in a claim is highlighted by adding a left border line with a certain width, the specification information generation unit 205b may For important words that are different from each other, for example, without adding a left border with a certain width, use the background-color property to create a color that is different from the text background color of the important words stated in the claim, but is the same as the text background color of the important words. The display control information etc. may be determined so that the text is displayed in a display mode with the text background color added. Accordingly, in the specification information, the important words and phrases stated in the claims can be sufficiently emphasized by having the left border line, compared to the important words and phrases that are not described in the claims. That is, for each of the plurality of important words and phrases in the specification information, if that important word is the same as any one of the plurality of important words and phrases in the claim information that are composed of mutually different character strings, The display control information regarding the important word/phrase in the specification information has a predetermined width in which the important word/phrase is displayed in the same display color as the left border line related to the display mode of the aforementioned one important word/phrase in the claim information. If the key word in the specification information is different from any of the plurality of key words in the claim information, the key word in the description information is controlled to be displayed in a display mode with a left border added. The display control information may be for controlling the important word/phrase to be displayed in a display mode without a left border line. FIG. 26 shows an example of specification information in patent application related information as a display document displayed on a browser. In the example shown in FIG. 26, among the important words and phrases in each line of specification information in the display document, important words written in the claims are indicated by a colored left border line expressed in a pattern in the figure. , for important words and phrases that are added and not stated in the claims, the left border is not added, but the same text background color, represented by a rectangular frame in the figure, is added.

The specification information generation unit 205b performs display control in order to add a GUI function to the display elements corresponding to the important words and phrases, for example, for performing display transitions related to the important words in response to operations on the browser. Information can be added. The specification information generation unit 205b, for example, adds a function as a GUI corresponding to an operation to an important word/phrase in the same manner as the complaint information generation unit 205a (that is, makes the display element of the important word/phrase a GUI element corresponding to an operation). display control information can be generated.

To give a specific example of a display element of one important term in the specification information in the display document generated by the specification information generation unit 205b, multiple important terms of the claim "equipment control section" are described in the specification information. For example, the display element corresponding to the 11th "device control unit" in the case where ")' title="...(Information added display content)...">Device control section</span>", etc. In this example, the identification code of the important term "equipment control unit" is "c003", and the identification code of the specification information "specinfo" is used as the value of the id attribute of the <span> tag of the display element. A character string including the identification code "c003" and the number "011" in the order of description is determined. Further, in this example, when a mouse click operation is performed on a display element, the wordProc function is called with the value of the id attribute as an argument. As a result, the wordProc function can, for example, specify the display element from the argument and refer to the character string of the important word related to the display element. Extract the word identification code "c003", the listing order number "011", etc., and specify that the ID attribute of the first display element of the same important word is "specinfo-c003-001". It is now possible to realize the process of displaying the first display element. For example, it is possible to display the display contents of the display elements of all important words "equipment control section" such as claim information, revised claim information, specification information, etc. to the user. It becomes possible to implement processing for replacing input character strings, etc.

For example, the specification information generation unit 205b generates a display element corresponding to a heading such as "Detailed Description of the Invention", each paragraph number, etc. included in the specification information part, for example, for highlighting or Display control information may be added in order to add a function as a GUI for performing a display transition to a display element corresponding to the immediately preceding or following heading, for example, in response to an operation on the browser. The specification information generation unit 205b generates display control information that adds a GUI function corresponding to an operation to a heading, for example, in the same manner as the complaint information generation unit 205a.

For example, when the inappropriate description indicated by the pointing information in the warning information generated by the warning information generating unit 2059 is included in the description information part, the description information generation unit 205b deletes the inappropriate description. For example, a display element may be sandwiched between a <span> tag and a </span> tag, and display control information for highlighting inappropriate descriptions may be added to the display element. Highlighting display elements related to strings with <span> tags can be done using display controls that adjust the values of text-decoration properties, font-weight properties, border properties, color properties, background-color properties, etc. in CSS, for example. This can be achieved with information.

The specification information generation unit 205b, which generates specification information based on the specification of the invention document that has been acquired by the invention document acquisition unit 2021 and is the analysis target of the analysis unit 205, generates the specification information based on the specification of the invention document that is the subject of analysis by the analysis unit 205. (PDF file) can be acquired by the drawing document acquisition unit 2022, when the specification information includes an important term that is the same as a signed element name included in the invention document, the important term is Adding display control information to the display element of the important word to make it a display element with a GUI function for displaying the part containing the code in the drawing document (for example, the page containing the code) according to the You can also use it as Further, the specification information generation unit 205b displays a code included in the specification information, which has a GUI function for displaying a portion including the code (for example, a page including the code) in a drawing document according to an operation. The display control information to be used as an element may be added to the display element of the code in the specification information. This display control information defines, for example, the function name of a script function that is activated with a code as an argument in response to an operation, or the function itself. For example, if the drawing document is the content of a searchable PDF file, or if a PDF file containing the drawing document can be converted to a searchable PDF file, the code can be extracted from the searchable PDF file. By searching, it is determined to be specific. For example, in response to an operation on a code, the specification information generation unit 205b generates a function related to display control information for displaying a portion of the drawing document that includes the code, for the code displayed on the browser and operated on. It may be determined that the portion of the drawing document that includes the code is displayed within a range where the display element is not hidden. For example, the specification information generation unit 205b divides the display screen of the browser into a plurality of display areas, and displays the specification information display area and the display area of a part of the drawing document displayed in response to an operation in parallel. may be displayed. For example, the specification information generation unit 205b uses display control information added to a display element of a code to activate a function that is activated in response to an operation on the code, and activates a function that is activated in response to an operation on the code, and a function that is activated in response to a further operation on a plurality of functions that include the code in the drawing document. It may be specified to perform a process of switching and displaying parts (for example, two or more pages that include that code), or all parts that include that code in a drawing document may be reduced as necessary and displayed all at once. It may also be determined that the process is executed. For example, in response to an operation on a code, the specification information generation unit 205b generates a function related to display control information for displaying a part of the drawing document that includes the code, and a function that displays the code in the part of the drawing document that includes the code. It is also possible to set the display to be highlighted. FIG. 27 shows an example of a pop-up screen displayed in response to a mouse click operation on a display element of a code added to an important phrase in the specification information of patent application related information displayed on the browser. The pop-up screen illustrated in FIG. 27 is displayed when a mouse click operation is performed on the code "12" following the important phrase "B part" in the specification information, and the pop-up screen shown in FIG. ", and the reference numeral "12" in the drawing "FIG. 1" is highlighted.

To give a specific example of the display element of one code in the specification information in the display document generated by the specification information generation unit 205b, 3 in the case where multiple codes "123" are described in the specification information. The display element corresponding to the th "123" is, for example, "<span class="sign" id="specinfo-sign123-003" onclick="refSignProc("specinfo-sign123-003")'>123</span ＞” etc. In this example, "sign" is defined as the class name of the <span> tag of the display element with the third code "123" in the specification information, and the identification symbol "specinfo" of the specification information is defined as the value of the id attribute. ”, the code identification information “sign123”, and the number “003” in the order of description. Further, in this example, when a mouse click operation is performed on a display element, the refSignProc function is set to be called using the value of the id attribute as an argument. As a result, the refSignProc function can, for example, specify the display element from the argument and refer to the character string of the code related to the display element. Replace the display contents of all display elements of the code "123" in book information, etc. with character strings, etc. representing the new code input by the user, and include the code "123" as the content of "code" in the attribute information. It becomes possible to realize a process of replacing the content of the "code" of the attribute information regarding an important phrase with a new code.

For example, when the specification information portion as a display document includes a column for "Brief explanation of drawings," the specification information generation unit 205b generates a column that is sandwiched between the starting corner bracket and the ending corner bracket of that column. The description part related to the drawing number is set as one display element, and the description of the drawing with the drawing number in the specification information is displayed on that display element in response to an operation (for example, a mouse click operation) on the drawing number. Display control information may be added in order to provide a function as a GUI for searching each line included and enumerating and displaying each line of the search results. In addition, for example, when a drawing number description including "drawing n" (n is a natural number) is included in the description information portion as a display document, the specification information generation unit 205b converts the drawing number description into one As a display element, search each line that includes a drawing number description that indicates the same drawing number as the drawing number in the specification information in response to an operation for that drawing number (for example, a mouse click operation, etc.). Display control information may be added to provide a GUI function for enumerating and displaying each row of results. Further, for example, when a drawing number description is included in the specification information portion as a display document, the specification information generation unit 205b treats the drawing number description as one display element and performs an operation ( It has a function as a GUI that searches each line containing the drawing number in the specification information, regardless of the drawing number, and lists and displays each line of the search results in response to a mouse click (for example, mouse click operation, etc.). Display control information may be added to make the display more visible.

For example, the specification information generation unit 205b adds information to the written content of claim n (n is a natural number) such as "■Claim n" in the specification information part as a display document, or to the written content of claim n (n is a natural number). It is a part that can be replaced with content that has undergone predetermined conversion processing (for example, conversion processing that deletes line breaks, period characters, and citations that quote claim sections). If there is a statement indicating that the "■Claim In order to have a function as a GUI for replacing and displaying the content of an element with the content described in claim n or with the content obtained by performing the above-described predetermined conversion process on the content of the description, Display control information may also be added.

The specification information generation unit 205b generates, for example, full-width or half-width blank characters, tabs, line breaks, The character codes for return, page break, or vertical tab may be removed, and the line feed at the end of the line may be represented by the HTML tag <br> for line feed.

[3-1-5-8-2. [Generation of statement heading list]
The specification information generation unit 205b realizes the specification heading list generation process by generating a specification heading list based on the specification included in the invention document acquired by the invention document acquisition unit 2021. The specification heading list is a list in which all lines containing headings in the specification of the invention document are listed in the order of their description in the specification of the invention document.

For example, the specification information generation unit 205b treats a line containing a description sandwiched between opening and closing corner brackets as a line containing a heading description. The specification information generation unit 205b further inserts one or more predetermined types of parentheses (for example, round brackets, square brackets, angle brackets, etc.) at the beginning of a line or at a position following one or more blank characters at the beginning of a line. ) shall be treated as a line containing a heading entry, either unconditionally or on the condition that there is no period character "." at the end of the line. It can also be treated as something to think about. As a result, for example, (Embodiment 1), (Embodiment 2), [1-1. Configuration of X device], [1-2. Various headings such as [Effects, etc.] will be listed in the specification heading list, and users viewing the specification heading list in the output patent application related information will be able to view the numbers and character strings written in the various headings. It is relatively easy to check whether the information is appropriate or not.

For example, the specification information generation unit 205b extracts a line containing a heading representing a paragraph number among the lines listed in the specification heading list by the important word extraction unit 2051 within the paragraph range of the paragraph number. If there are important words/phrases that have been identified, all of the important words/phrases may be added to that line. In this case, the specification information generation unit 205b may add only one of the same important phrases so that the additional important phrases do not overlap. Further, in this case, the specification information generation unit 205b may add only the important words and phrases described in the claims of the invention document among the important words and phrases within the paragraph range. Further, if there are multiple important words to be added, the specification information generation unit 205b may add the important words in the order of importance determined by the importance determining unit 2055, or within the paragraph range. The more important words and phrases that appear in the invention document, the more important words and phrases may be placed in the front. It may be added so that it is placed at the front. Further, the specification information generation unit 205b may not include a line including a description of a heading representing a paragraph number in the lines listed in the specification heading list. For example, if the setting information acquired by the setting information acquisition unit 201 defines whether or not to list paragraph numbers in the statement heading list, the statement information generation unit 205b Accordingly, the specification heading list generation process may be executed after determining whether or not to include a line including a paragraph number in the lines listed in the specification heading list. The specification information generation unit 205b also includes drawing number descriptions in lines that are not headings but include drawing number descriptions (for example, "Figure 1", "Figure 2", etc.) in the specification heading list. The line containing the content with the exceptions removed (for example, if there are multiple drawing number entries, a line listing the drawing number entries separated by a delimiter such as a comma) is mixed with the heading line, and the appearance in the specification of the invention document is They may be listed in order. For example, if the setting information acquired by the setting information acquisition unit 201 defines whether or not to list drawing numbers in the specification heading list, the specification information generation unit 205b Depending on the information, it may be determined whether or not to include drawing numbers in the lines listed in the specification heading list, and then the specification heading list generation process may be executed.

The specification heading list can be used, for example, to check the format of various headings in the specification in an invention document, to check whether each drawing number is properly written, to check important words and phrases in each paragraph, to give an overview of the specification, etc. can be useful.

For example, the specification information generation unit 205b treats a line that includes a paragraph number between opening and closing corner brackets as not a line that includes a heading. It can also be treated as something to think about. The specification information generation unit 205b also generates a predetermined number of characters from the beginning of each line that does not include a heading description in the specification of the invention document by interweaving lines that include a heading description in the specification heading list in the order of description. (For example, each line after deleting characters other than up to 30 characters) may be included.

In the invention document analysis system 20, when the output unit 208 outputs patent application related information including the specification heading list as a display document (for example, an HTML document), the specification information generation unit 205b outputs the patent application related information including the specification heading list as a display document. A list of specification headings included in patent application related information is generated.

The specification heading list portion of the patent application related information as a display document generated by the specification information generation unit 205b includes, for example, one or more display elements corresponding to the contents of each line of the specification heading list, If each line contains a heading (for example, a description with an opening corner bracket and an ending corner bracket), one or more display elements corresponding to the heading are included. As a specific example, when writing a specification heading list between an <article> tag and a </article> tag in an HTML document, the specification information generation unit 205b creates a specification heading list under the <article> tag. For each line of the list, describe the display element corresponding to that line by sandwiching the contents of the line between <span> and </span> tags, and write the <span> of the display element corresponding to that one line. Under the tag, a display element corresponding to the heading in that line is described by sandwiching the heading between <span> tag and </span> tag. For example, a browser may interpret a document for display and display each display element. Each display element corresponding to each line included in the specification heading list portion allows, for example, to display the contents of the corresponding line.

The specification information generation unit 205b controls the display mode (e.g., highlighted display, colored display, display transition, etc.) of the heading on the display element corresponding to the heading included in the specification heading list portion of the display document. ) Adds display control information to do so. The specification information generation unit 205b generates a part of the specification information corresponding to the heading (that is, a part of the specification information in which the heading is written) in a display element corresponding to the heading, for example, in response to an operation on the browser. Display control information may be added in order to add a function as a GUI for performing a display transition to a section (representing a line of the specification of an invention document).

28 and 29 show an example of a specification heading list in patent application related information as a display document displayed on a browser. FIG. 28 shows an example in which paragraph numbers are not listed in the specification heading list. In FIG. 29, paragraph numbers are listed in a specification heading list, drawing numbers are mixed with each heading and listed in the specification heading list in the order of appearance, and each line indicating a paragraph number in the specification heading list has specification information. An example is shown in which important words and phrases included in the paragraph with that paragraph number are listed.

When an important term is included in the specification heading list portion of the display document, the specification information generation unit 205b determines a display element corresponding to the important term and displays it in the display element for emphasis or for highlighting. Display control information can be added for adding a GUI function corresponding to an operation to important words.

[3-1-5-8-3. Generation of revised statement information]
The specification information generation unit 205b realizes the revised specification information generation process by generating revised specification information in which the specification included in the invention document acquired by the invention document acquisition unit 2021 is revised. The revised specification information is a part of the specification information generated by the specification information generation process. In the modified specification information generation process by the specification information generation unit 205b, among the modifications to claims in the revised claim information generation process by the claim information generation unit 205a, modifications that are applicable to the specification of the invention document are made. (For example, corrections when the numbers of mathematical formulas, chemical formulas, etc. are not consecutive numbers, etc.) can be similarly performed. For example, the user uses the content of the revised specification information as the specification of the invention document and causes the invention document analysis system 20 to analyze the invention document again, thereby generating new warning information including the invention document evaluation information. It becomes possible to obtain patent application related information consisting of etc.

[3-1-5-8-3-1. Correction of inappropriate descriptions and deletion of unnecessary descriptions]
The specification information generation unit 205b determines a correction method in advance when the description in the specification of the invention document acquired by the invention document acquisition unit 2021 is a target for generating indication information in the warning information generation unit 2059. When specified, the description is corrected according to the correction method to generate corrected version statement information. A specific example of generating revised specification information by modifying the description of the specification will be described below.

The specification information generation unit 205b may, for example, convert the description in the specification of the invention document acquired by the invention document acquisition unit 2021 into the invention document standard provisions for translation (Fig. If the pattern information of a condition item to which correction information has been added in (see 14) corresponds to the pattern indicated, the description in the specification is corrected based on the correction information, thereby making an amendment that represents the corrected specification. It is also possible to generate version specification information. As a specific example, the specification information generation unit 205b may, for example, add the following information to the specification of the invention document: “U.S. If there is a statement such as ``Characteristics of the present invention'' that corresponds to the pattern shown in the pattern information in the condition item with the condition item name ``Character strings that may be desirable to change,'' please include that statement in the condition item. According to the modification information, the information is modified to "The features of one embodiment of the present disclosure are", etc., and modified version specification information representing the modified specification is generated. For example, the specification information generation unit 205b may generate an invention document as an example that corresponds to a condition item related to "a character string that may be desirably changed to another expression in a U.S. application" in the invention document standard regulations for translation. In the specification of ``Figure 2 (1) is a diagram showing..., Figure 2 (2) is a figure showing...'', etc., there are multiple separate figures in one drawing. When there is a description that indicates the possibility that it exists, the description is changed according to the correction information specified in the condition item, for example, ``Figure 2A is a diagram representing..., Figure 2B is a diagram representing...'' ", etc., to generate revised version specification information representing the revised specification. For example, the specification information generation unit 205b determines whether the description in the specification of the invention document acquired by the invention document acquisition unit 2021 is “characters that are estimated to be desirably removed” according to the invention document standard regulations for translation (see FIG. 14). Search for descriptions of character strings (for example, "only", "preferably", etc.) that correspond to the pattern indicated by the pattern information for extracting inappropriate descriptions related to the condition items of "column", and search for characters that correspond to the pattern. When a column description is detected, the corrected version statement information is generated by removing that character string. Note that the content of the invention document standard provisions for translation is not limited to the content illustrated in FIG. By following the modification information of the condition item of the standard provisions for translation of invention documents that defines the condition item including the pattern information determined as applicable, for example, "preferably, if... It is also possible to generate the revised specification information by changing the statement "You may do this." to the statement "For example, you may do...".

For example, if the paragraph numbers of the specification of the invention document acquired by the invention document acquisition unit 2021 are not consecutive numbers along the line order, the specification information generation unit 205b generates consecutive four-digit numbers along the line order. By correcting the paragraph numbers of the specification so that the number becomes 4 digits, or 5 digits consecutive numbers if the total number of paragraph numbers is insufficient for 4 digits, a revised version of the specification information is generated. . For example, when half-width characters and full-width characters are mixed in the paragraph number of the specification of the invention document acquired by the invention document acquisition unit 2021, the specification information generation unit 205b generates characters of one type ( By correcting the paragraph number of the specification so that it becomes (for example, full-width characters), the revised specification information is generated.

For example, when the specification of an invention document acquired by the invention document acquisition unit 2021 does not include a paragraph number, the specification information generation unit 205b generates a paragraph number at a predetermined appropriate position and a predetermined paragraph number. The revised specification information may be generated by adding consecutive paragraph numbers to every predetermined number of lines, which is a reference for the number addition interval. The statement information generation unit 205b adds this paragraph number only when, for example, the setting information acquired by the setting information acquisition unit 201 includes information specifying the addition of a paragraph number. Good too. The information specifying the addition of paragraph numbers may include, for example, a specification of a predetermined number of lines that is a reference for the above-mentioned paragraph number addition interval. Further, the specification information generation unit 205b generates a paragraph number in the specification of the invention document when, for example, the setting information acquired by the setting information acquisition unit 201 includes information specifying deletion of a paragraph number. The revised specification information may be generated by deleting the description.

For example, when the specification of an invention document acquired by the invention document acquisition unit 2021 includes a “Brief explanation of drawings” column, the specification information generation unit 205b sorts the drawing numbers in that column in line order. If the numbers are not consecutive even after considering the branch numbers, the drawing numbers are corrected so that they are consecutive numbers in line order, thereby generating revised specification information. When the specification information generation unit 205b corrects the drawing number in the "Brief explanation of drawing" column, the specification information generation unit 205b corrects the drawing number written in the other columns of the specification (for example, the drawing that starts with "Drawing" and includes a branch number). (e.g., a part of a string with one or more consecutive characters that can be used in a number, etc.) by modifying it by character string replacement processing so that it corresponds to the modification of the drawing number in the "Brief explanation of the drawing" column. , the revised specification information may be generated. As a result, for example, a user who wants to change the drawing numbers in the specification can obtain each drawing number in the "Brief Description of Drawings" column of the specification in the invention document to be acquired by the invention document acquisition unit 2021 of the invention document analysis system 20. By simply correcting the arrangement order of each line corresponding to , it is possible to obtain revised specification information in which all drawing numbers of the specification have been corrected.

For example, when the specification of the invention document acquired by the invention document acquisition unit 2021 does not include a column of “Brief description of drawings”, the specification information generation unit 205b generates “Diagram n” ( (n is a natural number) is not a consecutive number in the order of the first appearance of the drawing number even after taking into account the branch number, the drawing number will be consecutively numbered in the order of the first appearance of the drawing number. Correct the drawing number so that it becomes , and add the "Brief explanation of the drawing" column at the appropriate position to include the content of the drawing number entry line or the content listing the incomplete message for writing support in order of drawing number. By doing so, revised specification information may be generated. For example, if the drawing number in the drawing number description part of the specification of the invention document acquired by the invention document acquisition unit 2021 contains a mixture of half-width characters and full-width characters, the specification information generation unit 205b By correcting the drawing number to use one type of character (for example, full-width character), revised specification information is generated.

For example, if the specification of the invention document acquired by the invention document acquisition unit 2021 includes half-width characters, the specification information generation unit 205b changes the half-width characters to corresponding full-width characters. , the revised specification information may be generated. In this case, the specification information generation unit 205b detects when half-width English characters are included in the "Non-patent Literature" column in the specification of the invention document, and when half-width English characters are included between opening parentheses and closing parentheses. If a half-width alphabetic character is included, the half-width alphabetic character may not be changed to a full-width character. Furthermore, if the paragraph numbers or drawing numbers of the specification of the invention document are all uniformly half-width characters, the specification information generation unit 205b may not change the numbers to full-width characters. If the specification of an invention document includes half-width characters, the setting information acquisition unit 201 determines the type of half-width characters (for example, alphabetic characters, katakana characters, symbols, etc.), character code range, etc. to be changed to full-width characters. The specification information generation unit 205b may change half-width characters in the specification to full-width characters in order to generate revised specification information in accordance with the specification, which can be specified using the acquired setting information or the like. . Conversely, when full-width characters are included in the specification of an invention document, the setting information acquired by the setting information acquisition unit 201 indicates the type of full-width characters to be changed to half-width characters, the character code range, etc. etc., and in accordance with the specification, the specification information generation unit 205b may change full-width characters in the specification to half-width characters in order to generate revised specification information.

For example, the specification information generation unit 205b may include, in the description of the invention document acquired by the invention document acquisition unit 2021, the description of claim n (n is a natural number), such as “■Claim n”, or the like. It is possible to replace the written content with content that has been subjected to predetermined conversion processing (for example, conversion processing that deletes line breaks, period characters, and citations that cite claim sections). If there is a statement indicating that it is a part, the part of "■Claim n" in this statement is added to the statement of claim n, or the above statement is added to the statement of claim n. The revised statement information may be generated by replacing the content with the content that has been subjected to a predetermined conversion process. The specification information generation unit 205b performs this replacement by, for example, replacing the description of the substitutable part to a claim with a description representing the content of the claim in the setting information acquired by the setting information acquisition unit 201. This may be performed only when the specified information is included. The specification information generation unit 205b extracts important words and phrases extracted by the important word extraction unit 2051 in the written content after the replacement, where there is a description in which a code is added to the important words in the invention document. For important words or phrases where the code corresponding to the important word is stated in the "Explanation of code" column of the specification of the invention document, insert the code corresponding to the important word immediately after the important word. You can also do it. Note that if there are multiple codes corresponding to the important term, the specification information generation unit 205b uses, for example, a list of the plural codes separated by commas etc. to insert after the important term. It is also possible to select one of the codes that comes first by sorting the plurality of codes in natural order, etc., and use that selected code to insert after the important word. good.

For example, the specification information generation unit 205b detects, in the specification of the invention document acquired by the invention document acquisition unit 2021, a description portion (that is, a so-called claim copy portion) that has substantially the same content as the content of the claim section. If included, a comment in the form of a character string that includes the item number n of the claim section in a predetermined format such as "■ (claim n)" immediately before or after that part. You may also insert . The specification information generation unit 205b inserts this comment, for example, when the setting information acquired by the setting information acquisition unit 201 includes information specifying insertion of a comment related to the item number to the claim copy portion. This may be done only in certain cases. Note that the information that may be included in the setting information that specifies insertion of a comment related to the item number in the claim copy portion may include information that specifies the format of the comment to be inserted.

[3-1-5-8-3-2. Generation of revised specification information part in display document]
In the invention document analysis system 20, when the output unit 208 outputs patent application related information including revised specification information as a display document (for example, an HTML document), the specification information generation unit 205b outputs the patent application related information including the revised specification information as a display document The revised specification information included in the patent application related information is generated.

The revised specification information portion of the patent application related information as a display document generated by the specification information generation unit 205b includes, for example, one revised specification information portion corresponding to the content of each line of the revised specification in the revised specification information. If there is one or more important words extracted by the important word extraction unit 2051 in each line, it includes one or more display elements corresponding to each important word, and in each line, , "Title of the Invention", "Details of Carrying Out the Invention", etc., if such a heading is included, one or more display elements corresponding to the heading, and one or more symbols in each line thereof. is included, one or more display elements corresponding to each code are included. As a specific example, when the revised specification information is written between an <article> tag and a </article> tag in an HTML document, the specification information generation unit 205b creates a revised version under the <article> tag. For each line of the revised specification as specification information, describe the display element corresponding to that line by sandwiching the content of the line between <span> tag and </span> tag, and Under the <span> tag of the display element corresponding to , for each important word in that line, create a display element corresponding to that important word by sandwiching the important word between <span> tag and </span> tag. For the heading in that line, write the display element that corresponds to that heading by sandwiching the heading between <span> tag and </span> tag, and for the code in that line, write the display element that corresponds to that heading. A display element is described by sandwiching its code between a <span> tag and a </span> tag. For example, a browser may interpret a document for display and display each display element. Each display element corresponding to each line included in the revised specification information portion allows, for example, to display the contents of the corresponding line.

To give a specific example of a display element of one important term in the revised specification information in the display document generated by the specification information generation unit 205b, the important term of the claim “device control unit” is displayed in the revised specification information. ", the display element corresponding to the ninth "device control unit" is, for example, "<span class="c003" id="revspecinfo-c003-009" onclick='wordProc("revspecinfo -c003-009")' title="...(Information added display content)...">Device control section</span>", etc. In this example, the identification code of the important phrase "equipment control section" is "c003", and the identification code "revspecinfo" for the revised specification information is set as the value of the id attribute of the <span> tag of the display element. A character string including a phrase identification code "c003" and a writing order number "009" is defined. Further, in this example, when a mouse click operation is performed on a display element, the wordProc function is called with the value of the id attribute as an argument. As a result, the wordProc function can, for example, specify the display element from the argument and refer to the character string of the important word related to the display element. Extract the word identification code "c003", the listing order number "009", etc., and specify that the id attribute of the first display element of the same important word is "revspecinfo-c003-001". It is now possible to realize the process of displaying the first display element, for example, displaying the display element of all important words "equipment control section" such as claim information, revised complaint information, specification information, revised specification information, etc. It becomes possible to perform a process of replacing the displayed content with a character string or the like input by the user.

When each display element of the revised specification information part in the display document is displayed by a browser, the displayed content can be copied and pasted by the user into an application program with document creation and editing functions (for example, text The document is configured so that it can be imported into a computer (editor, etc.) and used as a specification for a patent application or a Japanese manuscript for translation of a specification for a patent application. For example, the user can freely edit part or all of the imported revised statement information part or output it to a file by operating an application program that has a document creation and editing function. . Furthermore, through operations such as copying and pasting, editing, and outputting, the user can import part or all of the contents of the revised specification information section or the editing results into the patent application application program. can. In addition, the user can cause the invention document acquisition unit 2021 of the invention document analysis system 20 to acquire the invention document including the editing results of the revised specification information and analyze it again, thereby including the warning information etc. as the analysis result. Patent application related information can be used.

For example, the specification information generation unit 205b generates a display element corresponding to a heading, a code, or an important term included in the revised specification information portion of the display document, for example, the information about the heading, code, or important term. Display control information for controlling the display mode (for example, highlighted display, colored display, display transition, etc.) is added. The display control information is the same as, for example, the display control information added to the display element of the specification information portion of the patent application related information as the display document.

The specification information generation unit 205b uses a description part corrected from the specification of the invention document, which is included in the revised specification information part of the display document, as a display element, and emphasizes the display element by adding an underline or the like. Display control information for adjusting the text-decoration property, border-bottom property, etc. as display attributes may be added to the display element so that the display element is displayed in the display mode. In the revised specification information in the display document, for the description part that has been corrected by deleting it from the specification of the invention document, the display attribute is set in the display element composed of <span></span> tags that do not contain any character strings. It is also possible to add display control information such as adjusting the border property or the like to display a border line with a different width. In the revised specification information, the description part that has been modified from the specification of the invention document is not necessarily modified appropriately, so the revised description part is displayed in a display format that emphasizes the modified description part. By doing so, users can refer to the warning information and specification information, check whether the corrections in the revised specification information are appropriate, and then utilize the contents of the revised specification information. It becomes easier to target.

[3-1-5-8-4. Generate code list]
The specification information generation unit 205b realizes the code list generation process by generating a code list based on the specification etc. included in the invention document acquired by the invention document acquisition unit 2021.

The code list lists the codes written following the element name represented by the key word extracted from the specification of the invention document by the key word extraction unit 2051, and represents the code in association with the corresponding key word. It is a list. The specification information generation unit 205b generates a code list in a format that can be written in the "code explanation" column of a specification for a patent application in Japan, for example. The specification information generation unit 205b creates a code list in which, for example, one important word/phrase and one or more codes corresponding to the important word/phrase are arranged in the order of the one or more codes and the important word/phrase. Write with blank characters in between. When describing a plurality of codes corresponding to one important term in one line of the code list, the specification information generation unit 205b, for example, separates the plurality of codes by comma characters ",", delimiters such as commas, etc. They can be separated and listed in an order according to the result of sorting, such as natural order. The delimiter in the code list may be settable using the setting information acquired by the setting information acquisition unit 201. For example, the specification information generation unit 205b arranges each row of the code list in the row order according to the result of sorting, such as natural order, for the first code in the row.

For example, in a case where a description in which a code is added to the key word extracted by the key word extractor 2051 does not exist in the specification of the invention document, the specification information generation unit 205b adds a code to the key word or phrase extracted by the key word/phrase extraction unit 2051. When such a description exists in a claim of an invention document, the code added to the key word/phrase in the claim can be associated with the key word/phrase and included in the code list.

For example, when there is no description in the invention document in which a code is added to the key phrase extracted by the key phrase extraction unit 2051, the specification information generation unit 205b determines that the element represented by the key phrase is different from other elements. When included in the structure or included in another element, the code determined by the code determination process by the claim information generation unit 205a is associated with the important word and phrase and included in the code list. obtain.

For example, the specification information generation unit 205b identifies each important word associated with a code in the code list by referring to the “first claim number” specified by the attribute identification unit 2056 as an attribute of the important word. A code list may be generated to identifiably express whether a word or phrase is an important word or phrase described in a claim of an invention document.

For example, the specification information generation unit 205b may add a corresponding translation word or a set of translation word candidates for each important phrase associated with a code in the code list.

When the output unit 208 in the invention document analysis system 20 outputs patent application related information including a code list as a display document (for example, an HTML document), the specification information generation unit 205b outputs patent application related information including a code list as a display document (for example, an HTML document). Generate a code list included in application-related information.

The code list portion of the patent application related information as a display document generated by the specification information generation unit 205b includes, for example, one or more display elements corresponding to the contents of each line of the code list, and It includes one or more display elements corresponding to each code and key phrase. As a specific example, when a code list is written between an <article> tag and a </article> tag in an HTML document, the specification information generation unit 205b writes information for each line of the code list under the <article> tag. , describe the display element corresponding to that line by sandwiching the contents of that line between the <span> tag and </span> tag, and under the <span> tag of the display element corresponding to that one line, For each code or important phrase in the line, a display element corresponding to the code or important phrase is described by sandwiching the code or important phrase between <span> tag and </span> tag. For example, a browser may interpret a document for display and display each display element. Each display element corresponding to each line included in the code list portion allows, for example, to display the contents of its corresponding line. A part of FIG. 30 shows an example of a code list in patent application related information as a display document displayed on a browser. Note that FIG. 30 includes an example of a drawing code list to be described later.

The specification information generation unit 205b controls the display mode (e.g., highlighting, coloring, Display control information for display, display transition, etc.) is added.

For example, the specification information generation unit 205b adds display control information to display elements corresponding to important terms in the code list part of the display document to display elements corresponding to important terms in the specification information part of the display document. It can be generated in the same way as the additional display control information.

For example, the specification information generation unit 205b generates claim information or modified version claim information of a display element corresponding to a code in a code list portion of a display document in response to an operation (for example, a mouse click operation) on the display element. Display control information may be added to provide a function as a GUI that displays each line containing the code and the contents of each line including the code in the specification information or revised specification information.

To give a specific example of the display element of the code in the code list in the display document generated by the specification information generation unit 205b, the display element of the code "123" written in the code list is, for example, "< span class="sign" id="signlist-sign123" onclick='refSignProc("signlist-sign123")'>123</span>', etc. In this example, "sign" is defined as the class name of the <span> tag of the display element of the code "123" in the code list, and the identification symbol "signlist" of the code list and the code of the code are set as the value of the id attribute. A character string including identification information "sign123" is defined. Further, in this example, when a mouse click operation is performed on a display element, the refSignProc function is set to be called using the value of the id attribute as an argument. As a result, the refSignProc function can, for example, specify the display element from the argument and refer to the character string of the code related to the display element. The display contents of all the display elements of the code "123" in lists, statement information, etc. are replaced with character strings etc. representing the new code input by the user, and the code "123" is replaced with the "code" of the attribute information. It becomes possible to realize a process of replacing the content of the "code" of the attribute information of the important word/phrase included as the content with a new code.

For example, the specification information generation unit 205b may display information related to the invention document in accordance with the operation of the display element corresponding to the code in the code list portion of the display document generated based on the invention document. Display control information can be added to provide a function as a GUI for displaying the portion of the drawing document acquired by the drawing document acquisition unit 2022 in which the code is expressed. For example, the specification information generation unit 205b uses display control information added to a display element of a code so that when the code is operated on the GUI, a portion of the drawing document where the code is displayed is highlighted. , may be generated. FIG. 31 shows an example of a pop-up screen displayed in response to a mouse click operation on a code display element in the code list of patent application related information displayed on the browser (see FIG. 30). The pop-up screen illustrated in FIG. 31 is displayed when a mouse click operation is performed on the code "30a" that follows the important word "processing unit" in the code list, and the drawing "FIG. 5" that includes the code "30a" is displayed. The content is ``30a'' in the drawing ``FIG. 5'', and the symbol ``30a'' is highlighted.

[3-1-5-9. Claim tree generation section]
The claim tree generation unit 205c executes a claim tree generation process that generates a claim tree that can be included in the patent application related information output by the output unit 208 of the invention document analysis system 20. A claim tree is information that expresses, in a tree structure, the citation relationship between each claim term expressed in the claim information in the patent application related information.

The claim tree generation unit 205c can execute a modified claim tree generation process that generates a modified claim tree that can be included in the patent application related information output by the output unit 208 of the invention document analysis system 20. The revised claim tree is information that expresses, in a tree structure format, the citation relationship between each claim term expressed in the revised claim information in the patent application related information.

If the invention document acquired by the invention document acquisition unit 2021 does not include a claim, the claim tree generation unit 205c omits execution of the claim tree generation process and the modified claim tree generation process.

[3-1-5-9-1. Claim tree generation]
The claim tree generation unit 205c performs the claim tree generation process by referring to the claim citation relationship information held by the claim citation relationship detection unit 2057 and generating a claim tree expressing the citation relationship of each claim item in the claim information. Realize.

The claim tree generation unit 205c rearranges the lines of each claim in the claim information in descending order of item number, and adds information indicating which item number the item cites, such as a symbol, item number, etc. A claim tree is generated by notation.

For example, the claim tree generation unit 205c generates a claim starting point (that is, a claim that does not cite other terms, or a term that cites another term but does not refer to the cited source) among each claim in the claim information. A line in the claim tree that describes the item number of a claim (a claim that is different from the claim), or a line that describes the claim subject in the claim originating claim, or a line that describes the item number of the claim originating claim. Insert a line immediately above the line, and write the words or phrases that express the claimed subject matter in the claim originating claim on the inserted line.

For example, if the line immediately before a claim in claim information representing a claim of an invention document is a comment line, the claim tree generation unit 205c converts the contents of the comment line into that line in the claim tree. Add this to the line where the item number of the section is written. Further, the claim tree generation unit 205c also generates a heading (for example, between corner brackets) of each claim section in claim information representing a claim of an invention document when the line immediately preceding that section is not a comment line. If a comment is written in a prescribed format, such as between opening and closing parentheses, on the same line as "Claim n" (where n is a natural number), the content of the comment. is added to the line that describes the item number of that item in the claim tree. The complaint tree generation unit 205c can identify comment lines and comments based on comment identification information in the configuration information acquired by the configuration information acquisition unit 201, for example.

For example, the claim tree generation unit 205c detects whether, among each claim section in claim information representing a claim of an invention document, the line immediately before that section is not a comment line, and a comment is written on the same line as the heading of that line. If there are no important words and phrases extracted by the important word extraction unit 2051 in that section of the claim of the invention document, at least one of the one or more important words is included in the claim. Add this to the line that describes the item number of that item in the tree. As an example of this case, the claim tree generation unit 205c determines whether, among the important words written in that section and extracted by the important word extraction section 2051, they are written in any other section preceding that section of the claim. An important term that is not available may be preferentially selected and added to the line in which the item number of that term in the claim tree is written. Further, in this case, as an example when there are a plurality of important words and phrases written in that section and extracted by the important words and phrases extraction section 2051, the complaint tree generation section 205c may If there is a structural inclusion relationship between the represented elements, write a description that expresses the relationship between important words (for example, write the important words related to the included side in parentheses following the important words related to the containing side). , may be added to the line that describes the item number of that item in the claim tree.

[3-1-5-9-2. Generation of revised claim tree]
The claim tree generation unit 205c performs the modified claim tree generation process by referring to the revised claim citation relationship information held by the claim citation relationship detection unit 2057, and generates a modified claim tree that expresses the citation relationship of each claim item in the revised claim information. This is achieved by generating a version claim tree. Note that when the revised claim information indicates the citation relationship of each claim item in each of a plurality of claim sets, the claim tree generation unit 205c performs the revised claim tree generation process for each claim set of the revised claim information. This is achieved by generating a modified claim tree that expresses the citation relationships of each claim term in the claim set.

The method of generating a modified claim tree showing the citation relationship of each clause of a claim of the modified version claim information in the claim tree generation unit 205c is the method of generating a claim tree showing the citation relationship of each clause of a claim of the claim information described above. It is similar to

[3-1-5-9-3. Generation of constituent claim tree and revised constituent claim tree]
When the invention document acquired by the invention document acquisition unit 2021 includes a claim, the claim tree generation unit 205c generates a claim related to claim information included in the patent application related information output by the invention document analysis system 20 by the output unit 208. A constituent claim tree related to claim information and a modified claim tree related to the revised claim information can be regarded as a modified claim tree related to the tree and the modified claim information, or as something that is further added to the patent application related information. A modified configuration claim tree may also be generated.

The claim tree generation unit 205c generates a constituent claim tree expressing the citation relationship of each claim item in the claim information, for example, with reference to the claim citation relationship information held by the claim citation relationship detection unit 2057.

The claim tree generation unit 205c rearranges the lines for each claim term in the claim information in descending order of item number, and generates information indicating which item number the item cites for use in citation-related expressions. symbol (for example, "└"), item number, etc., and when there is one or more important words written in that item and extracted by the important word extraction unit 2051, each important word is A constituent claim tree is generated by adding this to the line in which the item number of that item in the claim tree is written. In this case, if there are a plurality of important words written in that section and extracted by the important words extraction section 2051, the claim tree generation section 205c generates If there is a structural inclusion relationship, write a description that expresses the relationship between important words (for example, write the important words related to the included side in parentheses following the important words related to the included side) in the constituent claim tree. Add it to the line where the item number of that section is written.

For example, the claim tree generation unit 205c refers to the revised claim citation relationship information held by the claim citation relationship detection unit 2057, and generates a revised configuration claim tree expressing the citation relationship of each claim item in the revised claim information. do.

The claim tree generation unit 205c rearranges the lines of each claim in the revised claim information in descending order of item number, and adds information indicating which item number the item cites to the citation-related item. It is expressed using expressive symbols, item numbers, etc., and when there is one or more important words written in that item and extracted by the important word extraction unit 2051, each important word is extracted from the modified version of the constituent claim tree. By adding this to the line where the item number of that section is written, a modified version of the constituent claim tree is generated. In this case, if there are a plurality of important words written in that section and extracted by the important word extraction section 2051, the claim tree generation section 205c generates If there is a structural inclusion relationship, write a description that expresses the relationship between important words (for example, write the important words related to the included side in parentheses following the important words related to the included side) in the revised version structure. Add this to the line that describes the item number of that item in the claim tree.

[3-1-5-9-4. Generation of various claim tree parts in display document]
In the invention document analysis system 20, the output unit 208 outputs patent application related information including at least one of various claim trees such as a claim tree, a modified version claim tree, a constituent claim tree, and a modified version constituent claim tree into a display document ( For example, when outputting as an HTML document), the claim tree generation unit 205c generates various claim trees included in the patent application related information as a display document.

Various claim tree portions of patent application related information as display documents generated by the claim tree generation unit 205c include, for example, one or more display elements corresponding to the contents of each line in various claim trees (for example, constituent claim trees). , and in each line, if there is an item number, it contains one or more display elements corresponding to each item number, and if there are symbols for expressing citation relationships, one or more display elements corresponding to each symbol. If there are one or more important words extracted by the important word extraction unit 2051, one or more display elements corresponding to each important word are included. As a specific example, when describing various claim trees (for example, constituent claim trees) in an HTML document by sandwiching them between an <article> tag and a </article> tag, the claim tree generation unit 205c creates a claim tree under the <article> tag. , for each line in the various claim trees (for example, the constituent claim tree), the display element corresponding to that line is described by sandwiching the contents of the line between <span> and </span> tags, and one of the Under the <span> tag of the display element corresponding to a line, for each item number, quotation-related expression symbol, or important phrase in that line, the corresponding display element is added to the <span> tag and </span. It is described by inserting the item number, quotation-related expression symbols, or important phrases between the > tags. For example, a browser may interpret a document for display and display each display element. Each display element corresponding to each line included in each part of the various claim trees allows, for example, to display the contents of its corresponding line.

The claim tree generation unit 205c adds, for example, item numbers and citation-related expressions to display elements corresponding to the item numbers, symbols for citation-related expressions, or important phrases included in various claim tree parts in the display document. Display control information for controlling the display mode (for example, highlighting, displaying in color, display transition, etc.) of symbols or important words and phrases is added. For example, display control information added to display elements corresponding to important words in the claim tree portion of the display document may be different from display control information added to display elements corresponding to important words in the claim information portion of the display document. The same is true.

For example, the claim tree generation unit 205c generates a display element for the item number of the claim tree part or the constituent claim tree part in the display document, in response to the operation of the display element. Display control information is added so that the display element has a function as a GUI for displaying the description part of the number section. For example, the claim tree generation unit 205c generates a modification in the display document according to the operation of the display element for the item number of the modified version claim tree part or the modified version constituent claim tree part in the display document. Display control information is added so that the display element has a function as a GUI for displaying the section described in the section of the version claim information with that section number.

For example, the claim tree generation unit 205c generates a display element for a symbol for expressing a citation relationship (for example, "└") in a claim tree portion or a constituent claim tree portion of a display document by operating the display element ( Add display control information that allows the display element to have a function as a GUI for executing a citation relationship disconnection process that disconnects the citation relationship represented by the symbol in response to a mouse click (for example, a mouse click operation). . The citation relationship cutting process that is executed in response to an operation on a display element regarding a symbol for expressing the citation relationship is, for example, a process that updates the contents of the revised claim information in the display document in order to disconnect the citation relationship. and the process of updating the contents of the revised claim tree part or the revised version component claim tree part so as to be consistent with the contents of the updated revised claim information, and the process of updating the contents of the revised version claim tree part or the revised version constituent claim tree part so as to match the contents of the revised claim information after the update, and This includes a process of changing the symbol to a predetermined symbol (for example, "x") to express that the citation relationship has been severed, and redisplaying the display element. In the process of updating the contents of the revised claim information in the display document in order to disconnect this citation relationship, for example, from the citation description including the citation destination item number in the relevant section of the revised claim information to the citation destination section number. , the statement of the referenced phrase (that is, the word that follows the reference character string) in any section of the claim represented by the amended claim information precedes that statement in that section. Only when there is no inappropriate situation in which a part or its section is not mentioned in the section to which it is cited or in the section to which it is chained, delete the statement related to the section number of the cited section, and delete such a section. If an inappropriate state occurs, display that fact, suppress the execution of the citation relationship cutting process, and return the patent application related information as a display document to the state before the citation relationship disconnection process was executed. . In this citation relationship disconnection process, if the description regarding the citation destination item number is deleted from the citation statement (for example, "any one of claims 1 to 3") that includes the citation destination item number (for example, 2), the above-mentioned Whether or not an inappropriate state occurs is determined by reference phrase detection processing performed by the warning information generation unit 2059 with reference to the claim citation relationship information held by the claim citation relationship detection unit 2057 (that is, with respect to the claim section, The condition that the referenced word or phrase that is the same as the word that follows the reference string must appear in the section that precedes the reference string or in the section from which it cites. This can be determined using a method similar to the process of detecting that a description related to a reference character string that does not satisfy the above is included.

FIG. 32 shows an example of a constituent claim tree in patent application related information as a display document displayed on a browser. Further, FIG. 33 shows that in response to a mouse click operation on a display element for a symbol for expressing a citation relationship in a claim tree that constitutes patent application related information displayed on a browser, the display element indicates that the citation relationship has been severed. Show examples of changes in what is being expressed. The example in FIG. 33 shows the state after the two symbols "└" for expressing citation relationships shown in FIG. 32 are sequentially clicked with the mouse, and the two symbols " "└" has changed to a symbol "x" in FIG. 33, which indicates that the citation relationship has been severed. In the constituent claim tree illustrated in FIG. 32, the fourth term of the claim cited each of the first to third terms, but after two mouse clicks illustrated in FIG. In the constituent claim tree of , it is shown that the fourth clause of the claim cites only the first clause. In the state illustrated in FIG. 33, the 4th clause of the claim is modified so as not to cite the 2nd and 3rd clauses, and the revised claim information is updated by the citation relationship disconnection process that corresponds to the mouse click operation. There will be.

[3-1-5-10. Classification section]
The classification unit 2053 is a partial modification of the classification unit 1053 shown in Embodiment 1, and is the same as the classification unit 1053 except for points not particularly shown here.

The classification unit 2053 performs group classification for each of the plurality of invention documents acquired by the invention document acquisition unit 2021.

For example, for each of the plurality of invention documents acquired by the invention document acquisition unit 2021, the classification unit 2053 uses an important word group that is a plurality of important words extracted by the important word extraction unit 2051, and classifies each of the extracted invention documents. Based on the correlation between the important word groups, group classification is performed to specify which of the plurality of groups each of the plurality of invention documents belongs to. The classification unit 2053 weights each important term in the important term group extracted from each of the plurality of invention documents in accordance with the degree of importance determined by the importance determining unit 2055. After that, it is possible to calculate the correlation between important word groups. For example, the classification unit 2053 sets the importance determined by the importance determining unit 2055 to a predetermined value for each important word in the important word group extracted from each of the plurality of invention documents. It is also possible to calculate the correlation between groups of important words while ignoring important words at a lower value. In order to evaluate the correlation between the important word groups, the classification unit 2053 uses an algorithm that counts the number of the same important words that are common to both important word groups and a synonym dictionary to identify the same or similar words in both important word groups. For each group of important words, each character string that is assumed to be used as a sufficiently large number (for example, several thousand, tens of thousands, etc.) of important words is used as an element, for example, to count the number of important words. A vector that defines an element value greater than 0 and less than or equal to 1, which corresponds to the importance of the important word when the character string of the element exists as an important word, and an element value of 0 when it does not exist as an important word. Any algorithm may be used, such as various algorithms that define data and calculate correlation coefficients between the vector data. Furthermore, the classification unit 2053 may classify the important word groups of each invention document into groups using any of various algorithms related to clustering techniques. The classification unit 2053 classifies, for example, a sufficiently large number (for example, tens, hundreds of thousands, etc.) of inventions as each character string that is assumed to be used as a sufficiently large number (for example, thousands, tens of thousands, etc.) of important words. A document is acquired by the invention document acquisition unit 2021 of the invention document analysis system 20, and each element of a set of important words is used, which is obtained by integrating the important words extracted from each invention document by the important word extraction unit 2051 without duplication. be able to.

The classification unit 2053 includes a technical field identification unit 20531 that performs technical field classification to identify which of a plurality of technical fields each invention document belongs to, as one of the group classifications of each invention document. The classification unit 2053 is not limited to technical fields as one of the group types, and can perform various group classifications. For example, the classification unit 2053 may classify the invention documents into groups according to the department to which the inventor belongs, according to the user's designation or the like. The group classification unit, method, etc. is determined by the operational unit of the invention document analysis system 20 (for example, invention documents, bilingual dictionary information, or information output by the invention document analysis system 20, which is used by each user to communicate with each other). It may be useful for each invention document analysis system 20) to be unified.

The technical field specifying section 20531 is a partial modification of the technical field specifying section 10531 shown in Embodiment 1, and is the same as the technical field specifying section 10531 except for points not particularly shown here.

For each of the plurality of technical fields, the technical field identification unit 20531 uses the invention document acquisition unit 2021 to acquire in the past one or more invention documents known to be related to the technical field, and uses the important term extraction unit 2051 to identify important invention documents. By using the results of past extraction of word groups, the important word extraction unit 2051 can extract important words from invention documents that have been acquired by the invention document acquisition unit 2021 in multiple technical fields and whose technical field has not been specified. Any technical field in which the correlation with the key phrase group is at least a predetermined degree is specified as a technical field corresponding to an invention document whose technical field is not specified. The technical field specifying unit 20531 weights each important term in the important term group extracted from each of the plurality of invention documents in accordance with the level of importance determined by the importance determining unit 2055 for that important term. After performing this, it is possible to calculate the correlation between important word groups. As a result, elements of correlation calculated between important words given relatively high weights are evaluated relatively highly, and elements of correlation calculated between important words given relatively low weights are evaluated relatively highly. elements are rated relatively low. Therefore, for example, when a group of important words with relatively high importance in invention document A and a group of important words with relatively high importance in invention document B are almost the same, the correlation between both invention documents is as follows: For example, a group of important words with a relatively high importance in invention document A and a group of important words with a relatively high importance in invention document B are completely different, and invention document A If the group of important words with relatively low importance in invention document B and the group of important words with relatively low importance in invention document B are almost the same, the correlation between the two invention documents is evaluated to be relatively low. . For example, the technical field specifying unit 20531 determines that, for each important term in the important term group extracted from each of the plurality of invention documents, the importance determined by the importance determining unit 2055 is determined in advance for each important term. Important words and phrases with values lower than a predetermined value may be ignored, and correlations between groups of important words may be calculated. A technical field is expressed by, for example, IPC. The technical field specifying unit 20531 may use classification up to any level of the IPC as the technical field, and may treat only sections, or only sections and classes, etc. as one technical field. Further, the technical field specifying unit 20531 may express the technical field using all of the IPC sections, classes, subclasses, main groups, and subgroups. For example, if a technical field is to be distinguished only by IPC sections, the technical field identification unit 20531 uses patent publications etc. to acquire a plurality of invention documents belonging to each section of the IPC, and Each invention document is acquired by the invention document analysis system 20, and a group of important words extracted by the important word extraction unit 2051 that corresponds to the section can be obtained. For example, by using the important words and phrases for each section obtained in the past, the technical field specifying unit 20531 can extract information from an invention document in which a section as a technical field is not specified, which is obtained by the invention document obtaining unit 2021. A section whose correlation with the important word group extracted by the important word phrase extraction unit 2051 is evaluated to be at least a predetermined level can be specified as the technical field of the invention document.

The technical field identification unit 20531 identifies the technical field to which the invention document belongs based on the important word group extracted from the invention document, for example, by performing predetermined arithmetic processing using a technical field identification learning model. It may also be a thing. In addition, the technical field identification unit 20531 may, for example, collect data corresponding to a group of items (that is, a group of headwords) of one or more existing technical term dictionaries for each of a plurality of technical fields, and important words and phrases extracted from the invention document. The technical field to which the invention document belongs may be identified by comparing the groups and counting and evaluating the number of matching words.

The technical field information, which is information indicating the technical field (for example, IPC) specified by the technical field specifying section 20531 corresponding to the invention document acquired by the invention document acquisition section 2021, is, for example, a patent output by the invention document analysis system 20. Included in application-related information.

In addition, the group specified by the classification unit 2053 to which the invention document belongs or the technical field specified corresponding to the invention document by the technical field specification unit 20531 is used for translation in the translation unit 2054, for example. , and various information such as bilingual dictionary information for each group or technical field. Furthermore, the group classification by the classification unit 2053, including the technical field classification by the technical field identification unit 20531, can be used by the machine learning unit 207 to generate a translation learning model for each group.

[3-1-5-11. Translation Department]
The translation unit 2054 receives as input the claim information and revised claim information generated by the claim information generation unit 205a, and the specification information and revised specification information generated by the specification information generation unit 205b. Through predetermined arithmetic processing for translation using a learning model, the contents of each claim section expressed in the claim information and revised claim information, as well as the content expressed in the specification information and revised specification information, are calculated. machine translation of the contents of the specification. Depending on the content of the invention document acquired by the invention document acquisition unit 2021, any one of claim information, revised claim information, specification information, and revised specification information may not be generated. Therefore, the translation unit 2054 performs machine translation on only generated claims information, revised claim information, specification information, and revised specification information as input.

A learning model for translation is a learning model for machine translation that is generated by machine learning using multiple sets of an invention document and a translation result document as a result of translation by a translator based on the invention document as training data. It is a learning model. Here, as an example, the learning model for translation will be explained as a learning model used to machine translate a Japanese invention document into an English document. It may also be a learning model used for machine translation of documents in Chinese or other languages.

In the case where a learning model for translation exists for each of a plurality of groups or technical fields, the translation unit 2054, for example, classifies the group or technical field to which the invention document acquired by the invention document acquisition unit 2021 belongs to the classification unit 2053 or the technical field. Claim information and revised claim information generated based on the invention document by selectively using a translation learning model that corresponds to the group or technical field to which the invention document belongs, which is specified by the field specifying unit 20531. It is also possible to perform machine translation of all or part of the claims expressed in , and the specification expressed in the specification information and revised specification information. In cases where the learning model for translation includes a learning model for translation of important words, machine translation is performed by calculating each word in the sentence using the learning model for translation of important words and phrases. , and the result is subjected to calculation processing using a translation learning model (for example, calculation processing that includes calculations according to the grammar of the language to be translated, such as selecting either singular or plural form for the translated word) This is achieved by generating sentences in the language to be translated. Note that a learning model for translation for each group or for a certain technical field may include a learning model for important phrase translation for that group or for that technical field.

The translation unit 2054 can acquire and use translation auxiliary information, for example, based on the information for specifying the translation auxiliary information in the setting information acquired by the setting information acquisition unit 201.

For example, the translation unit 2054 acquires translation auxiliary information corresponding to another invention document that is somewhat related to the invention document acquired by the invention document acquisition unit 2021, and acquires the translation auxiliary information by the invention document acquisition unit 2021. It may also be used for machine translation of claim information, revised version claim information, specification information, or revised version specification information related to invented invention documents. The degree of association between invention documents can be determined, for example, by determining the correlation between important word groups represented by the important word group information, based on important word group information indicating each important word extracted from each invention document by the important word extraction unit 2051. This can be determined based on the nature of the invention, the high proportion of the same important words in the important word groups, the relevance between the technical fields to which the invention documents identified from the important word groups belong, etc.

The translation unit 2054 may, for example, use translation auxiliary information generated in the past by the invention document analysis system 10 shown in Embodiment 1, translation auxiliary information generated in the past by the invention document analysis system 20, or these translation auxiliary information. acquires any of the translation auxiliary information as a result of collection and integration according to the system operation mode, and adds the acquired translation auxiliary information to the invention document acquired by the invention document acquisition unit 2021. It may also be used for machine translation of such claim information, revised claim information, specification information, or revised specification information. For example, the translation unit 2054 can acquire translation auxiliary information based on information for specifying translation auxiliary information in the setting information acquired by the setting information acquisition unit 201.

The translation unit 2054 performs predetermined arithmetic processing for translation of important words and phrases using an important word translation learning model for translation of important words and phrases, to obtain information expressed in the above-mentioned claim information and revised claim information. Executing machine translation of important words and phrases (that is, important words and phrases extracted by the important word and phrase extraction unit 2051) included in all or part of the specification represented in the claims, specification information, and revised specification information. You can also use it as Based on the results of this machine translation, corresponding translated words for the important words can be determined. In addition, the translation unit 2054 obtains the above-mentioned translation auxiliary information and uses the translation auxiliary information, or performs a corresponding translation for one or more important phrases acquired by the important phrase setting dictionary acquisition unit 203. By using an important word/phrase setting dictionary that associates words, you can specify all or part of the claims expressed in claim information and revised claim information, and the specification expressed in specification information and revised specification information. It may also be possible to determine corresponding translation words for the included important words. When the translation unit 2054 determines the corresponding translation word for the important word/phrase, the translation unit 2054 updates the important word/phrase management information held by the analysis unit 205 to reflect the determined corresponding translation word. Translation auxiliary information may be classified by group or technical field, and the translation unit 2054 selectively selects translation auxiliary information corresponding to the group or technical field to which the invention document acquired by the invention document acquisition unit 2021 belongs. It may also be used to determine corresponding translation words for important words and phrases, perform machine translation, etc.

For example, the translation unit 2054 converts important words and phrases included in each tree into the constituent claim tree and the modified version of the constituent claim tree generated by the claim tree generating unit 205c by a machine using arithmetic processing using a learning model for important term translation. A post-replacement claim tree may be included in the translation result or in the corresponding translated word based on the translation auxiliary information.

The translation unit 2054 translates all or part of the claims expressed in the above-mentioned claim information and revised version claim information, and the specification expressed in the specification information and revised version specification information. Machine translation may be executed using a translation learning model by inputting a machine translation result in which each of the included important words and phrases is replaced.

In addition, the translation unit 2054 may, for example, translate all or part of the claims expressed in the above-mentioned claim information and revised version claim information, and the specification expressed in the specification information and revised version specification information. Alternatively, machine translation is performed by replacing each important word or phrase included in a part with a machine translation result as input to a predetermined machine translation program, and obtaining the translation result output by this predetermined machine translation program. It may also be a thing. In this predetermined machine translation program, claims expressed in claim information and revised version claim information including machine translation results for each important term, and all or In machine translation of each sentence related to a part, the machine translation result for each important word is changed to the appropriate form of plural or singular depending on the content of each sentence, etc. Translation algorithms such as The predetermined machine translation program is, for example, a program that outputs a sentence as a translation result for an input Japanese sentence using existing machine translation technology, and even if it exists inside the invention document analysis system 20, May exist externally. For example, a predetermined machine translation program operates on a server device (that is, a computer) that provides a translation API, etc., and that can communicate with a terminal device that constitutes the invention document analysis system 20 through a network such as the Internet, The translation unit 2054 encrypts information input to a predetermined machine translation program and sends it to the server device, and the server device encrypts information output from the predetermined machine translation program and sends it to a terminal device. 2054 may also be receiving and decoding information output from a predetermined machine translation program. Any encryption method may be used in this encryption and decryption.

The translation learning model and important word/phrase translation learning model used by the translation unit 2054 for machine translation may be generated by any method. The translation learning model and important term translation learning model used by the translation unit 2054 for machine translation are, for example, those generated by the machine learning unit 207 in the past, or those generated by the output unit 208 by another invention document analysis system 20 in the past. It may be output from .

[3-1-5-11-1. Generation of various translation information]
The results of machine translation of the claims expressed in the claim information by the translation unit 2054 may be included in the patent application related information output by the invention document analysis system 20 as claim translation information, for example. Claim translation information includes, for example, each line of claim information, and for each line or claim section of the included claim information, the content of that line or claim section is immediately followed by the content of that line or claim section. It is also possible to insert a line representing the result of machine translation corresponding to , or a line of claim. In the claim translation information, the contents of one line or one paragraph representing the result of machine translation corresponding to the contents of the line or section of the claim information is a translated line consisting of one line or multiple lines, and the translated line is a translation line that is a translation line consisting of one line or multiple lines. It may be a complete translation line that expresses the entire line or section in the translation target language, or it may be a complete translation line that expresses the entire line or section of the claim information in the target language, or it may express only the important words and phrases that can be translated included in the line or section of the claim information in the corresponding translated language. It may be a partially translated line, leaving the parts other than the important words as written in Japanese. In addition, in a complete translation line in which the entire line or section containing an important phrase of claim information is expressed in the target language, the phrase corresponding to the important phrase expressed in the target language can be treated as a corresponding translated word. Therefore, the claim translation information can also be said to be information representing the result of replacing important words and phrases included in the content of the claim information with corresponding translated words corresponding to the important words and phrases. When the translation unit 2054 includes a complete translation line in the claim translation information, the translation unit 2054 further inserts a partial translation line corresponding to the translation target line of the complete translation line immediately before or after the complete translation line. Good too. The translation unit 2054 uses the settings acquired by the setting information acquisition unit 201 to determine whether each line of complaint information is mixed in the complaint translation information or only translated lines (that is, completely translated lines or partially translated lines) are mixed. The determination may be made based on the specification in the information, and claim translation information may be generated in accordance with the determination. Note that when the claim translation information includes each line of claim information, the invention document analysis system 20 may treat the claim translation information as claim information and include it in the patent application related information.

The results of machine translation of the claims expressed in the revised claim information by the translation unit 2054 may be included in the patent application related information output by the invention document analysis system 20, for example, as the revised claim translation information. For example, the revised claim translation information includes each line of the revised claim information, and for each line of the included revised claim information, immediately after that line, a line representing the result of machine translation corresponding to that line is added. It may be inserted. In the revised claim translation information, a line representing the result of machine translation corresponding to the line of the revised claim information may be a complete translation line that represents the entire line of the revised claim information in the target language. However, it is a partial translation line in which only the important words and phrases that can be translated contained in the line of the revised claim information are expressed in the corresponding translated words, and the parts other than the important words and phrases that can be translated are left as written in Japanese. Good too. When the revised claim translation information includes a complete translation line, the translation unit 2054 further inserts a partial translation line corresponding to the translation target line of the complete translation line immediately before or after the complete translation line. It may also be a thing. The translation unit 2054 uses the setting information acquired by the setting information acquisition unit 201 to determine whether each line of the revised claim information is to be mixed in the revised claim translation information, or not to be mixed, and only to be completely translated lines or partially translated lines. The revised claim translation information may be generated in accordance with the determination. Note that when the revised claim translation information includes each line of the revised claim information, the invention document analysis system 20 treats the revised claim translation information as the revised claim information and includes it in the patent application related information. It may also be possible to do so.

The result of machine translation of the specification represented by the specification information by the translation unit 2054 may be included in the patent application related information output by the invention document analysis system 20 as specification translation information, for example. For example, specification translation information includes each line of specification information, and for each line of included specification information, a line representing the result of machine translation corresponding to that line is inserted immediately after that line. It may be. In the specification translation information, a line that represents the result of machine translation corresponding to the specification information line is a translation line, and the translation line is a complete translation of the entire specification information line in the target language. It may be a translation line, or only the important words and phrases that can be translated included in the specification information line are expressed in corresponding translation words, and the parts other than the important words and phrases that can be translated are left as written in Japanese. It may be a partially translated line. In addition, in a complete translation line in which the entire line including an important phrase of specification information is expressed in the target language, the phrase corresponding to the important phrase expressed in the target language can be treated as a corresponding translated word. Therefore, the specification translation information can also be said to be information representing the result of replacing important words and phrases included in the contents of the specification information with corresponding translated words corresponding to the important words and phrases. When including a complete translation line in the specification translation information, the translation unit 2054 further inserts a partial translation line corresponding to the translation target line of the complete translation line immediately before or after the complete translation line. You can also use it as The translation unit 2054 allows the setting information acquisition unit 201 to acquire whether each line of specification information is to be mixed in the specification translation information, or whether to include only translated lines (that is, fully translated lines or partially translated lines) without mixing them. The translation information may be determined based on the specification in the setting information, and the specification translation information may be generated in accordance with the determination. In addition, when the specification translation information includes each line of specification information, the invention document analysis system 20 may treat the specification translation information as specification information and include it in the patent application related information. good.

The result of machine translation of the specification represented by the revised specification information by the translation unit 2054 may be included in the patent application related information output by the invention document analysis system 20, for example, as the revised specification translation information. . For example, the revised specification translation information includes each line of the revised specification information, and for each line of the included revised specification information, the machine translation result corresponding to that line is displayed immediately after that line. It may also be one in which a line representing the information is inserted. In the revised specification translation information, a line representing the result of machine translation corresponding to the line of the revised specification information is a complete translation line that represents the entire line of the revised specification information in the target language. Alternatively, a partial translation can be used, in which only the important words and phrases that can be translated in the lines of the revised specification information are expressed in corresponding translated words, and the parts other than the important words and phrases that can be translated are left as written in Japanese. It may be a row. When including a complete translation line in the revised specification translation information, the translation unit 2054 further inserts a partial translation line corresponding to the translation target line of the complete translation line immediately before or after the complete translation line. You can also do it. The translation unit 2054 determines whether each line of the revised specification information should be mixed in the revised specification translation information, or whether the lines should be completely translated or only partially translated without mixing them, when the setting information acquisition unit 201 has acquired the information. It may be determined based on the designation in the setting information, and the revised specification translation information may be generated in accordance with the determination. Note that when the revised specification translation information includes each line of the revised specification information, the invention document analysis system 20 treats the revised specification translation information as the revised specification information and files the patent application. It may also be included in related information.

For example, in generating a partial translation line, if a corresponding translation word corresponding to an important term in the line to be translated cannot be specified, the translation unit 2054 converts the important term in the line to be translated into a set of translation word candidates for the important term. The replaced line is generated as a partially translated line.

When the language of the translation destination is English, the translation unit 2054, for example, generates claim translation information, revised claim translation information, specification translation information, or revised specification translation information in the partial translation line or If the code in the complete translation line can be expressed in half-width characters, it can be written in half-width characters, and half-width blank characters can be inserted before and after the code.

When the language of the translation destination is English, the translation unit 2054, for example, generates claim translation information, revised claim translation information, specification translation information, or partial translation line or complete translation in the revised specification translation information. If there is a blank character at the beginning of the line to be translated, that blank character is replaced with a predetermined number of half-width blank characters and included at the beginning of the line in the partial translation line or complete translation line. It can be written as follows. The predetermined number of half-width blank characters that can be written at the beginning of this partial translation line or complete translation line may be determined by, for example, the setting information acquired by the setting information acquisition unit 201.

For example, the setting information acquired by the setting information acquisition unit 201 sets the format of the translated sentence in English (such as the number of half-width blank characters to be inserted at the end of the sentence) when the destination language is English. Alternatively, the translation unit 2054 may generate each translation in the specification translation information, revised specification translation information, etc. according to the format set in the setting information.

For example, the translation unit 2054 adds a heading ( For example, if there is a description that starts with an opening corner bracket and ends with a closing corner bracket, the heading is translated according to a predetermined bilingual table of headings to include the translation result of the heading. Alternatively, a partially translated line or a complete translated line may be generated. A bilingual table of predetermined headings (that is, a table in which predetermined translation results of the headings are associated with each of a plurality of Japanese headings) can be created, for example, based on the setting information acquired by the setting information acquisition unit 201. It may be stipulated. For example, when the target language is English, an example of a bilingual table of headings is to compare the Japanese headings (e.g. "Title of the Invention", "Mode for Carrying Out the Invention", etc.) with the Japanese headings. One example is the one that corresponds the heading with the English text of the heading for international applications based on the content agreed upon by the Trilateral Patent Offices.

The translation unit 2054 includes a translation word candidate set identification unit 20541. The translation word candidate set specifying unit 20541 is a partial modification of the translation word candidate set specifying unit 1052 shown in Embodiment 1, and is similar to the translation word candidate set specifying unit 1052 except for points not specifically mentioned here. be.

The translation word candidate set specifying unit 20541 can acquire and hold bilingual dictionary information, for example, based on the specifying information of bilingual dictionary information in the setting information. The translation word candidate set specifying unit 20541 may, for example, hold bilingual dictionary information in advance.

The translation word candidate set identification unit 20541 uses bilingual dictionary information for each of the important words extracted by the important word extraction unit 2051 from the invention document acquired by the invention document acquisition unit 2021, and determines the corresponding translation word for the important word. A translation word candidate set, which is a set of candidates, is identified. The translation word candidate set identification unit 20541 may perform various processes for each important word so that the number of elements in the translation word candidate set corresponding to the important word is one or more as much as possible, but as a result, There may be cases where the number of elements in the translation word candidate set corresponding to the important word/phrase becomes 0. Examples of these various types of processing include processing to break down an important word for which an element of the translation word candidate set is not found to be a compound word, then break it down into multiple words and synthesize bilingual translations obtained from bilingual dictionary information; If the stem of the important word for which no element is found contains a predetermined qualifier (for example, ordinal number, various numerals, ``predetermined'', etc.), the Japanese qualifier and its key word are A process of synthesizing parts other than the word stem with bilingual translations obtained from bilingual dictionary information, and adding a predetermined ending character string (e.g. "part", "means", etc.) to the end of the important word for which no element of the translation word candidate set is found. If included, the bilingual translation obtained from the bilingual dictionary information for the part other than the ending of the important word and the predetermined Japanese ending character string, or the bilingual dictionary information corresponding to the predetermined ending character string. Examples include a process of composing a bilingual translation and a bilingual translation. This predetermined ending character string may be one or more technical term-specific ending character strings indicated by technical term-specific ending information included in the setting information acquired by the setting information acquisition unit 201.

The elements of the translation word candidate set for important words and phrases identified by the translation word candidate set specifying unit 20541 may partially include Japanese words. For each important word, the translation word candidate set identification unit 20541 adds the corresponding translation word associated with the important word in the translation auxiliary information to the translation word candidate set corresponding to the important word, unless it overlaps with an existing element. It may be added as a new element. The translation word candidate set identification unit 20541 executes a predetermined calculation process using the important word translation learning model for each important word, and based on the calculation result, creates a translation word candidate set corresponding to the important word. It is also possible to specify or modify (for example, add a machine translation result generated by the calculation process). In addition, for each important term, the translation word candidate set identification unit 20541 extracts each line that includes the important term from the invention document from which the important term is to be extracted, and uses each line as input to a predetermined machine translation program. , obtain one or more bilingual translations that correspond to the translated portion of the important term among the translation results output by this predetermined machine translation program, and convert each of the bilingual translations into a set of translation word candidates corresponding to the important term. It may also be added as an element.

When the translation word candidate set specifying unit 20541 identifies or corrects the translation word candidate set for an important word, it updates the important word management information held by the analysis unit 205 to reflect the translation word candidate set. .

An example of a translation word candidate set that the translation word candidate set identification unit 20541 identifies as an important word based on bilingual dictionary information when the target language is English is the translation word for the important word "screwdriver". Candidate set ``screwdriver'', translation word candidate set ``processor, handler'' for the important word phrase ``processing part'', translation word candidate set ``judgment'' for the important word phrase ``judgment part'' that can be broken down into ``judgment'' and ``part''. ``part, decision part, adjudication part'', translation word candidate set ``variation, amount of change'' for the important phrase ``amount of change'', translation for the important phrase ``predetermined standard'' that can be broken down into ``predetermined'' and ``standard''. Word candidate set "predetermined standard, predetermined basis, predetermined criterion", translation word candidate set "support processing, backing processing, aid processing" for the important phrase "support processing" that can be broken down into "support" and "processing", important Examples of translation word candidates for the word "character" include "letter, character, writing."

The translation word candidate set specifying unit 20541 may, for example, classify the group specified by the classification unit 2053 to which the invention document belongs, or the technical field specified corresponding to the invention document by the technical field specifying unit 20531, by group or technology. It is used to select bilingual dictionary information for specifying a set of translation word candidates from among a plurality of bilingual dictionary information for each field. In each of the plurality of bilingual dictionary information, for example, the order in which it is estimated to be selected as a word corresponding to an important word in an invention document in the group or technical field corresponding to the bilingual dictionary information (that is, the corresponding The translation word candidate set specifying unit 20541 orders the words in the bilingual translation corresponding to the important word in descending order of validity as a translated word (in descending order of validity as a translated word), and the translation word candidate set specifying unit 20541 matches the order of the words in the selected bilingual dictionary information. Alternatively, the arrangement order of the elements of the translation word candidate set may be determined. When the translation word candidate set is output (for example, displayed) as part of patent application related information, the elements of the translation word candidate set may be output (for example, displayed) in that order. Each of the plurality of bilingual dictionary information by group or technical field is configured so that the more frequently used important words in that group or technical field are searched more quickly, in order to improve the search efficiency of the bilingual dictionary information. You may also do so.

An example of multiple bilingual dictionary information for each technical field when the target language is English is, for example, an element of a translation word candidate set for the important word "transplant" in bilingual dictionary information corresponding to the field of computer technology. For example, "porting" has relatively high validity, and "transplant" has relatively high validity among the elements of the translation word candidate set for the important phrase "transplant" in the bilingual dictionary information corresponding to the field of medical devices. Among the elements of the translation word candidate set for the important word ``molecule'' in the bilingual dictionary information corresponding to the fields of ``molecule'', the validity of ``molecule'' is relatively high. Among the elements of the translation word candidate set for the word "molecule", examples include "numerator" with relatively high validity.

When the target language is English, an example of multiple bilingual dictionary information grouped by the department to which the inventor belongs is, for example, the translation policy (for example, rules for translated expressions) regarding invention documents for each department. Due to differences, etc., in the elements of the translation word candidate set for the important phrases "analysis process" and "control process" in the bilingual dictionary information corresponding to the first category, the validity of "analysis process" and "control process" respectively In the bilingual dictionary information corresponding to the second category, the elements of the translation word candidate set for the important words ``analysis step'' and ``controlling step,'' respectively. In the bilingual dictionary information corresponding to the third category, the elements of the translation word candidate set for each of the important phrases "analysis process" and "control process" are set to have the highest validity. , an example that defines the validity of ``controlling'' at its highest.

[3-1-5-11-2. Generation of various translation information parts in display documents]
In the invention document analysis system 20, the output unit 208 outputs patent application related information including at least one of various types of translation information such as claim translation information, revised claim translation information, specification translation information, and revised specification translation information. When outputting as a display document (for example, an HTML document), the translation unit 2054 outputs claim translation information, revised claim translation information, specification translation information, or revised claim translation information included in the patent application related information as the display document. Generate version specification translation information.

The claim translation information, revised claim translation information, specification translation information, or revised specification translation information of the patent application related information as a display document generated by the translation unit 2054 is, for example, the claim translation information. , including one or more display elements corresponding to the content of each line in the revised claim translation information, the specification translation information, or the revised specification translation information, and in each line, the key words extracted by the important phrase extraction unit 2051 are included. If there are one or more important words or phrases, it contains one or more display elements corresponding to each important word, and if there is a machine translation result in which one or more important words are translated in each row, each machine Contains one or more display elements corresponding to translation results. As a specific example, the translation unit 2054 translates claim translation information, revised claim translation information, specification translation information, or revised specification translation information into an <article> tag and a </article> tag in an HTML document. In the case of writing between the <article> tags, for each line in the claim translation information, revised claim translation information, specification translation information, or revised specification translation information, display elements corresponding to that line. , the content of the line is written between the <span> tag and the </span> tag, and the important words or phrases in that line are written under the <span> tag of the display element corresponding to that one line. Regarding the machine translation result in which the phrase has been translated, the corresponding display element is described by sandwiching the important phrase or the machine translation result between the <span> tag and the </span> tag. For example, a browser may interpret a document for display and display each display element. Each display element corresponding to each line included in various translation information portions in a display document allows, for example, to display the contents of the corresponding line.

The translation unit 2054 displays, for example, important words and phrases included in various translation information parts of the display document, or display elements corresponding to the machine translation results into which the important words have been translated. , display control information for controlling the display mode (for example, highlighted display, colored display, display transition, etc.) is added. Regarding display control information added to display elements corresponding to important words and phrases in various translation information parts of a display document or machine translation results into which important words have been translated, for example, This is similar to the display control information added to display elements regarding words.

For example, the translation unit 2054 displays each of a plurality of important words and phrases whose character strings are different from each other in various translation information parts of the display document in different display modes (for example, in a mode in which different colors are added to each other). Display control information may be defined to be added to the display element corresponding to each important term so as to display it. In addition, the translation unit 2054 may, for example, display at least part of ( For example, display control information to be added to each display element can be determined so that the left border line (which is added to an important phrase or a machine translation result and has a certain width) is the same color. To give a specific example, the translation unit 2054 adds important words and phrases in claim translation information and specification translation information to a display element of a result of machine translation that targets important words and phrases in claim translation information. If the character strings constituting a word are the same key words as multiple different key words, at least part of the display format of the machine translation result will be the same as the key word in the claim information and specification information. The result of the machine translation so that at least a part of the display mode is the same (for example, so that the display color added to the result of machine translation of the important word is the same as the display color added to the important word) Define display control information for controlling the display mode. For example, display control information related to important words and phrases in complaint information is for controlling the important words and phrases to be displayed in a display mode with a left border line having a predetermined width. The display control information related to the machine translation results for the important words and phrases in the claim translation information and specification translation information that are to be translated is the display control information of the left border line related to the display mode of the important words and phrases. This is to control the display in a display mode in which a left border line having the same display color as the color and a predetermined width is added. As a result, in cases where the machine translation target and the machine translation result are alternately described line by line or claim term in various translation information parts of the display document, the results displayed by the browser. The user can relatively easily understand the correspondence between important words and phrases in various translation information parts and machine translation results obtained by translating the important words and phrases.

For example, the translation unit 2054 adds display control information related to a function as a GUI according to an operation to be added to display elements of important words and phrases of various translated information parts in a display document. It can be determined in the same way as the display control information related to the function as a GUI according to the operation, which is added to the display element of important words in the specification information part. In addition, the translation unit 2054 adds display control information related to the function as a GUI according to the operation, which is added to the display element of the machine translation result in which important words have been translated, in various translation information parts of the display document, for example. It can be determined in the same way as the display control information related to the function as a GUI according to the operation, which is added to the display element of the important phrase.

The translation unit 2054 provides, for example, a function as a GUI that displays important words and phrases in response to operations on the display elements of machine translation results in which important words and phrases have been translated in various translation information parts of the display document. Display control information may also be added to provide the information.

The translation unit 2054 displays the machine translation results for the important words and phrases in the display elements of the machine translation results for the important words and phrases in various translation information parts of the display document, for example, in accordance with the operation of the display elements. Display control information including translation word candidate information may be added to provide a GUI function for displaying different sets of translation word candidates. The translation word candidate set for the important phrase is specified by the translation word candidate set specifying unit 20541. The GUI function for displaying this set of translated word candidates may be to display a selection menu in which each element of the set of translated word candidates is an option. It may be displayed with an additional note, or each element may be displayed in order of relevance. The translation unit 2054 converts the content of the display element of the machine translation result into the content of the display element when an operation is performed to select one element from the set of translation word candidates for the important phrase displayed as a result of the operation on the GUI related to the display element. Display control information to be added to a display element may be determined so as to be changed to an element selected by a selection operation. The translation unit 2054 changes this machine translation result, for example, when the translation destination language is English, the selected element is changed to the same, This can be achieved by converting into singular or plural form and changing the converted element. The translation unit 2054 further uses the display control information to perform machine translation of the important word in various translation information parts of the display document when an operation is performed to select one element from the set of translation word candidates for the important word. It may be determined that the content of each display element of all machine translation results that are the same as the result is changed all at once to the element selected by the selection operation. In addition, when an operation is performed to select one element from a set of translation word candidates for an important phrase displayed as a result of an operation on the GUI related to the display element, the translation unit 2054 selects the important word and the selected word. Display control information to be added to a display element may be determined so that a pair with one element is transmitted to the machine learning unit 207 as training data for machine learning to generate a learning model for important word/phrase translation. . FIG. 34 shows an example of claim translation information in patent application related information as a display document displayed on a browser. In the example of FIG. 34, the contents of the claim and the partial translation lines corresponding to the contents are alternately displayed for each claim section. FIG. 35 shows translated words displayed in response to a mouse click operation on a display element of a translated word (for example, machine translation result) corresponding to an important word in a partial translation line of claim translation information of patent application related information displayed on a browser. An example of a selection menu for elements of a candidate set is shown. When one element is selected from the selection menu illustrated in FIG. 35, for example, the translated word/phrase for the important word/phrase in each partial translation line is replaced with the selected element and displayed.

For example, the translation unit 2054 performs machine translation in cases where machine translation targets and machine translation results are alternately written line by line or claim term in various translation information parts as a display document. At least a portion (for example, character background color) of the display mode of the display element of the target line and the display mode of the display element of the machine translation result line (for example, a partial translation line or a complete translation line) is made different, Display control information to be added to each display element can be determined. For example, the claim translation information and specification translation information include the display mode of display elements of at least a part of the machine translation target in the claim translation information and specification translation information, and at least part of the result of machine translation corresponding to the machine translation target. control so that the display mode of the display element is different at least in part (e.g., different in at least one of text background color, text color, text decoration, font size, font style, and font weight). In order to do this, display control information is added to each display element.

For example, the translation unit 2054 may display a partial translation line in a display element of various translation information portions as a display document, in response to an operation, by using a translation site (for example, a predetermined machine translation program) that can be accessed via a network such as the Internet. It is also possible to add display control information that functions as a GUI for translation by a server device that executes the .

[3-1-5-11-3. Generation of important word list]
The translation unit 2054 converts each important term extracted by the important term extracting unit 2051 from the invention document acquired by the invention document acquiring unit 2021 into a translation word candidate set specifying unit 20541 for the important term and the important term. An important word list generation process is executed to generate an important word list that is configured by associating the specified translation word candidate set (that is, the content expressed by the translation word candidate information corresponding to the important word). The important phrase list generated by the translation unit 2054 is included in the patent application related information output by the invention document analysis system 20, for example.

The translation unit 2054 generates the important phrase list as a list in which, for example, a new line is created for each important phrase, and each line describes one important phrase and a set of translation word candidates corresponding to the important phrase. For example, for an important phrase that is associated with a corresponding translation word in the important phrase setting dictionary acquired by the important phrase setting dictionary acquisition section 203, the translation unit 2054 adds a translation word candidate set to the line in which the important word is written. A corresponding translated word may be written instead. In addition, for example, for an important phrase to which a corresponding translation word is associated in the important phrase setting dictionary acquired by the important phrase setting dictionary acquisition section 203, the translation unit 2054 adds the corresponding translation to the line in which the important word is written. The word and the translation word candidate set may be written in distinguishable expressions. The translation unit 2054 limits the important words to be listed in the important word list to those for which the importance determined by the importance determination unit 2055 is higher than a predetermined degree. You can also use it as

For example, in each line of the important word list, if the number of elements of the translation word candidate set that is written in association with the important word in that line is 0, the translation unit 2054 adds the translation word candidate set to that line. Not listed. For example, in each line of the important word/phrase list, if the number of elements of the translation word candidate set that is written in association with the important word/phrase in that line is two or more, the translation unit 2054 may A validity value representing an evaluation value of the degree of validity as a translated word of the important phrase may be added to the key phrase, or the elements of the translation word candidate set may be listed in descending order of validity. This validity value is, for example, a plurality of translation auxiliary information output by the invention document analysis system 20 in the past (that is, information that associates important words and phrases with corresponding translated words selected by the user from a set of translation word candidates). It can be calculated through predetermined calculation processing based on data generated by collecting and integrating (for example, a learning model for machine translation of important phrases generated from translation auxiliary information using machine learning for machine translation of important phrases). It is. For example, the translation unit 2054 selects, among the elements of the translation word candidate set for one important phrase, the corresponding translated words for the important phrase based on data generated by collecting and integrating a plurality of pieces of translation auxiliary information. An element selected by a relatively large number of users may be associated with a relatively high value of validity.

For example, in each line of the important word/phrase list that is generated anew for each important word/phrase, the translation unit 2054 determines in advance that the importance level determined by the importance level determination unit 2055 for the important word/phrase written in that line is determined in advance. If the value is lower than the specified value, a predetermined notation (for example, a symbol "?") may be additionally written in that line to indicate that the usefulness of the important word is questionable. In each line of the important word list, the translation unit 2054 determines whether or not the important word written in that line is an important word stated in the claim, and whether the important word or phrase is included in the description of the specification. It is also possible to add a notation to make it possible to identify whether or not it is described in the form. The list of important words in this example makes it possible to understand, for example, what important words are stated in the "detailed description" of the specification in relation to the important words stated in the claims, Furthermore, it is possible to estimate important words and phrases that are not currently stated in the claims but may be stated in the claims in the future due to additions, amendments, divisions of applications, applications claiming priority, etc. Note that whether or not an important term is stated in a claim is determined by the "first appearance claim number" as an attribute of the important term specified by the attribute specifying unit 2056 among the attribute information in the important term management information held by the analysis unit 205. ", and whether or not an important word is stated in "Details for Carrying Out the Invention" can be determined by checking the "entire invention document, claim part, specification" as an attribute of the important word. It can be specified by the contents of the item "Number of parts used in parts and invention embodiment parts".

In each line of the important word phrase list, if there is a code corresponding to the important word written in that line, the translation unit 2054 adds one or more codes corresponding to the important word to that line. For example, in the specification of an invention document, if there are parts where the important words are written with codes added and parts where the important words are written without codes, the codes may be used. A notation indicating that a blank entry exists may be added to that line. Note that codes corresponding to important words or phrases, or a mixture of coded descriptions and unsigned descriptions of important words in the specification of an invention document, are determined by the attribute information in the important word management information held by the analysis unit 205. It can be specified based on the content of the "code" item as one attribute of the important word/phrase specified by the specifying unit 2056.

As determined by the inter-element relationship determination unit 2058, when the attribute information in the important term management information held by the analysis unit 205 includes the item “lower conceptual element of the element represented by the important term”, the translation unit 2054 By referring to the contents of that item, in each line of the important word list, if there is a subordinate concept element corresponding to the important word listed in that line, one or more elements corresponding to that important word are displayed in that line. It is also possible to add subordinate conceptual elements.

In each line of the important phrase list, the translation unit 2054 adds the content of one or more items as one attribute in the attribute information specified by the attribute specifying unit 2056 for the important phrase written in that line. Good too. For example, the translation unit 2054 can add, for each important phrase in the important phrase list, the contents of an item related to "number" as an attribute of the important phrase. In each line of the important phrase list, the translation unit 2054 may add the importance determined by the importance determining unit 2055 for the important phrase written in that line.

The various information that the translation unit 2054 adds to each important word in the important word list can be useful for a translator or the like to select and determine a corresponding translated word corresponding to the important word.

For example, the translation unit 2054 arranges each line of the important phrase list in a predetermined order. For example, the translation unit 2054 divides each line of the important word list into consecutive categories (for example, methods, objects, programs, etc.) to which the elements represented by the important words written in each line belong, and then divides the lines in a predetermined order. You can also arrange them by . Specific examples of the predetermined order include an example in which important words and phrases written in each line are arranged in order of importance determined by the importance determination unit 2055, and results in which important words and phrases written in each line are sorted in order of character code, etc. Examples include arranging important terms in order of the number of uses in the invention document or the number of uses in the claims of the invention document of the important words listed in each line. It's okay. The translation unit 2054 may generate, for example, a plurality of important word/phrase lists in which each line of the important word/phrase list is arranged in a different order as included in the patent application related information.

The translation unit 2054 uses the identifying information of the important term setting dictionary in the setting information acquired by the setting information acquisition unit 201 as the first line of the important term list as included in the patent application related information by the output unit 208. It is also possible to insert a predetermined heading (for example, "List of Important Words", etc.). As a result, it is possible to include the results of the user's editing of the important word list of patent application related information outputted from the invention document analysis system 20 in the invention document, and cause the invention document analysis system 20 to analyze the invention document again. can. In this case, in the invention document analysis system 20, the important word list edited by the user is treated as the important word setting dictionary, so the accuracy of extracting important words from the invention document may be improved, and the important word list after editing is treated as the important word setting dictionary. Since corresponding translation words for important words are determined based on the word list, the accuracy of machine translation of invention documents may be improved.

[3-1-5-11-4. Generation of important word list part in display document]
In the invention document analysis system 20, when the output unit 208 outputs patent application related information including a list of important words as a display document (for example, an HTML document), the translation unit 2054 outputs the patent application related information as a display document. Generate a list of important words included in the information.

The important word list part of the patent application related information as a display document generated by the translation unit 2054 includes, for example, one or more display elements corresponding to the contents of each line in the important word list, and in each line, , if there is an important word extracted by the important word extraction unit 2051, it includes a display element corresponding to that important word, and if there is a set of translation word candidates corresponding to that important word in each row, it corresponds to the set of translation word candidates. If there is a corresponding translated word corresponding to the important word/phrase in each row, a display element corresponding to the corresponding translated word is included. As a specific example, when writing an important word list between an <article> tag and a </article> tag in an HTML document, the translation unit 2054 writes each line of the important word list under the <article> tag. , write the display element corresponding to that line by sandwiching the contents of that line between the <span> tag and </span> tag, and write it under the <span> tag of the display element corresponding to that one line. , for the important word/phrase, translation word candidate set, or corresponding translation word in that line, display the corresponding display element with the <span> tag and </span> tag. Describe by putting words in between. For example, a browser may interpret a document for display and display each display element. Each display element corresponding to each line included in the important word list portion of the display document allows, for example, to display the contents of the corresponding line.

The contents of the row of the important phrase list of patent application related information generated by the translation unit 2054 include the importance of the important phrase, the number of descriptions in the "Details for carrying out the invention" in the specification, and the code. , when adding information such as the presence or absence of a mixture of descriptions without a code and descriptions with a code, the set of translation word candidates for the important word and information on the corresponding translation words should be added on the next line. may be written in. The translation unit 2054 may use multiple lines for one important phrase to represent information about that important phrase.

The translation unit 2054 adds, for example, the important words, the set of translation word candidates, or the corresponding translation to the display elements corresponding to the important words, the set of translation word candidates, or the corresponding translation words included in the list of important words in the display document. Display control information for controlling the display mode (for example, highlighting, colored display, display transition, etc.) of the word is added. Regarding display control information added to display elements corresponding to important words in the important word list part of the display document, for example, display control information added to display elements corresponding to important words in the specification information part of the display document It is similar to

To give a specific example of the display elements of the important words included in the important words list in the display document generated by the translation unit 2054, the display element corresponding to the important words "equipment control section" of the claim in the important words list is as follows. For example, “<span class=”c003” id=”keywordlist-c003” onclick=’wordProc(”keywordlist-c003”)’ title=”...(information-added display content)...”>Device control unit< /span>” etc. In this example, the identification code of the important phrase "equipment control section" is "c003", and the identification code of the important phrase list "keywordlist" and the value of the id attribute of the <span> tag of the display element are A character string including the identification code "c003" is defined. Furthermore, in this example, when a mouse click operation is performed on a display element, the wordProc function is called with the value of the id attribute as an argument. As a result, the wordProc function can, for example, specify the display element from the argument and refer to the string of important words related to the display element. Extract the word identification code "c003" and display the display contents of all the important word "device control section" display elements such as claim information, revised claim information, specification information, revised specification information, etc. It becomes possible to realize a process of replacing the character string etc. input by the user.

For example, the translation unit 2054 selectably displays each element of the translation word candidate set on a display element corresponding to the translation word candidate set including a plurality of elements, which is included in the important word list in the display document. Display control information is added to provide a function as a GUI that executes translation word candidate set selection processing corresponding to the selected element. This display control information is, for example, a function of a script that is activated in response to a user's selection operation for a set of translation word candidates, using elements of the selected translation word candidate set and important words related to the set of translation word candidates as arguments. The function name or the function itself is defined, and the function is defined, for example, by associating the elements of the argument with the important words and phrases, and selecting the important word/phrase setting dictionary (for example, This is a function for executing translation word candidate set selection processing, which includes processing to list important words and phrases and reflect them in a file (such as a file that can add corresponding translated words to the important words). The process of associating elements of the translation word candidate set with important words and reflecting them in the important word setting dictionary is, for example, when the important word is already included in the important word setting dictionary. It may include a process that allows the user to select whether or not to overwrite the information related to the important phrase. In addition, the translation word candidate set selection process executed by the function uses pairs of elements and important words that are arguments of the function as training data for machine learning to generate a learning model for important word translation. It may also include processing for transmitting information to the machine learning unit 207.

FIG. 36 shows an example of an important word list in patent application related information as a display document displayed on a browser. FIG. 37 shows an example in which a set of translation word candidates in an important word list of patent application related information displayed on a browser is changed to a corresponding translation word in response to a mouse click operation to select one element. The example shown in Figure 37 is important because the element "processor" is selected from among the elements "processor" and "handler" of the translation word candidate set corresponding to the important word "processing part" illustrated in Figure 36. A state in which "processor" has been determined as a corresponding translation word corresponding to the phrase "processing unit" is shown. For example, if the user has set the specification information of the important word/phrase setting dictionary in the setting information to use the <important word list> in the invention document as the important word/phrase setting information, By copying and pasting the important word list into the invention document file, you can include the important word list in the invention document file and edit that invention document file with a text editor, etc. It becomes possible to have the invention document analysis system 20 analyze the invention document file containing the word list again.

FIG. 38 shows another example of an important word list in patent application related information as a display document displayed on a browser. In the example of FIG. 38, the contents of each line of the important word list include an important word, the degree of importance of the important word, and the number of times the important word is described in the "Detailed Description" of the specification. , if there is a code corresponding to the important word or phrase, and information indicating whether there is a mixture of descriptions without a code and descriptions with a code for the important word by the presence or absence of the symbol "●". , and a set of translation word candidates for the important word/phrase. In this example, the importance level is expressed as a six-level value from 0 to 5 with the word "Level" attached, and the set of translation word candidates for important words that could not be retrieved from the bilingual dictionary information is omitted. Each line is arranged in alphabetical order of important words (that is, sorted in ascending order of character codes). In the example of FIG. 38, the important phrase "control program" may be a typo in the important phrase "control program." Although the possibility of such a typo may be pointed out in the warning information, for example, the important phrase list By including a list in which important words are arranged in alphabetical order in whole or in part, it may become easier to confirm such descriptions that may contain typographical errors.

For example, the translation unit 2054 has a function as a GUI that executes a process to convert important words into non-important words according to an operation on a display element corresponding to an important word included in a list of important words in a display document. Display control information may also be added for display purposes. This display control information defines, for example, the function name or the function itself of a script function that is activated with the important word as an argument in response to the user's operation on the important word. A menu is displayed that allows the user to select to make the important words of the argument into non-important words, and when the user selects to make the important words into non-important words, the important words are acquired by the important word setting dictionary acquisition unit 203. This is a function for executing a process of deleting important words from the acquired important words and phrases setting dictionary and adding them to the unimportant words and phrases setting dictionary. , may include processing for re-analyzing the invention document acquired by the invention document acquisition unit 2021.

[3-1-5-11-5. Generation of drawing code list]
The translation unit 2054 can, for example, use the content of the drawing document acquired by the drawing document acquisition unit 2022 and the attribute information about each important phrase extracted by the important phrase extraction unit 2051 from the invention document acquired by the invention document acquisition unit 2021. A drawing code list generation process is executed to generate a drawing code list based on the "code" item and the like. The drawing code list generated by the translation unit 2054 is included in the patent application related information output by the invention document analysis system 20, for example.

As a drawing code list generation process, the translation unit 2054 performs drawing code list generation processing for each drawing in the drawing document acquired by the drawing document acquisition unit 2022 from, for example, a searchable PDF file, in order of drawing number, for example, the drawing number and the number used in the drawing. A drawing code list is generated that enumerates each code, an important word/phrase corresponding to each code, and a corresponding translation word or a set of translation word candidates corresponding to the important word/phrase. Each code for each drawing in the drawing code list is written in a predetermined order (for example, an order according to a sorting result such as natural order). As a specific example, the first line of the drawing code list is ``Figure 1: 1 biological information processing system, 10 biometric sensor, 20 storage unit, 30 processing unit, . . . ", and the second line is "FIG. 2: S101 sensing step, S102 recording step, storing step, S103 processing step, . . . ".

The translation unit 2054 adds character strings written in the drawing other than the important words corresponding to the code and the character strings to the drawing code list for each drawing in the drawing document using a learning model for translation or the like. The translated results may also be added. Thereby, the drawing code list can be useful information for the translation of the drawings (that is, the translation of the explanations in the drawings).

[3-1-5-11-6. Generation of drawing code list part in display document]
In the invention document analysis system 20, when the output unit 208 outputs the patent application related information including the drawing code list as a display document (for example, an HTML document), the translation unit 2054 outputs the patent application related information as the display document. Generate a drawing code list included in the information.

The drawing code list portion of the patent application related information as a display document generated by the translation unit 2054 includes, for example, one or more display elements corresponding to the contents of each line in the drawing code list, and in each line. , if there is a drawing number, it includes a display element corresponding to that drawing number, and if there is an important term extracted by the important term extraction unit 2051 in each line, it includes a display element corresponding to the important term, and in each line, contains a display element corresponding to the corresponding translation word or translation word candidate set if there is a corresponding translation word or translation word candidate set corresponding to the important term, and if there is a code in each line, a display corresponding to that code. Contains elements. As a specific example, when a drawing code list is written between an <article> tag and a </article> tag in an HTML document, the translation unit 2054 writes each line of the drawing code list under the <article> tag. , write the display element corresponding to that line by sandwiching the contents of that line between the <span> tag and </span> tag, and write it under the <span> tag of the display element corresponding to that one line. , for the drawing number, important phrase, corresponding translation word, translation candidate set, or code in that line, the corresponding display element is displayed using the <span> tag and </span> tag. It is described by inserting a translated word, a set of translated word candidates, or a code. For example, a browser may interpret a document for display and display each display element. Each display element corresponding to each line included in the drawing code list portion of the display document allows, for example, to display the contents of the corresponding line. A part of FIG. 30 shows an example of a drawing code list in patent application related information as a display document displayed on a browser.

The translation unit 2054 adds, for example, the drawing number, important words, corresponding translation words, translation word candidate set, or display element corresponding to the code included in the drawing code list in the display document. Display control information for controlling the display mode (for example, highlighted display, colored display, display transition, etc.) of the translated word, the set of translated word candidates, or the code is added. Regarding display control information added to display elements corresponding to important words or codes in the drawing code list part of the display document, for example, display control information added to display elements corresponding to important words or codes in the code list part of the display document This is similar to display control information. Regarding the display control information added to the display element corresponding to the drawing number in the drawing code list part of the display document, for example, the display added to the display element regarding the description of the drawing number in the specification information part of the display document It is similar to control information. In addition, regarding the display control information added to the display element corresponding to the corresponding translation word or translation word candidate set in the drawing code list part in the display document, for example, the corresponding translation word or translation in the various translation information parts in the display document. This is similar to the display control information added to the display element regarding the word candidate set.

[3-1-5-11-7. Generation of translation information related to important words]
For example, based on the important word/phrase relationship information generated by the inter-element relationship determining unit 2058, the translation unit 2054 converts the important word/phrase in the important word/phrase relationship information into a corresponding translation word or translation word candidate set for the important word/phrase. Generates translated information related to replaced important words. For example, the translation unit 2054 may add a corresponding important phrase to the corresponding translation word or translation word candidate set in the important phrase related translation information. The important word/phrase related translation information generated by the translation unit 2054 is included in, for example, the patent application related information output by the invention document analysis system 20.

The important phrase relationship translation information generated by the translation unit 2054 includes inclusion relationship translation information in which the important phrases in the inclusion relationship information generated by the inter-element relationship determination unit 2058 are replaced with corresponding translation words or a set of translation word candidates. The important word/phrase relationship translation information generated by the translation unit 2054 may include, for example, record relationship translation information in which the important word/phrase in the record relationship information generated by the inter-element relationship determination unit 2058 is replaced with a corresponding translation word or a set of translation word candidates. . The important term related translation information generated by the translation unit 2054 is, for example, program realization related translation information in which important terms in the program realization related information generated by the inter-element relationship determining unit 2058 are replaced with corresponding translated words or a set of translation word candidates. may be included. The important term relationship translation information generated by the translation unit 2054 may include, for example, equivalence relationship translation information in which the important term in the equivalence relationship information generated by the inter-element relationship determination unit 2058 is replaced with a corresponding translation word or a set of translation word candidates. . The important word/phrase relation translation information generated by the translation unit 2054 is, for example, the upper/lower concept relation translation information in which the important word/phrase in the upper/lower concept relation information generated by the inter-element relation determining unit 2058 is replaced with a corresponding translation word or a set of translation word candidates. may be included.

In the invention document analysis system 20, when the output unit 208 outputs patent application related information including important word related translation information as a display document (for example, an HTML document), the translation unit 2054 outputs patent application related information including important word related translation information as a display document (for example, an HTML document). Generate translation information related to important words included in application-related information. The important word/phrase related translation information portion of the patent application related information as a display document generated by the translation unit 2054 includes, for example, display elements corresponding to each corresponding translation word or translation word candidate set for the important word/phrase. For example, a browser may interpret a document for display and display each display element. For example, the translation unit 2054 can add display control information for controlling the display mode, etc. to the display element corresponding to each corresponding translation word or translation word candidate set for the important phrase.

[3-1-5-12. Generation of aggregate information]
For example, the analysis unit 205 can generate aggregate information representing a value calculated based on the invention document acquired by the invention document acquisition unit 2021. The aggregate information generated by the analysis unit 205 may be included in the patent application related information output by the output unit 208.

The analysis unit 205, for example, expresses count results such as the number of claim terms, the number of independent claims, and the number of characters for each term in the invention document acquired by the invention document acquisition unit 2021 in a distinguishable manner, and aggregates the results. Include in information.

For example, the analysis unit 205 identifies the counting results such as the number of lines, number of sentences, number of paragraphs, total number of characters, average number of characters per sentence, etc. for the specification of the invention document acquired by the invention document acquisition unit 2021. Express it as possible and include it in the aggregate information.

The analysis unit 205 calculates, for example, the total number of words in all completely translated lines in the claim translation information or revised version claim translation information, or the total number of words in all complete translated lines in the specification translation information or revised version specification translation information. Counting results such as the total number of words and the average number of words per sentence of all completely translated lines in the specification translation information or revised specification translation information are expressed in a distinguishable manner and included in the total information.

For example, the analysis unit 205 calculates, for each important term extracted by the important term extracting unit 2051 from the claims in the invention document acquired by the invention document acquiring unit 2021, the “for carrying out the invention” in the specification of the invention document. The count result of the number of entries of the important word/phrase in the column "Format" is included in the tally information.

The analysis unit 205 analyzes the above-mentioned various information (such as The calculation results of the average values (number of claims, average number of characters per sentence in the specification, etc.) are included in the total information.

For example, the analysis unit 205 extracts important words from each invention document among a plurality of invention documents acquired by the invention document acquisition unit 2021 based on a designation by the user or a designation in the configuration information acquired by the configuration information acquisition unit 201. The correlation evaluation value (for example, the correlation coefficient calculation result) calculated based on each important word group extracted by the extraction unit 2051 is included in the total information. For example, in calculating the correlation evaluation value, the analysis unit 205 may calculate the evaluation value by weighting each important phrase in the important phrase group according to the importance determined by the importance determining unit 2055. .

For example, the analysis unit 205 analyzes important words and phrases for which corresponding translation words have been determined by a selection operation, etc., among the important words and phrases extracted by the important word extraction unit 2051 from the invention document acquired by the invention document acquisition unit 2021. Showing the corresponding translated words, each corresponding translated word determined by a selection operation etc. corresponding to the important word/phrase in another plurality of invention documents belonging to the same classification group as the invention document, and the ratio of the determination. The corresponding translated word/phrase comparison information may be included in the tally information.

[3-1-5-13. Generation of charge information]
For example, the analysis unit 205 can calculate various costs required when filing a patent application in Japan for the invention document acquired by the invention document acquisition unit 2021, and generate fee information indicating the calculation results. . The fee information generated by the analysis unit 205 may be included in the patent application related information output by the output unit 208.

For example, the analysis unit 205 includes patent application fees in the fee information. For example, the analysis unit 205 calculates the application examination request fee according to the number of claims in the invention document and includes it in the fee information. For example, the analysis unit 205 calculates the application examination request fee when the revised claim information is used in the application, which is calculated according to the number of claims in the revised claim information that represents the revised claims of the invention document. It may be added to the charge information. For example, the analysis unit 205 calculates the patent fee for each period according to the number of terms in the claims of the invention document and includes it in the fee information. For example, the analysis unit 205 calculates the patent fee for each period when the revised claim information is used in the application, which is calculated according to the number of terms in the claims in the revised claim information that is expressed by modifying the claims of the invention document. may be added to the charge information.

For example, the analysis unit 205 uses the invention document acquired by the invention document acquisition unit 2021 to determine whether the invention document is to be exported to the United States or other regions, using the number of claims based on the invention document, the number of independent claims, and other various aggregated results. The various costs required when filing a patent application may be calculated, and the calculation results may be expressed in a way that can be identified as the various costs associated with filing a patent application in Japan, and added to the fee information. .

[3-1-6. Machine Learning Department]
The machine learning unit 207 is a partial modification of the machine learning unit 107 shown in Embodiment 1, and is the same as the machine learning unit 107 except for points not particularly shown here.

The machine learning unit 207 uses the analysis results of the analysis unit 205 for each of the plurality of invention documents to generate a learning model related to the processing of the invention document by machine learning, and holds and manages the learning model. This machine learning, for example, uses a general machine learning algorithm to create an invention document or a sentence in an invention document that has a certain correlation, and a corresponding translation result document or a sentence in the translation result document. The important words and phrases are realized by arithmetic processing using a set of important words and phrases as training data, or realized by arithmetic processing using a set of important words and phrases and corresponding translation words as training data, or important words and phrases extracted from invention documents. This is realized by arithmetic processing using a set of the group and the technical field to which the invention document belongs as training data.

The learning model generated by the machine learning unit 207 is, for example, a learning model for translation for machine translation of an invention document, or a learning model for machine translation of important words extracted from an invention document by the important word extraction unit 2051. This is a learning model or a technical field identification learning model for identifying the technical field to which an invention document belongs. The learning model generated by the machine learning unit 207 is, for example, a data group that can be used for processing such as calculation, and can be implemented, for example, as a data group in a database.

The machine learning unit 207 uses, for example, various libraries for machine learning in Python (registered trademark) (for example, the scikit-learn library provided by a third party) for processing such as generating a learning model for machine learning.

The machine learning unit 207 includes a document evaluation unit 2071, a teacher data storage unit 2072, and a learning model management unit 2073.

[3-1-6-1. Generation of learning model for translation]
The machine learning unit 207 and related units will be described below, focusing on the translation learning model generated by the machine learning unit 207.

The document evaluation unit 2071 evaluates the invention document acquired by the invention document acquisition unit 2021 based on the invention document evaluation information generated as an analysis result by the analysis unit 205. For example, if the evaluation value included in the invention document evaluation information generated by the warning information generation unit 2059 for the invention document is expressed as a numerical value from 0 to 1, the document evaluation unit 2071 determines whether the evaluation value is By evaluating whether the invention document is appropriate enough to be used for generating a learning model for translation based on whether it exceeds a predetermined threshold (for example, 0.8), the document can be used to generate a learning model for translation. Select invention documents that can be used. An inappropriate invention document may hinder accurate translation, and as a result, the translation may generate a translated document with content that is somewhat different from the content of the invention document, and it may not be used as training data. It may not be suitable for use. For this reason, the machine learning unit 207 uses the selected invention documents (that is, invention documents evaluated as appropriate) by the document evaluation unit 2071 as part of training data for machine learning to generate a learning model for translation. Use as.

The teacher data storage unit 2072 assumes that the document evaluation unit 2071 operates a system that provides an invention document that has been evaluated as appropriate to the extent that it can be used to generate a learning model for translation, and a translation result document that corresponds to the invention document, for example. As such, a set of translation result documents that are assumed to correspond to the invention document acquired by the translation result document acquisition unit 2023 is stored on a recording medium, etc. as training data for machine learning to generate a learning model for translation. to be stored and retained. For example, if the invention document acquisition unit 2021 and the translation result document acquisition unit 2023 acquire an invention document that has been evaluated as appropriate and a translation result document generated by a translator based on the invention document, the teacher The data storage unit 2072 stores the combination of the invention document and the translation result document as teacher data. In addition, when the language of the translation destination is English, for example, the invention document acquisition unit 2021 acquires the international publication of the international application written in Japanese as an invention document, and the invention document is determined by the document evaluation unit 2071 to be When the international application is evaluated to be ``translated into English'' and the international application is transferred to the national phase of a designated country that uses English such as the United States, a document containing the content equivalent to the translation result document translated into English is published by the designated country's office, etc. By causing the translation result document acquisition unit 2023 to acquire the published document, the teacher data storage unit 2072 stores the combination of the invention document and the translation result document as teacher data.

When translating an inappropriate invention document, there is a possibility that the translator may correct the inappropriate descriptions in the invention document and translate it. There is a possibility of mistranslation due to not understanding the content of the part. Therefore, an inappropriate invention document and a translation result document generated by translating the invention document may not only be in different languages but also have different contents. On the other hand, there is a relatively high possibility that the invention document evaluated as appropriate by the document evaluation unit 2071 and the translation result document that is the result of its translation have the same content except for the language difference. , the combination of the invention document and the translation result document is relatively likely to be suitable as training data for machine learning.

By repeatedly using the invention document analysis system 20, a plurality of (for example, hundreds, thousands, etc.) invention documents and translation result documents are acquired, and the teacher data storage unit 2072 stores a plurality of teacher data. It can be stored on a recording medium or the like. Note that the teacher data storage unit 2072 may collect and additionally store teacher data accumulated in several other invention document analysis systems 20 in a recording medium that stores teacher data. is also possible. Further, the combination of an invention document and a translation result document as training data for machine learning may be set in units of sentences. In other words, the teacher data storage unit 2072 may store, for each sentence of the invention document, a set of that sentence and a corresponding sentence of the translation result document as the teacher data for machine learning. One sentence in an invention document can be separated by, for example, a line break, a period character ".", etc., and one sentence in an English translation result document can be separated, for example, by a line break, a period ".", etc.

For example, the teacher data storage unit 2072 writes a code corresponding to one of the important words extracted by the important word extraction unit 2051 in each of the set of invention documents and translation result documents as each teacher data. If so, the code may be removed and the removed invention document and translation result document may be stored in a recording medium or the like as training data for machine learning. This allows the machine to generate a learning model for translation in cases where codes are omitted during the translation stage, or when codes are added during the translation stage in response to the omission of codes in invention documents. It may be possible to prevent this from having a negative impact on learning.

The machine learning unit 207 uses a plurality of training data for machine learning to generate a learning model for translation, which is stored in the training data storage unit 2072, to perform machine learning for machine translation of invention documents. Generate a learning model. The translation learning model generated by the machine learning unit 207 can be used to predict a sentence as a translation result based on the sentence of the invention document, and can be used, for example, for machine translation in the translation unit 2054.

The teacher data storage unit 2072 stores sets of invention documents and translation result documents to be stored as teacher data, a set of claims of the invention document and claims of the translation result document, and a specification of the invention document and a specification of the translation result document. It is also possible to distinguish between the pair and the pair. In this case, the machine learning unit 207 generates a translation learning model for translating the claims of the invention document by machine learning using training data related to a plurality of claims accumulated in the training data storage unit 2072, A translation learning model for translating the specification of an invention document can be generated by machine learning using the teacher data related to a plurality of specifications stored in the teacher data storage unit 2072. The machine learning unit 207 may generate the translation learning model by machine learning using teacher data at any time. For example, the machine learning unit 207 performs translation learning by machine learning using each teacher data every time there is a change in the teacher data as the content of the recording area of the recording medium that stores the teacher data in the teacher data storage unit 2072. This may be done by generating a model, or by checking the fluctuations in the training data as the contents of the recording area at regular intervals, and when the fluctuations are confirmed, using machine learning using each training data to create a translation model. It is also possible to generate a learning model.

Instead of a set of claims of an invention document and a claim of a translation result document, the teacher data storage unit 2072 generates a set of claim information generated by the claim information generation unit 205a for the invention document and a claim of a translation result document. , may be stored as training data, or a set of the modified version of claim information generated by the claim information generation unit 205a for the invention document and the claims of the translation result document may be stored as training data. . The teacher data storage unit 2072 selects between the claim information and the revised claim information the claims of the translation result document based on predetermined commonality criteria such as the number of each claim and the citation relationship between each claim. It is also possible to select the one with higher commonality and store the selected one and the complaint of the translation result document as training data. In addition, the teacher data storage unit 2072 performs machine translation on the complaint information and the revised version of the complaint information using a translation learning model that has already been generated, and the difference between the machine translation result and the complaint in the translation result document is small. A set of one of the claim information and the revised version of the claim information, which served as the basis for the other machine translation, and the claims of the translation result document may be stored as training data.

Instead of the set of the specification of the invention document and the specification of the translation result document to be stored as teacher data, the teacher data storage unit 2072 stores the specification information generated by the specification information generation unit 205b for the invention document, A combination of the translation result document and the specification may be stored as training data, or the revised specification information generated by the specification information generation unit 205b for the invention document and the specification of the translation result document may be stored as training data. It is also possible to store the set of data as teacher data. The teacher data storage unit 2072 determines whether the specification information and the revised specification information are similar to the specification of the translation result document based on predetermined commonality criteria such as the total number of lines of the specification and paragraph numbers. It is also possible to select the one with higher commonality and store the selected one and the specification of the translation result document as training data. In addition, the teacher data storage unit 2072 performs machine translation on the specification information and the revised specification information using the already generated translation learning model, and compares the machine translation result with the specification of the translation result document. A set of one of the specification information and the revised specification information, which is the basis of the machine translation with the smaller difference, and the specification of the translation result document may be stored as training data.

The learning model management unit 2073 stores and manages the translation learning model generated by the machine learning unit 207 in a recording medium or the like. When the machine learning unit 207 generates a translation learning model for claim translation and a translation learning model for specification translation, the learning model management unit 2073 records them separately. Store and manage on media etc. The learning model management unit 2073 may overwrite and update the already held translation learning model with the generated translation learning model every time the translation learning model is generated by the machine learning unit 207. , a plurality of generated translation learning models (for example, a predetermined number of translation learning models in chronological order), including the latest translation learning model, may be stored in association with each generation date and time. . The learning model management unit 2073 may, for example, register and manage the translation learning model in a database. The translation learning model held by the learning model management unit 2073 can be output by the output unit 208.

[3-1-6-2. Generation of learning model for important word translation]
The machine learning unit 207 and related units will be described below, focusing on the important term translation learning model generated by the machine learning unit 207.

The teacher data storage unit 2072 stores important words and phrases with corresponding translations added in the important words and phrases setting dictionary (for example, a part of an invention document, other files, etc.) acquired by the important words and phrases setting dictionary acquisition unit 203 and their corresponding translations. The pairs of words are stored and retained in a recording medium or the like as teacher data for machine learning to generate a learning model for important word/phrase translation. For example, the teacher data storage unit 2072 acquires the translation assistance information based on the identification information of the translation assistance information in the setting information acquired by the configuration information acquisition unit 201, and identifies the important words and phrases in the translation assistance information and the corresponding translated words. The set may be stored and held in a recording medium or the like as teacher data for machine learning to generate a learning model for important word/phrase translation.

By the way, in the invention document analysis system 20, the translation word candidate set specified by the translation word candidate set specifying unit 20541 for the important words extracted by the important word extraction unit 2051 from the invention document acquired by the invention document acquisition unit 2021. A display document including the display element as a display element is generated, and the display element is displayed on a browser or the like. The translation unit 2054 adds display control information to the display elements for executing a GUI function that selects elements of the translation word candidate set according to the operation, and the user selects the translation word candidate set for important words. A selection operation of one element can be performed from . Correspondingly, the teacher data storage unit 2072 generates a display document based on the invention document selected by the document evaluation unit 2071 (that is, the invention document evaluated as appropriate) and adds it to the translation word candidate set. As a result of displaying such GUI elements, a set of an important term and one element selected from a set of translation word candidates for the important term is used for machine learning to generate a learning model for important term translation. The data can be stored and retained in a recording medium, etc., as teacher data. Note that, for example, the teacher data storage unit 2072 stores important words and phrases as a result of the display document being generated based on the invention document and GUI elements related to the translation word candidate set, regardless of the evaluation by the document evaluation unit 2071. and one element selected from the set of translation word candidates for the important word/phrase are stored in a storage medium, etc. as training data for machine learning to generate a learning model for important word/phrase translation. , may be retained. Since it is assumed that an invention document evaluated as appropriate by the document evaluation unit 2071 has a relatively high possibility of being appropriately understood, key words in the invention document evaluated as appropriate are selected from a set of translation word candidates. The selection related to the operation may be a selection with relatively high validity, and using the selection result as training data may be useful for more appropriately generating a learning model for important word/phrase translation.

Further, the teacher data storage unit 2072 translates each important term extracted by the important term extraction unit 2051 from the invention document evaluated as appropriate by the document evaluation unit 2071 based on the important term and the invention document. A machine learning teacher for generating a learning model for important word translation uses a pair of translated words and phrases corresponding to the important words extracted from the translation result document generated by the person and acquired by the translation result document acquisition unit 2023. It can be stored and retained as data on a recording medium or the like. The teacher data storage unit 2072 extracts translated words and phrases from the translation result document that correspond to important words and phrases in the invention document, based on, for example, existing syntax analysis technology. As a specific example when the language of the translated phrase is English, the sentence in a certain line of an invention document in which two important phrases are represented by A and B, respectively, is "A does B...". , it can be assumed that this sentence is translated into the SVO (that is, subject, verb, object) sentence pattern, and the sentence in the line at the corresponding position in the translation result document, where the translated words are expressed as S, V, and O, is "SVO". .'', an example is given in which it is estimated that the translated word S corresponds to the important word A, and the translated word O corresponds to the important word B. Based on such estimation results, the teacher data storage unit 2072 can extract translated words and phrases from the translation result document that correspond to important words and phrases. In addition, the teacher data storage unit 2072 extracts a translated word in the translation result document that corresponds to an important word that is followed by a code in the invention document. This can be done by searching for words that follow.

By repeatedly using the invention document analysis system 20, the teacher data storage unit 2072 accumulates a plurality (for example, thousands, tens of thousands, etc.) of pairs of important words and corresponding translated words in a recording medium, etc. as teacher data. obtain. Note that the teacher data storage unit 2072 may collect and additionally store teacher data accumulated in several other invention document analysis systems 20 in a recording medium that stores teacher data. is also possible.

The teacher data storage unit 2072 stores, for example, teacher data for machine learning for generating a learning model for important word/phrase translation, which includes the important words extracted from the invention document by the important word/phrase extracting unit 2051, when the invention document is It may be stored in a recording medium in association with a group classified by the classification unit 2053 or a technical field specified by the technical field specifying unit 20531. Thereby, the teacher data storage unit 2072 can store, on the recording medium, each piece of teacher data for machine learning for generating a learning model for important word/phrase translation, for example, for each group or technical field.

The machine learning unit 207 performs machine learning for machine translation of important words and phrases by machine learning using a plurality of training data for machine learning for generating a learning model for important words and phrases translation, which is accumulated in the teacher data storage unit 2072. Generate a learning model for important phrase translation. In the case where the training data storage unit 2072 stores a plurality of training data for machine learning for generating a learning model for important word/phrase translation in a recording medium for each group or technical field, the machine learning unit 207 , for each group or technical field, by machine learning using a plurality of training data for machine learning to generate a learning model for important word/phrase translation for each group or technical field, which is stored in the training data storage unit 2072. Generate a learning model for important phrase translation for machine translation of important phrases.

The learning model for important word/phrase translation generated by the machine learning unit 207 can be used to predict the word/phrase of the translation result corresponding to the important word/phrase of the invention document, for example, in the translation unit 2054. , it can be used to determine corresponding translation words through machine translation of important words and phrases. The machine learning unit 207 may generate the learning model for important word/phrase translation by machine learning using teacher data at any time. For example, the machine learning unit 207 detects fluctuations in the training data for machine learning for generating the learning model for important word/phrase translation, which is the content of the recording area of the recording medium that stores the training data in the training data storage unit 2072. It may be possible to generate a learning model for translating important words or phrases by machine learning using each training data, or to generate a learning model for translating important words or phrases based on the contents of the recording area at regular intervals. It is also possible to check the fluctuations in the training data for machine learning, and when the fluctuations are confirmed, to generate a learning model for important phrase translation by machine learning using each training data. In addition, the machine learning unit 207 uses a learning model for translation of important words (for example, multiple sets of important words and results of machine translation based on the learning model for translation of important words for various important words). A learning model for translation may be generated by using it as training data for machine learning for model generation. Note that the machine learning unit 207 may generate a learning model for translation that includes a learning model for important word/phrase translation.

The learning model management unit 2073 stores and manages the important term translation learning model generated by the machine learning unit 207 in a recording medium or the like. When the machine learning unit 207 generates the learning model for important word/phrase translation by distinguishing each group or technical field, the learning model management unit 2073 separates the learning model for important word/phrase translation by group or technical field. Store and manage the information on a recording medium, etc. Every time the machine learning unit 207 generates a learning model for important word/phrase translation, the learning model management unit 2073 overwrites and updates the already held learning model for important word/phrase translation with the generated learning model for important word/phrase translation. Alternatively, a plurality of generated important word/phrase translation learning models, including the latest important word/phrase translation learning model, may be created in association with each generation date and time (for example, a predetermined number of important word/phrase translations in chronological order). It is also possible to maintain a translation learning model). The learning model management unit 2073 may, for example, register and manage the learning model for important word/phrase translation in a database. The learning model for important word/phrase translation held by the learning model management unit 2073 can be output by the output unit 208.

[3-1-6-3. Generation of learning model for specifying technical field]
The machine learning unit 207 and related units will be described below, focusing on the technical field identification learning model generated by the machine learning unit 207.

The possibility that the same important words or phrases are commonly described in multiple invention documents describing inventions in the same technical field is lower than the possibility that the same important words or phrases are commonly described in multiple invention documents describing inventions in different technical fields. higher than the probability that the phrase will be written. The machine learning unit 207 performs machine learning based on the correlation between a technical field and a group of important words in an invention document. Through predetermined arithmetic processing using the technical field identification learning model, it is possible to identify the technical field to which the invention document belongs based on the group of important words extracted from the invention document. This specified technical field may be expressed in IPC or F-terms according to the F-term classification system, or may be classified up to any level of IPC (for example, classification only by sections and classes). It may also be expressed using

When the invention document acquired by the invention document acquisition unit 2021 includes technical field information such as IPC and F-term, the teacher data storage unit 2072 stores each important term extracted by the important term extraction unit 2051 from the invention document. A set of important words and phrases and technical field information such as IPC and F-terms written in the invention document is stored on a recording medium as training data for machine learning to generate a learning model for specifying a technical field. etc., and store it. Examples of cases in which an invention document includes technical field information such as IPC are: If it is a patent publication, or the publication number (for example, the publication number of a patent application) is stated in the "Patent Document" column of the prior art document in the specification of the invention document, and the publication number is based on that publication number. , cases where IPC etc. can be obtained as a result of a patent publication search via the Internet, etc.

The group of important words extracted from the invention document in the training data for machine learning for generating a learning model for specifying a technical field, which is stored in the training data storage unit 2072, is determined by the importance determining unit 2055 for each important word. It may be a data group to which the determined importance is added. A specific example of a group of important words extracted from an invention document is a predetermined number of predetermined character strings (e.g., thousands, tens of thousands, etc.) each of which is an element, and the character string of the element is an invention. A predetermined number of element values in an ordered manner, with an element value that is a degree of importance expressed as a value greater than 0 and less than or equal to 1 when it exists as an important word in the document, and an element value of 0 when it does not exist as an important word. This is vector data containing .

The machine learning unit 207 identifies a technical field from a group of important words by machine learning using a plurality of training data for machine learning to generate a learning model for identifying a technical field, which is stored in the training data storage unit 2072. Generate a learning model for important word translation.

If the important word group in the machine learning teacher data for generating the technical field specific learning model accumulated in the teacher data storage unit 2072 is data in which importance is added to each important word, the machine For example, the learning unit 207 assigns a corresponding importance level to each important word based on a plurality of training data for machine learning for generating a technical field specific learning model, which is accumulated in the training data storage unit 2072. A learning model for important phrase translation is generated by performing machine learning using multiple sets of weighted data and technical fields as training data.

The technical field identification learning model generated by the machine learning unit 207 can be used to predict the technical field from a group of important words extracted from the invention document. It can be used to identify a technical field from a group of important words related to. The machine learning unit 207 may generate the technical field identification learning model by machine learning using teacher data at any time. For example, the machine learning unit 207 detects fluctuations in the training data for machine learning for generating the technical field specific learning model, which is the content of the recording area of the recording medium that stores the training data in the training data storage unit 2072. It may be possible to generate a learning model for specifying a technical field by machine learning using each teacher data, or to generate a learning model for specifying a technical field as the contents of the recording area at a certain period. It is also possible to check the fluctuations in the training data for machine learning, and when the fluctuations are confirmed, to generate the technical field identification learning model by machine learning using each training data.

The learning model management unit 2073 stores and manages the technical field identification learning model generated by the machine learning unit 207 in a recording medium or the like. Every time the machine learning unit 207 generates a technical field specific learning model, the learning model management unit 2073 overwrites and updates the already held technical field specific learning model with the generated technical field specific learning model. Alternatively, a plurality of generated technical field specific learning models including the latest technical field specific learning model (for example, a predetermined number of technical field specific learning models in order of newest A specific learning model) may also be retained. The learning model management unit 2073 may, for example, register and manage the technical field identification learning model in a database. The technical field identification learning model held by the learning model management unit 2073 can be output by the output unit 208.

[3-1-7. Output section]
The output unit 208 is a partial modification of the output unit 108 shown in Embodiment 1, and is the same as the output unit 108 except for points not particularly shown here.

The output unit 208 outputs information generated based on the analysis result of the analysis unit 205 regarding the invention document acquired by the invention document acquisition unit 2021. The output unit 208 outputs, for example, information generated based on each of a plurality of important terms extracted by the important term extraction unit 2051 from the invention document acquired by the invention document acquisition unit 2021. The output unit 208 uses a plug-in module that performs processing to output the results of performing predetermined arithmetic processing on various analysis results (various information regarding important words, codes, etc.) generated by the analysis unit 205. You can also output it by doing so. For example, the output unit 208 provides an API that allows the plug-in module to use various analysis results generated by the analysis unit 205 in the plug-in module, and outputs information generated by the plug-in module. It is also possible to provide an API or the like for adding to a part of the patent application related information.

The output unit 208 includes a learning model output unit 2081, a patent application related information output unit 2082, a display document generation unit 2083, and a display unit 2084 that functions as a browser for displaying display documents.

[3-1-7-1. Learning model output section]
The learning model output unit 2081 outputs various learning models held and managed by the learning model management unit 2073.

For example, the learning model output unit 2081 outputs a learning model for translation, a learning model for important phrase translation, or a learning model for technical field identification to an external source of the invention document analysis system 20 in response to an instruction from a user who has specified a destination. It can be output by sending it to the specified destination. For example, the learning model output unit 2081 converts a learning model for translation, a learning model for important phrase translation, or a learning model for technical field identification into a designated learning model in response to an instruction from a user who has specified a learning model output file. It can be output as the contents of an output file.

The learning model for translation or the learning model for important phrase translation output by the learning model output unit 2081 is, for example, used in another invention document analysis system 20 or a machine translation system for translation of invention documents or a machine translation system for machine translation. It can be used in programs, general machine translation systems, etc. The technical field identification learning model output by the learning model output unit 2081 can be used, for example, in another invention document analysis system 20 or a system for identifying a technical field from an invention document.

The learning model output unit 2081 may output, for example, teacher data for machine learning that is stored in the teacher data storage unit 2072 and serves as a basis for generating various learning models. The teaching data output by the learning model output unit 2081 of the invention document analysis system 20 may be used to be stored in the teacher data storage unit 2072 of another invention document analysis system 20, or may be used for machine learning in other systems. may be used for the execution of

[3-1-7-2. Patent application related information output section]
The patent application related information output unit 2082 outputs patent application related information (see FIG. 15) generated based on the analysis result by the analysis unit 205 of the invention document acquired by the invention document acquisition unit 2021.

For example, the patent application related information output unit 2082 includes, in the patent application related information, information in which a line number is added to each line of the invention document acquired by the invention document acquisition unit 2021 and analyzed by the analysis unit 205. The line number in the information in which line numbers are added to each line of this invention document is the same as the indication in the warning information and the invention document when the indication information in the warning information includes the line number of the section of the invention document that is the target of the indication. This is useful for making it possible to understand the correspondence with the written part of the text. For example, when the patent application related information output unit 2082 includes a predetermined control code (for example, a code within the range of "0x00" to "0x1f") in the information in which a line number is added to each line of the invention document, In this case, the predetermined control code may be converted into a predetermined character such as "●".

For example, the patent application related information output unit 2082 includes warning information including invention document evaluation information in the patent application related information.

The patent application related information output unit 2082 includes, for example, an important word/phrase list, important word/phrase related information, and important word/phrase related translation information in the patent application related information. The patent application related information output unit 2082 can include, for example, inclusion related information, recording related information, program implementation related information, equivalence related information, and superordinate concept related information in the important word/phrase related information. The patent application related information output unit 2082 may include, for example, important word/phrase related translation information, inclusion related translation information, recording related translation information, program implementation related translation information, equivalence related translation information, and upper/lower concept related translation information. .

For example, the patent application related information output unit 2082 outputs claim information, claim translation information, revised claim information, revised claim translation information, claim base drawing information, claim base specification information, and claim base as patent application related information. Abstract information may be included.

For example, the patent application related information output unit 2082 can include a claim tree, a modified version claim tree, a constituent claim tree, and a modified version constituent claim tree in the patent application related information.

For example, the patent application related information output unit 2082 can include specification information, specification translation information, revised specification information, revised specification translation information, and specification heading list in the patent application related information.

The patent application related information output unit 2082 can include, for example, a code list, a drawing code list, technical field information, fee information, and total information in the patent application related information.

The patent application related information output unit 2082 outputs the patent application related information, for example, as the contents of an output file. The output file may be recorded on a recording medium or the like. The output file may be a file in a format mainly used for various application programs having document creation and editing functions such as Microsoft (registered trademark) Word, a text file, or an XML file. It may be a PDF file. Further, the output file may be a display document (for example, an HTML document) composed of one or more files. Further, the patent application related information output unit 2082 may output the patent application related information as a plurality of output files (for example, a text file and an HTML document) in different formats. The display document may include a document written in a markup language. The display document may be an HTML document whose contents are interpreted by a browser or the like, or an XML document whose contents are interpreted and displayed by a display application program (for example, one XML file or The document may be a collection of one or more files referenced from the XML file, or may be an XML document that can be converted into an HTML document using an existing conversion application program.

For example, the patent application related information output unit 2082 refers to the information for specifying the storage destination of the output file in the setting information acquired by the setting information acquisition unit 201, and specifies a predetermined folder in the file system. Output files containing patent application related information may be stored in a predetermined folder.

When generating patent application related information as a display document (for example, an HTML document), the patent application related information output unit 2082 outputs each part ( For example, the content of the important word list part, claim information part, specification information part, etc.) is transmitted to the display document generation unit 2083, and the output of patent application related information as a display document is sent to the display document generation unit 2083. entrust. The patent application related information as a display document generated by the display document generation unit 2083 can be displayed and output by a display unit 2084 having a function as a browser.

The patent application related information as a display document generated by the display document generation unit 2083 may be recorded on a recording medium, etc., as a file group as an HTML document (for example, an HTML file, a CSS file, a script file, etc.). It can be output as The output HTML document may be interpreted and displayed by the display unit 2084 of the invention document analysis system 20, or may be interpreted and displayed by a browser external to the invention document analysis system 20.

In the case where the invention document analysis system 20 has a configuration as shown in FIG. It is interpreted and displayed by an application program for realizing a browser function under the OS of the terminal device. In this case, the HTML document may be interpreted and displayed by reading a local file, which is an HTML file of the HTML document output by the display document generation unit 2083, by an application program for realizing the browser function. In addition, the display document generation unit 2083 provides an HTML document at a predetermined port on the local host used as a web server, and an application program for realizing the browser function is sent to that port on the local host. The HTML document may be obtained, interpreted, and displayed by accessing through HTTP (Hypertext Transfer Protocol) communication.

When the invention document analysis system 20 has the configuration shown in FIG. 10, for example, the patent application related information as an HTML document generated by the display document generation unit 2083 installed in the server device is stored in the server device. The information will be acquired, interpreted, and displayed by the browser on the terminal device that can communicate with the .

Note that the output of patent application related information by the patent application related information output unit 2082 is not limited to output as an output file, but may also be output to a recording medium such as a memory card that is detachable from the devices that constitute the invention document analysis system 20. Exporting patent application related information, writing patent application related information to a recording medium external to the invention document analysis system 20, transmitting a signal indicating patent application related information, displaying patent application related information on a display device, patent application related information This may include printing information, outputting audio representing patent application related information using speech synthesis technology, or causing other physical phenomena representing patent application related information. For example, when all or part of the invention document analysis processing program that realizes the functions of the invention document analysis system 20 is executed on the command prompt of Microsoft (registered trademark) Windows (registered trademark), patent application related information may be output, for example, by displaying it in a command prompt window.

The HTML document as an example of the display document generated by the display document generation unit 2083 includes an HTML file, and may further include a CSS file corresponding to the specification in the HTML file, or may further include an HTML file. It may also include a script file corresponding to the specifications in the file. When the display document generation unit 2083 generates an HTML document as an example of a display document related to patent application related information, the display document generation unit 2083 generates a display document according to each invention document to be analyzed by the invention document analysis system 20. Although the content of the HTML file that makes up the HTML document generated by the section 2083 may change, for example, the content of the CSS file and script file that make up the HTML document remain unchanged regardless of differences between invention documents. obtain.

For example, the display document generation unit 2083 generates an HTML file containing each part of patent application related information (see FIG. 15) as a display document generated by the analysis unit 205, and controls the display mode of display elements in the HTML file. A CSS file for , and a script file including functions written in JavaScript (registered trademark) are generated. For example, the display document generation unit 2083 generates a plurality of parts of patent application related information (for example, invention document, warning information, important word list, Claim information, specification information, etc.) are included in the <nav> tag and </nav> tag under the <body> tag in the HTML file that contains the descriptions sandwiched between the <article> tag and </article> tag, respectively. Write the menu between them. This menu includes, for example, menu items representing the title, name, etc. of the content under each <article> tag (for example, "Invention Document", "Warning Information", "Important Word List", "Claim Information", "Specification This is a list listing multiple display elements (such as "Information"), and it is a list that lists multiple display elements such as "Information", and is a list that lists multiple display elements such as It is configured by adding display control information that defines a function for switching the contents under each <article> tag between a hidden state and a displayed state, for example, to the display element of the menu item. In the display screen illustrated in FIG. 17, a menu written between <nav> and </nav> tags is displayed in the menu column, and the contents under the <article> tag selected according to the menu are It is displayed in the content display field of the item of patent application related information according to the menu selection. The menu may include a menu item that links to a settings information display document (for example, an HTML file as a settings information display document) that displays and edits the settings information acquired by the settings information acquisition unit 201. .

For example, the display document generation unit 2083 sets the background color to white in a CSS file, etc., as a default value of the display attribute of the content under the <article> tag corresponding to each of the plurality of parts of patent application related information in the HTML document. , the background color, font color, etc., such as black font color, etc. may be defined. For example, in a script file, etc., a predetermined user operation (for example, a GUI corresponding to a predetermined operation) when an HTML document is displayed may be defined. You may also define functions that change the background color, font color, etc., such as changing the background color from white to black and the font color from black to white, etc., in response to a mouse click on an element (mouse click operation on an element, etc.). . In this way, in display documents, using a bright font color (for example, white) against a dark background color (for example, black) or a dark font color against a light background color improves the visibility of displayed content. can be done. Furthermore, the user can switch between a pair of a dark background color and a bright font color and a pair of a bright background color and a dark font color depending on the brightness and darkness of the environment surrounding the display device in the invention document analysis system 20. can be useful for improving the visibility of displayed content.

In the display unit 2084 or other browsers that interpret and display HTML documents, important words and phrases under the <article> tag are displayed based on the default background color, font color, etc. of the contents under the <article> tag. The background color, text color, etc. defined for each display element, such as the following, may be preferentially applied.

The display document generation unit 2083 may generate a display document so that the contents under a plurality of <article> tags are displayed side by side. The content under each <article> tag in a display document can be made visible or hidden by processing functions written in a script file when the display document is interpreted and displayed by the browser. In addition, you can specify transparency, display size, display position, etc. for display. When the display document is interpreted and displayed by the browser, the display document generation unit 2083 generates the Internet by, for example, operating display elements as GUI elements that support operations for important words under various <article> tags. Processes such as searching for web pages that include the important word on a search site on a network such as Internet, etc., or translating the important word on a translation site on a network such as the Internet, etc. are written in a script file. If this is done using a function, etc., the screen related to the search site, translation site, etc. is moved to a different display area from that for displaying the display document (for example, in the display screen shown in FIG. 17). The display document may be generated so as to be controlled to be displayed in a separate window or a separate tab ("display column for items of operation-compatible GUI elements").

For example, the display document generation unit 2083 generates a display document related to patent application related information with display control information added thereto to make the display element a GUI element for re-analyzing the invention document in response to an operation. It is also possible to include elements. This makes it possible to re-analyze the invention document in response to updates such as the setting information acquired by the setting information acquisition unit 201, the important term setting dictionary acquired by the important term setting dictionary acquisition unit 203, etc. It can be easy.

In this embodiment, as an example, each part of patent application related information (for example, claim information, specification information, claim tree, important word list, etc.) as an HTML document compliant with HTML5 is placed under a separate <article> tag. Although various explanations have been given using examples of describing each part of patent application related information, the tags used to describe each part of patent application related information are not limited to the <article> tag, and other tags (for example, <section> tag, etc.). Further, the HTML document related to patent application related information generated by the display document generation unit 2083 does not necessarily have to be compliant with HTML5.

If there is a code that cannot be used in an HTML file in each part of the patent application related information as a display document generated by the analysis unit 205, the display document generation unit 2083 converts the code and converts the code into an HTML file. generate. For example, when the display document generation unit 2083 is realized by a computer that executes a program written in the Python (registered trademark) language, this code conversion is performed using functions of the standard library of Python (registered trademark). xml. sax. saxutils. It can be implemented using escape. In order to ensure that when patent application related information is displayed as a display document in some existing browser applications, copying and pasting of multiple lines of displayed content will be executed appropriately. For example, when each part of the patent application related information as a display document generated by the analysis unit 205 includes a half-width space character, the generation unit 2083 converts the half-width space character into a predetermined code "&nbsp;". For example, if the <br> tag is not added to the end of each line of each part of the patent application related information as a display document generated by the analysis unit 205, the <br> tag may be replaced with It is also possible to add .

As a specific example of the display document generation unit 2083 outputting the generated display document, the display document generation unit 2083 may include information indicating the analysis date and time of the invention document in a predetermined folder. A subfolder with a specified subfolder name is created, and an HTML file containing patent application related information, which is the main part of an HTML document as a display document, is stored in the subfolder. This predetermined folder may be, for example, a folder indicated by the information for specifying the storage location of the output file in the setting information acquired by the setting information acquisition unit 201, or a folder determined by the user of the invention document analysis system 20. It may be a folder, or it may be a folder in which an invention document analysis processing program is installed in a terminal device or the like that includes a microprocessor that constitutes the invention document analysis system 20. In addition, in the subfolder name that includes information indicating the analysis date and time of the invention document, for example, if the invention document is acquired from a file by the invention document acquisition unit 2021, the file name of the file is added. Alternatively, if the invention document acquisition unit 2021 acquires the invention document from a clipboard, a phrase indicating the clipboard may be included. The display document generation unit 2083 further defines one subfolder within the predetermined folder as a storage destination for common resources, and displays a display added to a display element regarding patent application related information as a part of the HTML document. A CSS file for display mode control and a script file, which are resources that function in accordance with control information, are stored in the subfolder where the common resource is stored, only if they are not already stored. According to this example, in the invention document analysis system 20, each subfolder within a predetermined folder is sorted by subfolder name, and, for example, a subfolder that stores an HTML file containing patent application related information based on the analysis result of an invention document is created. It is possible to display folder names in order from the newest to the oldest. For example, each HTML file in each subfolder may contain patent application related information as a result of an analysis of an invention document with different content, but each HTML file may be in one subfolder where common resources are stored. An HTML document is composed of a CSS file and a script file. Note that the display unit 2084 displays a list enumerating each subfolder within the above-mentioned predetermined folder, and in response to a selection operation for one subfolder, displays information including HTML files in the subfolder related to the selection operation. It may also be possible to display a GUI screen in which the configured HTML document is interpreted and displayed by the browser, and within that GUI screen, depending on the user's operation, one of the subfolders can be displayed with the contents under it. It is also possible to display an operation-compatible GUI element for performing the process of including and deleting. The display of this GUI screen may be realized by an HTML document such as an HTML file, or a hyperlink may be created to transition the displayed content to the corresponding HTML file in response to a selection operation for one subfolder. Note that the operations of the display document generation unit 2083, display unit 2084, etc. are performed when the invention document analysis processing program described above is installed and executed on a terminal device equipped with a microprocessor, etc. It may also constitute part of the analysis process.

FIG. 39 is a diagram showing an example of a display screen as a GUI screen displayed by the display unit 2084. The display screen illustrated in FIG. 39 displays GUI elements for acquiring an invention document to be analyzed by selecting an invention document file or from the clipboard. Selection of the invention document file can be realized, for example, by specifying file in the type attribute of the HTML <input> tag. On the display screen illustrated in FIG. 39, the list of analyzed invention documents includes the names of subfolders that store HTML files containing patent application related information (for example, the date and time of analysis of invention documents, and the file name or characters of invention documents that were the subject of analysis). column "clipboard") are listed, and a push button for starting a function for deleting that subfolder is displayed to the left of each subfolder name. For example, when a mouse click operation is performed on a subfolder name displayed on this display screen, a display transition related to the link occurs, and an HTML document containing patent application related information stored in that subfolder is displayed. Display contents based on the file (see FIG. 17) will be displayed. When display elements of important terms in claim information, specification information, etc. in patent application related information stored in one subfolder in a predetermined folder as an HTML document are displayed by a browser, etc. The above-mentioned important word inclusion description enumeration process, which can be executed in response to an operation on a GUI element corresponding to the operation, can be performed, for example, on an HTML document in each subfolder other than the one subfolder in the predetermined folder. From the claim information and specification information in the patent application related information, the contents of each line in which the important term is written are collected by searching for the important term, and each line that is the collection result is listed and each line is It may also be a process of emphasizing and displaying important words and phrases. As a result, the user can check claim information and specifications in one or more pieces of patent application-related information that have already been generated based on invention documents that have been analyzed by the invention document analysis system 20 in the past, regarding important words and phrases in the patent application-related information. It becomes possible to refer to each line containing important words and phrases included in the book information. The subfolder deletion function is written, for example, in JavaScript (registered trademark), and calls a subfolder deletion processing function written in, for example, Python (registered trademark) that runs on the local host.

As another specific example, the display document generation unit 2083 includes information indicating the analysis date and time of the invention document in a predetermined folder indicated by the output file storage destination identification information in the setting information. A subfolder with a specified subfolder name is created, and an HTML file, a CSS file, and a script file constituting an HTML document as a display document related to patent application related information are stored in the subfolder. According to this example, by recording the contents of the subfolder on a portable recording medium and transferring it, or by compressing the contents of the subfolder and transmitting it, the contents of the subfolder can be transferred to the destination or destination. Patent application related information can also be viewed and utilized as a display document on various terminals (for example, desktop PCs, notebook PCs, smartphones) having the aforementioned browser function.

If the inventor, etc. or patent attorney determines that the patent application related information outputted from the invention document analysis system 20 based on the analysis result of the invention document is not appropriate, the inventor, etc. By modifying the document, the setting information acquired by the setting information acquisition unit 201, the important term setting dictionary acquired by the important term setting dictionary acquisition unit 203, etc., and having the invention document analysis system 20 reanalyze the invention document, a patent can be obtained. Application-related information can be updated to have more appropriate content. A translator who receives patent application-related information from an inventor or a patent attorney, for example, in the form of a display document that has been updated to have appropriate content, may, based on the appropriate patent application-related information, It may become possible to perform translation work appropriately or efficiently. In addition, when a translator who can use the invention document analysis system 20 receives an invention document, setting information, important word setting dictionary, etc. from an inventor or a patent attorney, the translator can use the invention document analysis system 20. Patent application related information can be obtained, and based on the patent application related information, it may be possible to perform translation work appropriately or efficiently.

When generating patent application related information including information in which line numbers are added to each line of an invention document as a display document, the display document generation unit 2083, for example, generates information about each display element corresponding to each line of the invention document. Display control information that specifies that the display mode of the line (for example, the color of the character string in the line) is displayed differently depending on whether the line is the line that has been pointed out in the warning information. It may also be added. In addition, instead of this, when generating patent application related information including information in which line numbers are added to each line of the invention document as a display document, the display document generation unit 2083 generates, for example, the line number of each line of the invention document. is one display element, and the display element corresponding to each line number displays the line number of the line pointed out in the warning information and the line number of the line pointed out in the warning information. It is also possible to add display control information that specifies that the lines are displayed in different formats (for example, the font colors of the line numbers). In addition, when generating patent application related information including information in which line numbers are added to each line of the invention document as a display document, the display document generation unit 2083 may, for example, add warning information to each line of the invention document. When an inappropriate description that is the subject of a It is also possible to add display control information that specifies that the inappropriately written character string be displayed (for example, highlighted by changing the character color of the inappropriately written character string to red, etc.).

The display unit 2084 displays patent application related information as a display document generated by the display document generation unit 2083. The display unit 2084 is, for example, a browser that interprets an HTML document in accordance with HTML5 and displays display elements defined in the HTML file in accordance with a CSS file or the like. When the display document generation section 2083 generates a display document related to patent application related information as an HTML document that is not compliant with HTML5 (that is, an HTML document that is compliant with other standards other than HTML5), The display unit 2084 has a function as a browser compatible with a standard that can interpret and display the HTML document generated by the display document generation unit 2083. In addition, all or part of the content displayed by the browser in the patent application related information as a display document can be selected by the user's operation and can be copied and pasted or cut and pasted. By pasting it, it can be included as the content of a document or the like used by a certain application program that is being executed on the terminal device that constitutes the invention document analysis system 20. In addition, if the content displayed by the browser changes as a result of processing executed in response to user operations on GUI elements related to patent application related information, all or part of the changed display content may be changed by the user. Depending on the operation, it can be selected and can be copied and pasted.

In addition to having the function of a browser that interprets and displays a display document, the display unit 2084 also has a function of displaying patent application related information output as a file other than a display document (for example, a text file). Furthermore, like an application program (for example, a text editor) that has a conventional document creation and editing function, it has the function of editing and outputting the displayed patent application related information according to the user's operations. You may do so. When executing the function of editing and outputting patent application related information in accordance with user operations, the display unit 2084 displays a plurality of important words included in the list of important words in the patent application related information in order to support user input. A function may be provided that allows the user to input one selected important word or phrase at the cursor position from a pop-up screen or the like in which important words are listed and displayed in a predetermined order (for example, in order of importance).

The display section 2084 includes an operation reception section 20841. For example, if the display control information added to a display element defined in an HTML file specifies that the display element is an operation-compatible GUI element (that is, a GUI element that starts processing in response to an operation), the operation reception unit 20841 (in the case where the display element is to have a function as a display element), the display element receives the user's operation and executes the function etc. determined by the display control information.

As a specific example, the operation reception unit 20841 selects a translation word candidate set from display elements related to a translation word candidate set added to important words in the important word list part of patent application related information as a display document. When a selection operation for one element is received, a predetermined function is activated using the element as an argument, and a process for determining a corresponding translation word for the important phrase is executed. When the operation reception unit 20841 accepts a user's selection operation for an important phrase from a set of translation word candidates, the selected element forms a pair with the important phrase, for example, and generates a learning model for important phrase translation. can be stored in the teacher data storage unit 2072 as teacher data for machine learning. The output unit 208 also outputs, for each of the one or more important words, the important word and the corresponding translation word determined by the operation reception unit 20841 according to the user's selection from the translation word candidate set for the important word. It is possible to generate and output translation auxiliary information including a set of .

In addition, the operation reception unit 20841 receives, for example, an operation for a display element corresponding to an important term in the important term list part of patent application related information as a display document, and inputs the important term as an argument according to the operation. A predetermined function is activated to convert the important word to a non-important word. In the non-important word/phrase processing, a part of the sentence in which the important word is written in the invention document (for example, a part separated by a comma character or a period character) is associated with the important word/phrase, and a non-important word/phrase setting dictionary is created. It may also be included in In the invention document analysis system 20, from the contents of the important word/phrase setting dictionary and the non-important word/phrase setting dictionary that are updated in this non-important word/phrase processing, the part of the sentence or the word/phrase to be inputted is determined. Generates training data for machine learning that generates a learning model for extracting important words to predict whether or not it contains important words, and if it contains important words, the important words. 207 may generate a learning model for extracting important words, or the learning model output unit 2081 may output a learning model for extracting important words. This important word extraction learning model can be used, for example, in the important word extraction unit 2051 to extract important words from an invention document.

[3-1-7-3. [Display document for displaying settings information]
The display document generation unit 2083 generates a setting information display document, such as an HTML document, for displaying all or part of the setting information acquired by the setting information acquisition unit 201. Good too. In a setting information display document, for example, various types of information included in the setting information are included as display elements, and display elements for some of the various types of information may be configured as GUI elements that support operation. display control information may be added. The display document generation unit 2083 generates the invention document standard for translation acquired by the invention document standard provision acquisition unit 204 based on the identification information of the invention document standard provision for translation in the setting information, in the setting information display document. All or part of the content of each condition item in the regulations may be included as a display element, and display control information for making the display element an operation-compatible GUI element may be added to the display element.

An example of a display element that becomes an operation-compatible GUI element in the setting information display document generated by the display document generation unit 2083 is a modification used when generating revised claim information for the area where the application is planned. This includes modification information that defines the method. A specific example of a GUI element that corresponds to the operation of this correction information is to display a plurality of correction methods with a check box attached to each one, and select the correction method whose check box is checked as the content of this correction information. An example of updating setting information is: Examples of these multiple modification methods include ``if there is a clause in which the claimed object of a claim is a program, change that term to a term in which the claimed subject is a program recording medium''; ``When any of the cited terms in a claim cites multiple terms, that one term is changed to a single term that cites only one of the multiple terms'', ``Changing multiple terms in a claim "Change a cited term into a single term that cites only one of the multiple terms"; "Citing only one different term or changing the terms to the same number of terms", "If a code corresponding to an element represented by an important term is written in the specification, a code is added to the important term in the claim" Examples include "I do". The configuration information display document can be interpreted and displayed by a browser.

The display unit 2084 can display a setting information display document, for example. When the display element of the setting information display document is displayed on the display unit 2084, the user can edit the contents of various information (for example, condition information) included in the setting information using the operation-compatible GUI element. Among the condition items in the invention document standard provisions for translation, the user may be able to select a condition item in order to narrow down the condition items to be used for evaluation as a basis for generating invention document evaluation information. The operation reception unit 20841 receives operations such as editing and selection by the user, and executes predetermined processing according to the display item to be operated and the operation. This processing includes, for example, defining the contents of setting information (e.g., reference character string pattern information, specific particle character string pattern information, and the modification method used when generating revised claim information for the area where the application is planned. modification information, etc.) (for example, updating a configuration information file containing configuration information), updating conditions for inappropriate descriptions of invention documents used for generating warning information, translations used for generating invention document evaluation information Narrowing down of condition items in the invention document standard provisions for use can be performed.

[3-1-7-4. Output of important word group information]
The output unit 208 can output, for example, important word group information that lists a plurality of important words extracted by the important word extraction unit 2051 from the invention document acquired by the invention document acquisition unit 2021.

The output unit 208 may add the importance level determined by the importance level determination unit 2055 to each important term in the important term group information by referring to the important term management information held by the analysis unit 205. Attribute information specified by the attribute specifying unit 2056 may be added. The output unit 208 may include information regarding only important words whose importance exceeds a predetermined value among the important words extracted by the important word extraction unit 2051 in the important word group information. good.

For each important word in the important word group information, the output unit 208 outputs a corresponding translation word determined by the translation unit 2054 corresponding to the important word or a translation word candidate set specifying unit 20541 corresponding to the important word. The element selected by the operation reception unit 20841 from among the elements of the specified translation word candidate set may be added.

The important word group information output based on the invention documents can be used, for example, to classify multiple invention documents, determine the correlation between invention documents, generate an important word setting dictionary, generate translation auxiliary information, etc. .

[3-2. Invention document analysis processing example according to Embodiment 2]
An example of an invention document analysis process executed by the invention document analysis system 20 having the above-described configuration will be described.

FIG. 40 is a flowchart illustrating an example of invention document analysis processing.

As shown in FIG. 40, the invention document analysis system 20 performs the invention document analysis process, for example, as shown in FIG. Acquisition of regulations (step S13), acquisition of invention documents (step S14), acquisition of drawing documents (step S15), citation relationship analysis processing regarding claims (step S16), and important phrase extraction processing (step S17) , inter-element relationship determination processing (step S18), identification of attribute information of each important term (step S19), determination of the importance of each important term (step S20), and identification of a set of translation word candidates ( step S21), technical field identification processing (step S22), display color determination processing for important words and phrases (step S23), warning information generation processing (step S24), and complaint information generation processing (step S25). , claim tree generation processing (step S26), specification information generation processing (step S27), various translation information generation processing (step S28), important word list generation processing (step S29), patent application generation processing A related information generation process (step S30), a display document generation process (step S31), and a patent application related information output process (step S32) are executed. Note that the example in FIG. 40 is merely an example, and the invention document analysis system 20 may execute a process in which some of the steps shown in FIG. 40 are omitted or changed as the invention document analysis process, or You may add some processing steps to the steps shown in FIG. 40 and execute them, or you may change the order of execution of the steps shown in FIG. All or some of the steps may be performed in parallel.

Each process shown in FIG. 40 will be described below with reference to the drawings.

In the setting information acquisition process (step S11), the setting information acquisition unit 201 acquires setting information.

In acquiring the important word/phrase setting dictionary (step S12), the important word/phrase setting dictionary acquisition unit 203 obtains the important word/phrase setting dictionary based on, for example, setting information.

In acquiring the invention document standard provisions for translation (step S13), the invention document standard provisions for translation acquisition unit 204 acquires the invention document standard provisions for translation based on, for example, the setting information.

In the acquisition of an invention document (step S14), the invention document acquisition unit 2021 acquires an invention document to be analyzed by the analysis unit 205 of the invention document analysis system 20, according to the user's designation of the invention document analysis system 20.

In the drawing document acquisition (step S15), the drawing document acquisition unit 2022 acquires a drawing document including a drawing corresponding to the invention document acquired by the invention document acquisition unit 2021, for example, according to a user's designation.

In the claim-related citation relationship analysis process (step S16), the claim citation relationship detection unit 2057 executes a citation relationship analysis process that analyzes the citation relationship between each clause of the claim of the invention document acquired in step S14.

FIG. 41 is a flowchart showing an example of the citation relationship analysis process (step S16). As a citation relationship analysis process, the claim citation relationship detection unit 2057 searches, for example, for each clause of a claim in an invention document, a preceding item number description part that corresponds to a predetermined cited item number description pattern, and detects the preceding item number description. If there is a part, it is extracted (step S161), and one or more item numbers are specified from the extracted preceding item number description part (step S162). Then, as a citation relationship analysis process, the claim citation relationship detection unit 2057 updates, for each item of the claim, claim citation relationship information that includes information associating that term with the set of item numbers identified in step S162. By storing the information, the claim citation related information is stored and managed (step S163).

In the important term extraction process (step S17), the important term extraction unit 2051 extracts important terms from the claim of the invention document acquired in step S14 based on the setting information acquired in step S11. and a specification important word/phrase extraction process for extracting important words/phrases from the specification of the invention document.

FIG. 42 is a flowchart illustrating an example of complaint important phrase extraction processing. As the claim important phrase extraction process, the important phrase extraction unit 2051 extracts, for example, the same important phrases as the important phrases included in the important phrase setting dictionary acquired in step S12 from the claims of the invention document (step S1711), The subject of the claim is extracted as an important word/phrase (step S1712), and the character string corresponding to the important word/phrase candidate character string pattern immediately before the character string (i.e. code) corresponding to the code string pattern in the claim is extracted as the important word/phrase. (Step S1713), extracts a character string corresponding to the important word candidate character string pattern following the character string corresponding to the reference character string pattern in the claim as an important word (step S1714), and extracts the character string corresponding to the important word phrase candidate character string pattern in the claim Among the character strings that correspond to the column pattern (that is, candidates for important words), those that satisfy certain conditions (for example, when the range of descriptions of multiple candidates for important words have an inclusion relationship in the description, the relationship between the including side and the included side) (inclusive side) are extracted as important words and phrases (step S1715). Note that the example in FIG. 42 is merely an example, and the important word/phrase extraction unit 2051 may execute a process in which some of the steps shown in FIG. 42 are omitted or changed as the complaint important word/phrase extraction process, or 42, some processing steps may be added to the steps shown in FIG. 42, the order of execution of the steps shown in FIG. 42 may be changed, or the steps shown in FIG. All or some of the steps may be performed in parallel.

FIG. 43 is a flowchart illustrating an example of specification important word extraction processing. As the specification important phrase extraction process, the important phrase extraction unit 2051 extracts, for example, the same important phrases as the important phrases included in the important phrase setting dictionary acquired in step S12 from the specification of the invention document (step S1721). ), if there is an important phrase extracted by the claim important phrase extraction process in the specification, the important phrase is extracted (step S1722), and the key phrase is extracted immediately before the character string (that is, the code) that corresponds to the code character string pattern in the specification. The character string that corresponds to the important word/phrase candidate character string pattern is extracted as an important word (step S1723), and the important word/phrase candidate character immediately before the character string that corresponds to the specific particle character string pattern (that is, the predetermined particle) in the specification is extracted as an important word/phrase. The character strings that correspond to the column pattern are extracted as important words (step S1724), and the character strings that meet a certain condition (for example, multiple Among the key word candidates, those containing the key word extracted in the complaint key word extraction process at the end of the word) are extracted as key words (step S1725). Note that the example in FIG. 43 is only an example, and the important word/phrase extraction unit 2051 may execute a process in which some of the steps shown in FIG. 43 are omitted or changed as the specification important word/phrase extraction process, or Some processing steps may be added and executed to the plurality of steps shown in FIG. 43, the execution order of the plurality of steps shown in FIG. 43 may be changed, or the steps shown in FIG. All or some of the steps shown may be performed in parallel.

In the inter-element relationship determination process (step S18), the inter-element relationship determination unit 2058 determines the relationship between elements represented by each of the plurality of important phrases extracted in the important phrase extraction process (step S17).

FIG. 44 is a flowchart showing an example of the inter-element relationship determination process (step S18). As the inter-element relationship determination process (step S18), the inter-element relationship determination unit 2058 determines, for example, the structural inclusion relationship between elements represented by the plurality of important words extracted in the important word extraction process (step S17). Inclusion relationship determination processing is executed (step S181), and recording relationship determination processing is executed to determine whether the elements represented by each of a plurality of important terms have a relationship with a recording medium and the contents of the recording medium (step S182). , executes a program realization relationship determination process that determines whether the elements represented by each of the plurality of important words have a relationship between the program and the content realized by the program (step S183), and determines whether the elements represented by the plurality of important words are equivalent. An equivalent relationship determination process is executed to determine that there is a relationship (step S184), and a hierarchical conceptual relationship determination process is executed to determine that there is a conceptual hierarchical relationship between elements expressed by a plurality of important words (step S185). ). The inter-element relationship determination unit 2058 then extracts inclusion relationship information indicating the result of the inclusion relationship determination process, recording relationship information indicating the result of the recording relationship determination process, and program realization relationship information indicating the result of the program implementation relationship determination process. , generates important term/phrase relationship information including equivalence relationship information indicating the result of the equivalence relationship determination process and superior/superior concept relationship information indicating the result of the superior/superior concept relationship determination process (step S186). Note that the example in FIG. 44 is only an example, and the inter-element relationship determination unit 2058 may execute a process in which some of the steps shown in FIG. 44 are omitted or changed as the inter-element relationship determination process, or Some processing steps may be added and executed to the plurality of steps shown in FIG. 44, the execution order of the plurality of steps shown in FIG. 44 may be changed, or the steps shown in FIG. All or some of the steps shown may be performed in parallel.

In specifying the attribute information of each important term (step S19), the attribute specifying unit 2056 identifies the results of the inter-element relationship determination process (step S18) for each important term extracted in the important term extraction process (step S17). The attribute information (see FIG. 19) of the important word/phrase is specified.

In the process of determining the importance of each important word/phrase (step S20), the importance determining unit 2055 determines the degree of importance of each important word/phrase extracted in the important word/phrase extraction process (step S17).

FIG. 45 is a flowchart illustrating an example of the process of determining the importance of each important phrase (step S20). As the process of determining the importance of each important term (step S20), the importance determining unit 2055 determines, for example, that the attribute specifying unit 2056 specifies the information in step S19 for each important term extracted in the important term extracting process (step S17). The importance of the important word/phrase is determined based on the attribute information (step S201), and further, between important words representing elements having a structural inclusion relationship, the including side is determined to have the same or higher importance as the included side. The importance level is re-determined so that (step S202). Note that the example in FIG. 45 is merely an example, and the importance level determining unit 2055 may also execute a process in which some of the steps shown in FIG. 45 are omitted or changed as the process for determining the importance level of each important term. Alternatively, some processing steps may be added to the steps shown in FIG. 45, or the order of execution of the steps shown in FIG. 45 may be changed. All or some of the steps shown in FIG. 45 may be performed in parallel.

In the translation word candidate set identification process (step S21), the translation word candidate set identification unit 20541 uses the bilingual dictionary information to identify each important word extracted in the important word extraction process (step S17). A translation word candidate set, which is a collection of corresponding translation word candidates, is identified. In the translation word candidate set identification process (step S21), the translation word candidate set identification unit 20541 performs calculation processing for each important word using, for example, a learning model for important word translation with the important word as input. The generated machine translation result can be added to a set of translation word candidates for the important word/phrase.

In the technical field identification process (step S22), the technical field identification unit 20531 performs a technical field identification process based on the important word group, which is a set of important words extracted from the invention document obtained in step S14 in the important word extraction process (step S17). to identify the technical field of the invention document. In the technical field identification process (step S22), the technical field specifying unit 20531 identifies technical fields and important words extracted in the past from the invention documents in each of the plurality of invention documents whose correspondence with the technical field is known. The technical field of an invention document for which the technical field is unknown is identified by using the relationship with the important word group. In the technical field identification process (step S22), the technical field identification unit 20531 performs a predetermined calculation process using a learning model for technical field identification, for example, based on the important word group extracted from the invention document. can identify the technical field to which the invention document belongs.

In the process of determining the display color of important words and phrases (step S23), for example, the analysis unit 205 uses the important words and phrases extracted in the important word and phrase extraction process (step S17) as part of the patent application related information as a display document. The display control information added to the display element includes an identification code that includes a serial number determined in descending order based on the first appearance line number of the important word, and analysis. The unit 205 or the display document generating unit 2083 determines the display control information so that display colors as display modes of display elements whose serial number portions of their identification codes are relatively close to each other have relatively large differences. . Here, for example, for a plurality of important words and phrases in which the character strings constituting the important words in claim information expressing a claim are different from each other, the serial number included in the identification code of the important words is, for example, the same character string as the important words. The key terms are arranged in columns in ascending order of their earliest position in the claim (for example, the line number in which the key terms first appear). In this example, the display control information relating to each of a plurality of important words and phrases in which the character strings constituting the important words in the complaint information are different from each other is the serial number associated with the identification code in the identification information for each important word and phrase. Each important word is set so that the difference in display color between consecutive important words (for example, difference in hue component, difference in RGB value components, etc.) is greater than the difference in display color between important words whose serial number is not consecutive. It may also be used to control the display color added to words. For example, the RGB values (in this case, each RGB component This will be explained using an example that can have 256 values from 0 to 255 (0xff). The RGB values of the display colors added to the important word according to the display control information are red 0, green "255" (0xff), and blue "0", and the display control information of the important word whose serial number is 3 indicates that the important word The RGB values of the display colors added to are red "0", green "0", and blue "255" (0xff), and the display added to the important word according to the display control information of the important word whose serial number is 4. The RGB values of the colors are red "153" (0x99), green "0", and blue "0", and the RGB value of the display color added to the important word is determined by the display control information of the important word whose serial number is 5. If red is "0", green is "153" (0x99), and blue is "0", then two important words out of the plurality of important words in the first group whose serial numbers are less than or equal to a predetermined number (3 in this example) The minimum value of the difference in the display color added to each important word in each set of words for all sets is the difference between a plurality of important words in the second group whose serial number is larger than a predetermined number and a plurality of important words in the first group. Among a plurality of important words and phrases in a mixed group in which words and phrases are mixed, the difference in display color added to each important word in each set of two important words is greater than the minimum value of all the sets.For example, simple Assuming that the RGB values are each component of three orthogonal axes, the difference in display color can be calculated by calculating the distance between the RGB values of the display colors added to each of the two important words. View the claim information when each important word in the claim information in the patent application related information as a display document is displayed using the important word display control information that controls the difference in display colors added to the display. It may be possible for translators, etc., to quickly identify each important phrase in the complaint information.Adjusting the display color based on the serial number related to the identification information of the important phrase (i.e., the first line number of the important phrase) as described above , is useful for improving the distinguishability between important words in a relatively important part near the beginning of claim information (for example, the description part of a claim with a relatively small item number).
As another example, for example, the display color added to the first group of important words whose serial numbers are from 1 to a first predetermined number (for example, 200) of 216 or less is "0" as the value of each component of the RGB value, Use one of "51" (0x33), "102" (0x66), "153" (0x99), "204" (0xcc), and "255" (0xff) to be different from each other. The display color added to the important words and phrases of the second group whose serial numbers are greater than the first predetermined number (for example, 200) and up to the second predetermined number (for example, 300) is "34" as the value of each component of the RGB value. (0x22), "85" (0x55), "136" (0x88), "187" (0xbb), and "238" (0xee) to be different from each other, The display color added to the important words and phrases of the third group whose serial numbers are greater than the second predetermined number (for example, 300) and up to the third predetermined number (for example, 400) is "17" (0x11) as the value of each component of the RGB value. ), "68" (0x44), "119" (0x77), "170" (0xaa), and "221" (0xdd) to be different from each other. good. In this example as well, among the plurality of important words and phrases in the first group whose serial numbers are equal to or less than the first predetermined number (200 in this example), the difference in the display color added to each important word in each set of two important words is shown. , the minimum value of all the sets is the second group whose serial number is larger than a predetermined number, or a mixture of a plurality of important words from the second and third groups and a plurality of important words from the first group. Among the plurality of important words and phrases in the group, the difference in display color added to each important word in each of the two sets of important words is greater than the minimum value of all the sets. In a group in which the minimum value of the difference between the display colors of each important term is large, each important term can be identified relatively quickly and correctly based on the difference between the display colors. In this example, when each important phrase in the claim information in the patent application related information as a display document is displayed, for a translator etc. viewing the claim information, it is written relatively near the beginning of the claim information. It becomes possible to relatively quickly or correctly identify the important words and phrases of the first group that have been identified from each other.
Display control information added to various display elements in patent application related information as a display document includes changing the display color of the display element to emphasize the display element, bolding the character string of the content of the display element, Display elements may be emphasized by changing the font size, etc., or information may be displayed that adds information to the content of display elements (for example, adding importance, attribute information, etc. to important words). The display element may be a GUI element for adding and displaying information in response to an operation, or the display element may be a GUI element for displaying additional information in response to an operation. , it may be a GUI element for performing processing such as translation, or it may be a GUI element for performing display transition to display information related to the content of the display element according to the operation of the display element. It may be something that does.

In the warning information generation process (step S24), the warning information generation unit 2059 generates the invention acquired in step S14 based on the setting information acquired in step S11, the invention document standard regulations for translation acquired in step S13, etc. Generate warning information including pointing information pointing out inappropriate descriptions in a document.

FIG. 46 is a flowchart showing an example of the warning information generation process (step S24). As the warning information generation process (step S24), the warning information generation unit 2059 includes, for example, pointing information in the warning information that points out if the description of the invention document does not conform to the "standard provisions for invention documents for translation" (step S24). S241), the claim citation related information is referred to, and if there is no preceding phrase corresponding to the reference statement in the claim of the invention document, warning information is included to point out that fact (step S242), and the invention document If there are deficiencies in the description format such as claims, headings of the specification, drawing numbers, paragraph numbers, etc., the warning information includes pointing information to that effect (step S243), If there is any one or more of the following: deficiencies, inconsistency in notation, and possibility of typographical errors, the warning information includes pointing information to that effect (step S244). Display control for display elements related to each pointed information to function as GUI elements for display transition to the line of the invention document targeted for pointing out, the section of claim information related to the pointing out, or the paragraph of specification information Information is added (step S245), and invention document evaluation information is generated and added to the warning information (step S246). Note that the example in FIG. 46 is only an example, and the warning information generation unit 2059 may execute a process in which some of the steps shown in FIG. 46 are omitted or changed as the warning information generation process, or It is also possible to add some processing steps to the plurality of steps shown in FIG. All or some of the steps may be performed in parallel.

In the claim information generation process (step S25), the claim information generation unit 205a generates claim information etc. to be included in the patent application related information based on the claims of the invention document acquired in step S14.

FIG. 47 is a flowchart showing an example of the complaint information generation process (step S25). As the complaint information generation process (step S25), the complaint information generation unit 205a executes, for example, a complaint information generation process (step S251), a code determination process (step S252), and a claim-based drawing information generation process. (Step S253), executes claim-based statement information generation processing (Step S254), executes claim-based summary information generation processing (Step S255), and executes revised claim information generation processing (Step S256). .

FIG. 48 is a flowchart showing an example of the complaint information generation process (step S251). As a claim information generation process (step S251), the claim information generation unit 205a generates claim information representing a claim based on the claim in the invention document acquired in step S14 (step S2511), and displays it as a display document. Basically, a different display color is added to each important word to the display element of each important word in the claim information in the patent application related information, and information related to the important word is added and displayed according to the operation. Display control information is added to the complaint information (step S2512) in order to provide a GUI function for performing processing such as searching and translating important words in response to other operations. Display control information is added to the display elements of the headings of each section in order to provide a GUI function that supports operations such as highlighting, display transition to the next heading, etc. (step S2513), and complaint information is added. Display control information is added to the display element of the reference character string in the content for highlighting, etc. (step S2514), and display control information is added to the display element of the comment in the complaint information for highlighting, etc. Control information is added (step S2515), and display control information is added for highlighting, etc. to the display element of the portion of the complaint information that is the target of the warning information (step S2516).

FIG. 49 is a flowchart showing an example of the code determination process (step S252). As the code determination process (step S252), the claim information generation unit 205a, for example, selects the important words extracted from the claims in the invention document obtained in step S14 by the important word extraction process (step S17). If the category type of the element represented by the word is either product or method, it is selected as a code determination target (step S2521), and the important word selected as a code determination target corresponds to the code in the "code explanation" in the invention document. If so, that code is adopted with the highest priority (step S2522), and if the important word selected as the target for code determination is given a code in the specification of the invention document, that code is adopted as the next highest priority. (step S2523), and for each important word selected as a code determination target, based on the structural inclusion relationship of the elements represented by the important word, the character string of the code of the included important word and the included The code is determined so that the character string of the code of the important word on the other side has a predetermined relationship such as an inclusive relationship between the character strings (step S2524).

FIG. 50 is a flowchart illustrating an example of claim-based drawing information generation processing (step S253). As the claim-based drawing information generation process (step S253), the claim information generation unit 205a generates, for example, a claim represented by the important words extracted from the claims in the invention document acquired in step S14 by the important word extraction process (step S17). When the target includes elements expressed by other important terms, a configuration diagram is generated in which rectangular frames surrounding each important term are arranged according to the inclusion relationship (step S2531), and a configuration diagram is generated in which a rectangular frame surrounding each important term is arranged according to the inclusion relationship (step S2531). If such an element records an element expressed by another important term, a diagram listing the recorded contents of the recording medium is generated (step S2532), and if the element related to the program expressed by the important term of the claim records an element expressed by another important term. When realizing the function, process, etc. of the element represented by , a diagram showing a set of elements related to the function to be realized is generated (step S2533), and an important element whose sign is determined by the sign determination process (step S252) is generated. For words and phrases, a code connected by a leader line from the drawn content (for example, a rectangular frame) related to the important words in the figure is added (step S2534), and for each figure, the smallest first claim number of the important word in each figure is added. Drawing numbers are determined in order (step S2535), and display control is performed by code search corresponding to the display element of each code in the claim-based drawing information (for example, control to highlight the code by performing a display transition to display the code) Display control information to enable this is added (step S2536).

FIG. 51 is a flowchart showing an example of the claim-based specification information generation process (step S254). As the claim-based specification information generation process (step S254), the claim information generation unit 205a generates a "name of the invention" based on the claim subject of each claim originating claim of the claim in the invention document acquired in step S14, for example. ”, “problem to be solved by the invention”, and “means for solving the problem” (step S2541), and generate the contents of the claim-based drawing information generated in the claim-based drawing information generation process (step S253). The content of the "brief description of the drawings" is generated based on the content (step S2542), and the content of the "brief explanation of the drawings" is generated for each claim, and the "for carrying out the invention" is generated, which includes the content and an unfinished message for supporting description of the corresponding effect for each claim. The content of the "format" is generated (step S2543), and the explanation order of each important phrase is determined based on the important phrase related information generated in step S186 and the claim base drawing information (step S2544). In accordance with the order of explanation, each important term is the subject, a code is added to the important term whose code has been determined, and an unfinished message for writing support is added as necessary. ” (step S2545), and control the display mode, such as highlighting, on the display elements of important phrases, codes, and incomplete messages for writing support in the claim-based specification information in the patent application related information as the display document. , display control information is added for control and the like to have a function as an operation-compatible GUI element (step S2546).

FIG. 52 is a flowchart illustrating an example of claim-based summary information generation processing (step S255). As the claim-based abstract information generation process (step S255), the claim information generation unit 205a, for example, acquires setting information in the "Issue" column under "Abstract" in the same format as the abstract of a patent application in Japan. Write an unfinished message for description support based on the unfinished message information of the setting information acquired in (step S11) (step S2551), and write the claim in the invention document acquired in step S14 in the "Solution" column. Include the content of the first term, add a code to important words whose codes have been determined in the code determination process (step S252) (step S2552), and write "Figure 1" in the "selected figure" column. (Step S2553), describes the contents of the first section of the claims as a Japanese version of the abstract for US applications (Step S2554), and writes the contents of the first section of the claims as an abstract of the Japanese version of the international application form (Step S2554). , and for important words whose codes have been determined in the code determination process (step S252), the codes are added in the following parentheses (step S2555), and the patent application related information is displayed as a display document. Display control information for highlighting, etc. is added to display elements such as each important phrase in the claim-based summary information (step S2556).

FIG. 53 is a flowchart illustrating an example of the modified complaint information generation process (step S256). As the modified complaint information generation process (step S256), the complaint information generation unit 205a generates, for example, warning information generation process (step S24) based on the complaint information generated in the complaint information generation process (step S251). Modified claim information is generated by modifying the inappropriate description pointed out by the warning information based on the modification information etc. (step S2561), and the claim information is written in the modified claim information in step S11. The program is modified to conform to the format specified by the acquired setting information (for example, specifying numbers as full-width characters, codes added to important words, etc.) (step S2562), and the program is modified based on the setting information. Modified claim information, such as a section describing the claimed subject matter belonging to the category, is added to the claim (step S2563), and each section of the claim (for example, claim information After processing, such as changing a clause that cites multiple clauses to a clause that cites a singular clause, etc.) is added to the revised claim information (step S2564), and the patent application is displayed as a display document. Adding display control information to display elements such as important words in the revised claim information in related information in order to highlight the display element or to make the display element an operation-compatible GUI element ( Step S2565).

Note that the examples shown in FIGS. 47 to 53 are merely examples, and the complaint information generation unit 205a may perform processing in which some of the steps shown in each of FIGS. 47 to 53 are omitted or changed. , some processing steps may be added and executed to the plurality of steps shown in each of FIGS. 47 to 53, or the execution order of the plurality of steps shown in each of FIGS. 47 to 53 may be changed. Alternatively, all or part of the plurality of steps shown in each of FIGS. 47 to 53 may be performed in parallel.

In the claim tree generation process (step S26), the claim tree generation unit 205c generates, for example, a claim tree, a modified claim tree, a constituent claim tree, and a modified claim tree.

FIG. 54 is a flowchart showing an example of the claim tree etc. generation process (step S26). As the claim tree etc. generation process (step S26), the claim tree generation unit 205c, for example, refers to the claim citation related information and generates the claim information generated in the claim information generation process (step S251) and the modified version of the claim information generation process ( A claim tree and a revised claim tree are generated based on each of the revised claim information generated in step S256) (step S261), and a constituent claim tree is generated based on the claim information and inclusion relationship information (step S262). ), a revised version of the constituent claim tree is generated based on the revised version of the claim information and the inclusion relation information (step S263), and a modified version of the constituent claim tree is generated based on the modified version of the claim information and the inclusion relation information (step S263), and is used to express the citation relations in the various generated claim trees in the patent application related information as a display document. Display control information is added to each display element of the symbol and important phrase in order to highlight the display element or to make the display element an operation-compatible GUI element (step S264). Note that the example in FIG. 54 is merely an example, and the complaint tree generation unit 205c may execute a process in which some of the steps shown in FIG. 54 are omitted or changed as the complaint tree generation process, or 54, some processing steps may be added to the steps shown in FIG. 54, the order of execution of the steps shown in FIG. 54 may be changed, or the steps shown in FIG. All or some of the steps may be performed in parallel.

In the specification information generation process (step S27), the specification information generation unit 205b generates, for example, specification information, a specification heading list, a code list, and revised specification information.

FIG. 55 is a flowchart showing an example of the statement information generation process (step S27). As the specification information generation process (step S27), the specification information generation unit 205b performs, for example, a specification information generation process that generates specification information representing the specification based on the specification of the invention document acquired in step S14. (step S271), and executes a specification heading list generation process (step S272) that extracts the headings of the specification information and generates a specification heading list (step S272). A code list generation process is performed to generate a code list in which the existing codes are sorted and enumerated (step S273), and the specification information is updated in response to the indication information in the warning information generated in the warning information generation process (step S24). A revised statement information generation process is performed to generate revised statement information by correcting inappropriate descriptions in the contents (step S274). Note that the example shown in FIG. 55 is merely an example, and the statement information generation unit 205b may execute a process in which some of the steps shown in FIG. 55 are omitted or changed as the statement information etc. generation process. , some processing steps may be added and executed to the plurality of steps shown in FIG. 55, or the execution order of the plurality of steps shown in FIG. 55 may be changed and executed. All or some of the steps shown in the following may be executed in parallel.

In the process of generating various translation information (step S28), the translation unit 2054, for example, corresponds to each of claim information, revised claim information, specification information, and revised specification information, and converts the contents of the information. Claim translation information, revised claim translation information, specification translation information, and revised specification translation information including at least some translation results are generated.

FIG. 56 is a flowchart showing an example of the generation process (step S28) of various translation information. As the process of generating various translation information (step S28), the translation unit 2054, for example, creates a section or line of claim information, etc. (that is, claim information, revised claim information, specification information, and revised specification information), For that term or line, claim translation information, etc. (i.e. Claim translation information, revised claim translation information, specification translation information, and revised specification translation information) are generated (step S281), and the partial translation lines of the claim translation information, etc. are generated using the invention document analysis system 20 in the past. A complete translation line is generated as a result of machine translation using the generated translation learning model or machine translation using another machine translation program (step S282), and partial translation of claim translation information, etc. The line is replaced with a complete translation line or a complete translation line is added to the partial translation line (step S283), and the display element of the corresponding translated word corresponding to each important word in claim translation information of patent application related information as a display document Display control information for displaying in a display mode that has commonality with the display element of the original important word is added to the display element of the original important word (step S284), and the display element of the corresponding translated word corresponding to each important word in the complaint translation information etc. Display control information is added to the GUI element for making it an operation-compatible GUI element that includes the GUI function related to the original important phrase and displays a set of translation word candidates for the important phrase (step S285). Note that the example in FIG. 56 is merely an example, and the translation unit 2054 may execute a process in which some of the steps shown in FIG. 56 are omitted or changed as the process for generating various translation information, or You may add some processing steps to the steps shown in FIG. 56 and execute them, or you may change the order of execution of the steps shown in FIG. All or some of the steps may be performed in parallel.

In the important phrase list generation process (step S29), the translation unit 2054 generates, for example, an important phrase list, a drawing code list, and important phrase related translation information.

FIG. 57 is a flowchart showing an example of the important word list etc. generation process (step S29). As the important phrase list generation process (step S29), the translation unit 2054, for example, for each important phrase extracted in the important phrase extraction process (step S17) from the invention document acquired in step S14, An important word list generation process is executed to generate a list of important words by associating a set of translation word candidates or corresponding translation words (step S291), and the important words in the patent application related information as a display document are executed. Display control information for controlling the display mode, etc. is added to each display element of the important words and phrases in the list, the corresponding translation words, and the translation word candidate set (step S292), and the drawing document obtained in step S15 is Drawing code list generation that generates a drawing code list based on the "code" item of the attribute information specified in step S19 for each important term extracted from the content and invention document in the important term extraction process (step S17) The process is executed (step S293), and the display mode of each display element of the drawing number, important word/phrase, corresponding translation word, translation word candidate set, and code in the drawing code list in the patent application related information as the display document is determined. (Step S294), and display control information for controlling the important words and phrases is added (step S294), and important words and phrases in the important word-phrase relationship information generated in step S186 are replaced with corresponding translation words or translation word candidate sets corresponding to the important words. Phrase-related translation information is generated (step S295). Note that the example in FIG. 57 is merely an example, and the translation unit 2054 may execute a process in which some of the steps shown in FIG. 57 are omitted or changed as important word list etc. generation process, or You may add some processing steps to the steps shown in FIG. 57 and execute them, or you may change the order of execution of the steps shown in FIG. All or some of the steps may be performed in parallel.

In the patent application related information generation process (step S30), the patent application related information output unit 2082 generates information in which a line number is added to each line of the invention document acquired in step S14 and analyzed by the analysis unit 205, for example. , important word/phrase relationship information generated in the inter-element relationship determination process (step S18), technical field information indicating the technical field specified in the technical field identification process (step S22), and warning information generation process (step S24). The warning information generated in step S25, the claim information generated in the claim information generation process (step S25), claim-based drawing information, claim-based specification information, claim-based summary information, and revised claim information, The claim tree, revised claim tree, constituent claim tree, and revised constituent claim tree generated in the tree generation process (step S26) and the specification information generated in the specification information generation process (step S27) , the specification heading list, the code list, the revised specification information, and the claim translation information generated in the process of generating various translation information (step S28), the revised claim translation information, the specification translation information, and the corrections. It is generated by the analysis unit 205 based on the version specification translation information, the important word list, the drawing code list, the important word related translation information generated in the important word list etc. generation process (step S29), and the invention document etc. The patent application related information (see FIG. 15) is generated by including the collected fee information and total information.

In the display document generation process (step S31), each piece of information generated by the analysis unit 205 and which is part of the patent application related information as a display document is integrated to create a display document (for example, an HTML document). generate.

In the process of outputting patent application related information (step S32), for example, the display unit 2084 displays the patent application related information as a display document generated in the display document generation process (step S31). The patent application related information output unit 2082 outputs the display document related to the patent application related information generated in the display document generation process (step S31) as one or more files. Alternatively, the patent application related information output unit 2082 outputs the patent application related information generated in the patent application related information generation process (step S30) as a text file or other file.

[3-3. Example of important word/phrase translation learning model generation process according to Embodiment 2]
The invention document analysis system 20 can execute an important term translation learning model generation process that generates an important term translation learning model for machine translation of important terms in an invention document. An example of a learning model generation process for important word/phrase translation will be described below.

FIG. 58 is a flowchart illustrating an example of a learning model generation process for important word/phrase translation.

As shown in FIG. 58, the invention document analysis system 20 performs a learning model generation process for important word/phrase translation, for example, as shown in FIG. acquisition of invention document standard provisions (step S13), acquisition of invention documents (step S14), acquisition of drawing documents (step S15), citation relationship analysis processing for claims (step S16), and important word extraction processing. (Step S17), inter-element relationship determination processing (Step S18), identification of attribute information for each important phrase (Step S19), processing for determining the importance of each important phrase (Step S20), and translation word candidate set identification processing (step S21), technical field identification processing (step S22), display color determination processing for important words (step S23), warning information generation processing (step S24), and warning information generation processing (step S22). Evaluation (step S51) is performed based on the evaluation value based on the invention document evaluation information in the warning information generated in step S24). For example, if the evaluation in step S51 is inappropriate (that is, the evaluation value does not exceed the threshold) case), the process returns to step S14 and another invention document is acquired. If the evaluation in step S51 is appropriate (that is, the evaluation value exceeds the threshold), the invention document analysis system 20 performs a process of generating patent application related information (step S52) and a process of generating a display document (step S52). S53), group classification of invention documents (step S54), display of patent application related information (step S55), acceptance of selection of translation word candidate set elements (step S56), and input of important words and selection results into training data. (step S57), generation of a learning model for important word/phrase translation (step S58), and output of a learning model for important word/phrase translation (step S59). Note that the example in FIG. 58 is only an example, and the invention document analysis system 20 may execute a process in which some of the steps shown in FIG. 58 are omitted or changed as the important term translation learning model generation process. However, some processing steps may be added and executed to the plurality of steps shown in FIG. 58, or the execution order of the plurality of steps shown in FIG. All or some of the steps shown in 58 may be performed in parallel. The invention document analysis system 20 can, for example, repeat the processing in step S56 and step S57 in FIG. 58 multiple times depending on the user's operation. Further, the invention document analysis system 20 may repeatedly execute the processes in steps S14 to S57 in FIG. 58 a certain number of times or for a certain period of time, and then execute the processes in steps S58 and S59.

The processing from step S11 to step S24 in the example of FIG. 58 is similar to the processing from step S11 to step S24 in the invention document analysis process described above with reference to the example of FIG.

In the evaluation in step S51 regarding the evaluation value based on the invention document evaluation information in the warning information generated in the warning information generation process (step S24), the document evaluation unit 2071 of the machine learning unit 207 Based on whether or not the evaluation value included in the invention document evaluation information exceeds a threshold value, it is evaluated whether the invention document is appropriate enough to be used for generating a learning model for important word/phrase translation.

The generation of patent application related information (step S52) in the example of FIG. 58 is the process of step S30 that is executed after the processes of steps S25 to S29 in the invention document analysis process described above in accordance with the example of FIG. It is similar to

The display document generation process (step S53) in the example of FIG. 58 is similar to the process in step S31 in the invention document analysis process described above with reference to the example of FIG.

In the invention document group classification (step S54), the classification unit 2053 classifies the important word group extracted in the important word extraction process (step S17) for the invention document currently acquired in step S14 and from each of the plurality of invention documents. Group classification is performed based on the correlation with the extracted important word group, and the group to which the currently acquired invention document belongs is specified in step S14, or the group is classified by a user (for example, a translator). The group to which the invention document belongs is specified according to the specification. Note that the group specified for the invention document by the group classification in step S54 may be, for example, a technical field.

In displaying patent application related information (step S55), the display unit 2084 having a browser function interprets the display document (for example, an HTML document) for patent application related information, and displays the patent application on the display device. Displays application related information (see Figure 17).

In accepting the selection of a translation word candidate set element (step S56), the operation reception unit 20841 selects the user ( For example, a selection operation by a translator is accepted. For example, the process in step S57 may be executed each time an operation for selecting an element from the translation word candidate set for each important phrase is accepted.

In storing important words and selection results as training data (step S57), the training data storage unit 2072 of the machine learning unit 207 accepts important words and phrases from the translation word candidate set corresponding to the important words in step S56. A set of elements (that is, corresponding translated words corresponding to the important word/phrase) selected by the operation is stored in a recording medium or the like as teacher data for machine learning to generate a learning model for important word/phrase translation. Note that the teacher data storage unit 2072 may store the teacher data in association with the group to which the invention document identified in step S54 belongs. Further, the teacher data storage unit 2072 generates a learning model for important word/phrase translation by using, for example, a set of an important word/phrase with a corresponding translation word and its corresponding translated word in the important word/phrase setting dictionary acquired in step S12. It may be stored in a recording medium etc. as teacher data for machine learning.

In the generation of the learning model for important word/phrase translation (step S58), the machine learning unit 207 generates a plurality of machine learning models for generating the learning model for important word/phrase translation, which are stored in the recording medium etc. by the teacher data storage unit 2072. A learning model for machine translation of important words and phrases is generated by machine learning using training data. When the training data storage unit 2072 stores a plurality of training data for machine learning for generating a learning model for important word/phrase translation in a storage medium or the like in association with groups, the machine learning unit 207 By machine learning using multiple training data for machine learning for generating learning models for important word translation for each group, stored in the teacher data storage unit 2072, machine learning for machine translation of important words for each group is performed. Generate a learning model for translating important words. For example, in a case where groups are classified by the department to which the inventor belongs, the important term ``receiving unit'' and the corresponding translated word ``receiving unit,'' and the important term ``controlling unit'' and the corresponding translated word ``controlling unit.'' The learning model for important word translation, which is generated by machine learning using training data such as pairs of important words "transmitting unit" and the corresponding translation word "transmitting unit", is based on the translation policy of the department that is the group. For example, if the key word "processing unit" is input and machine translated using the learning model for important word translation, a machine translation result of "processing unit" will be generated. This can be predicted. In addition, for groups related to different departments that have different translation policies, etc., we have created a combination of the important word "receiving section" and the corresponding translated word "receiving circuit", and the important word "control section" and the corresponding translated word "controlling circuit". The training model for important word translation, which is generated by machine learning using training data such as pairs of important words ``transmitting part'' and the corresponding translated word ``transmitting circuit'', reflects the translation policy of the group. If the key word "processing part" is input and machine translated using the learning model for important word translation, it is predicted that a machine translation result of "processing circuit" will be generated, for example. obtain. In addition, for groups related to different departments that have different translation policies, etc., the important word "receiving section" and the corresponding translated word "receiver" are combined, and the important word "control section" and the corresponding translated word "controller" are combined. The group's translation policy is reflected in the training model for important word translation, which is generated by machine learning using training data such as the pair of important words "transmitter" and the corresponding translated word "transmitter", etc. There is a possibility that if the key word phrase "processor" is input and machine translated by arithmetic processing using the key word translation learning model, it can be predicted that, for example, a machine translation result of "processor" will be generated. The machine learning unit 207 stores and manages the generated learning model for important word/phrase translation in a recording medium or the like by the learning model management unit 2073. The generated important word/phrase translation learning model is used when the invention document analysis system 20 executes the translation word candidate set identification process (step S21) in the invention document analysis process (see FIG. 40). It can be used to determine the elements of a candidate set, or to machine translate important phrases when performing various translation information generation processes (step S28).

In outputting the learning model for important word/phrase translation (step S59), the learning model output unit 2081 outputs the learning model for important word/phrase translation held by the learning model management unit 2073 as the contents of a file according to the user's specifications, for example. Output. The output learning model for important word/phrase translation is used when another invention document analysis system 20 executes the generation process (step S28) of various translation information in the invention document analysis process (see FIG. 40). A machine translation system or machine translation system for the translation of invention documents that can be used for machine translation, and that machine translates the important words and phrases in the invention document and then uses the results to machine translate the entire invention document. program or general machine translation system.

[3-4. Translation learning model generation processing example according to Embodiment 2]
The invention document analysis system 20 can execute a translation learning model generation process that generates a translation learning model for machine translation of an invention document. An example of the translation learning model generation process will be described below.

FIG. 59 is a flowchart illustrating an example of a translation learning model generation process.

As shown in FIG. 59, the invention document analysis system 20 performs, for example, a setting information acquisition process (step S11), an important phrase setting dictionary (step S12), and a translation learning model generation process. Acquisition of document standard regulations (step S13), acquisition of invention documents (step S14), acquisition of drawing documents (step S15), citation relationship analysis processing for claims (step S16), and important phrase extraction processing (step S17), inter-element relationship determination processing (step S18), identification of attribute information for each important term (step S19), importance determination process for each important term (step S20), and warning information generation process (step S17). S24) and an evaluation (step S51) related to the evaluation value based on the invention document evaluation information in the warning information generated in the warning information generation process (step S24), and for example, in the evaluation in step S51, an inappropriate (in other words, when the evaluation value does not exceed the threshold value), the process returns to step S14 and another invention document is acquired. If the evaluation in step S51 is appropriate (that is, the evaluation value exceeds the threshold), the invention document analysis system 20 generates a revised version of the invention document (step S61) and obtains the translation result document (step S62). , group classification of invention documents (step S63), storage of revised invention documents and translation result documents as teacher data (step S64), generation of a learning model for translation (step S65), and generation of a learning model for translation. output (step S66). Note that the example in FIG. 59 is merely an example, and the invention document analysis system 20 may execute a process in which some of the steps shown in FIG. 59 are omitted or changed as the translation learning model generation process, or Some processing steps may be added to the plurality of steps shown in FIG. 59 and executed, the execution order of the plurality of steps shown in FIG. 59 may be changed and executed, or the steps shown in FIG. All or some of the steps shown may be performed in parallel. The invention document analysis system 20 may, for example, repeatedly execute the processes in steps S14 to S64 in FIG. 59 a certain number of times or for a certain period of time, and then execute the processes in steps S65 and S66.

Steps S11 to S20 and step S24 in the example of FIG. 59 are the same as steps S11 to S20 and step S24 in the invention document analysis process described above with reference to the example of FIG. The process in step S63 in the example of FIG. 59 is similar to the process in step S54 in the example of FIG.

In the evaluation in step S51 regarding the evaluation value based on the invention document evaluation information in the warning information generated in the warning information generation process (step S24), the document evaluation unit 2071 of the machine learning unit 207 Based on whether or not the evaluation value included in the invention document evaluation information exceeds a threshold value, it is evaluated whether the invention document is appropriate enough to be used for generating a learning model for important word/phrase translation.

In generating a revised invention document (step S61), the machine learning unit 207 integrates the revised claim information generated by the claim information generation unit 205a and the revised specification information generated by the specification information generation unit 205b. This generates a revised invention document. The revised invention document is generated based on the invention document for which the evaluation value included in the invention document evaluation information exceeds the threshold in step S51, but depending on the predetermined threshold, the revised invention document Since there is a possibility that the content of the invention document has higher consistency with the translation result document, a modified version of the invention document is generated.

In the acquisition of the translation result document (step S62), the translation result document acquisition unit 2023 obtains a translation that is the translation result corresponding to the invention document currently acquired in step S14 by receiving a specification from the user (for example, a translator). Get the result document.

In storing the revised invention document and the translation result document as training data (step S64), the training data storage unit 2072 of the machine learning unit 207 stores the revised invention document generated in step S61 and the translation result document acquired in step S62. The set with the translation result document is stored in a recording medium or the like as teacher data for machine learning to generate a learning model for translation. The teacher data storage unit 2072 may store the teacher data in association with the group to which the invention document identified in step S63 belongs. If the revised invention document generated in step S61 has a plurality of claim sets in the revised claim information generated by the claim information generation unit 205a, the teacher data storage unit 2072 stores multiple claim sets in the revised invention document. After removing all claim sets from the revised invention document except for one claim set that is highly consistent with the translation result document, the combination of the revised invention document and the translation result document is used as a translation learning model. It is stored in a recording medium, etc. as training data for machine learning for generation. As an example of a specific method for identifying one claim set that is highly consistent with the translation result document among multiple claim sets in the revised version of the invention document, The sameness of the number of terms, the sameness of the number of each cited term in the statement of each term of the claim, and the sameness of whether each term of the claim contains a statement that includes a symbol in parentheses. An example of identifying the claim set in the revised invention document that has the highest overall degree of consistency, for example, all of its identity items are the same, as a claim set with high consistency. . The teacher data storage unit 2072 uses a set of a revised invention document with the code description removed and a translation result document with the code description removed as a machine learning teacher for generating a translation learning model. It may also be stored as data on a recording medium or the like. In addition, the teacher data storage unit 2072 stores, for each sentence in the revised invention document, a set of that sentence and a sentence in the translation result document corresponding to the sentence as training data for machine learning to generate a learning model for translation. It may also be stored on a recording medium or the like.

In the generation of the learning model for translation (step S65), the machine learning unit 207 collects a plurality of teaching data for machine learning for generating the learning model for translation, which is stored in a recording medium etc. by the teaching data storage unit 2072. Using machine learning, we generate a translation learning model for machine translation of invention documents. When the teacher data storage unit 2072 stores a plurality of pieces of teacher data for machine learning for generating a learning model for translation in a storage medium or the like in association with groups, the machine learning unit 207 stores the teacher data Translation learning for machine translation of invention documents is performed for each group by machine learning using a plurality of training data for machine learning for generating translation learning models for each group, which are accumulated in the storage unit 2072. Generate the model. The machine learning unit 207 stores and manages the generated translation learning model in a recording medium or the like by the learning model management unit 2073. The generated translation learning model is used when the invention document analysis system 20 executes the generation process (step S28) of various translation information in the invention document analysis process (see FIG. 40), the claim information related to the invention document, It can be used for machine translation of specification information, etc.

In outputting the learning model for translation (step S66), the learning model output unit 2081 outputs the learning model for translation held by the learning model management unit 2073, for example, as the contents of a file according to the user's designation. The output learning model for translation is used when another invention document analysis system 20 executes the generation process (step S28) of various translation information in the invention document analysis process (see FIG. 40), etc. It can be used to machine translate information, specification information, etc., and it can be used in a machine translation system or a machine translation program for translating invention documents, or in a general machine translation system.

[3-5. Effects of invention document analysis system according to Embodiment 2]
Since the invention document analysis system 20 is a partial modification of the invention document analysis system 10 shown in the first embodiment, it can exhibit many of the effects of the invention document analysis system 10 described above.

The invention document analysis system 20 uses an analysis unit 205 to extract important terms from an invention document created by an inventor or a patent attorney in an important term extraction unit 2051 based on setting information, and performs various analyzes based on each important term. The output unit 208 outputs patent application related information including various information generated as analysis results. Therefore, for translators, etc., the understanding of the invention document is facilitated by referring to the patent application related information. There is a possibility that Facilitating the understanding of invention documents will be useful for translators to appropriately or efficiently translate invention documents.

The patent application related information generated based on the analysis result of the invention document in the invention document analysis system 20 is displayed by the display document generation unit 2083 or the like, for example, by highlighting headings, comments, important words, codes, etc. The output unit 208 outputs a display document such as an HTML document to which display control information for displaying GUI elements that performs various searches based on important words or the like or selection of corresponding translation words from a set of translation word candidates is added. Translators and others can view patent application-related information on various terminal devices with a browser function, and can improve their invention by highlighting important words, etc., and operating GUI elements related to important words, etc. It may be possible to quickly understand a document, and it is possible to efficiently determine corresponding translation words corresponding to important words and phrases. Further, since the patent application related information as a display document (for example, an HTML document) generated based on the analysis result of the invention document in the invention document analysis system 20 is displayed on the display unit 2084, it is useful for translators etc. By being able to view patent application related information, highlighting important words, etc., and operating GUI elements related to important words, it is possible to quickly understand invention documents, and it is possible to determine corresponding translation words corresponding to important words. It can be done efficiently. Note that the content of patent application related information displayed by a browser etc. can be subject to copy and paste, so translators etc. can copy and paste the necessary content and paste it into the text. It can be edited using an application program such as an editor that has a document creation and editing function.

The invention document analysis system 20 outputs, for example, patent application related information generated based on the analysis result of the invention document as an HTML document (for example, an HTML file, a file specified by the HTML file, etc.). This makes it possible to view patent application-related information using various browsers that support HTML, so translators who receive patent application-related information as an HTML document from an inventor or patent attorney can use special Patent application related information can be viewed using a terminal device having a browser function (for example, a PC running a browser as an application program) without having a system, a special application program, or the like.

Patent application related information as a display document (for example, an HTML document) generated based on the analysis result of an invention document in the invention document analysis system 20 includes, for example, claim information representing a claim in an invention document, and a specification in an invention document. Each important word/phrase stated in the specification information, etc. is displayed in a display mode (for example, display color) to identify whether it is an important word/phrase stated in the claim, and important words with different contents are displayed in different colors. Since the words are displayed with a left border line, etc., it may be possible for a translator or the like to quickly grasp each important word or phrase.

The patent application related information generated and outputted by the invention document analysis system 20 based on the analysis results of the invention document includes, for example, warning information, and the warning information indicates that the invention document conforms to the invention document standard provisions for translation. It includes indication information that points out when an inappropriate description such as not being included is included, and invention document evaluation information that includes an evaluation value related to the conformity of the invention document to the invention document standard provisions for translation. For the inventor or patent attorney who created the invention document, by referring to the warning information, it may be possible to modify the invention document to make it easier to translate. In addition, for example, a translator who receives an invention document and patent application related information from an inventor or a patent attorney and is asked to translate it can expect to be provided with an appropriate invention document, and can translate the invention document. It may be possible to translate properly or efficiently. For example, if the translator has the invention document analysis system 20, the translator may receive and use the invention document, setting information, important word setting dictionary, and drawing document from the inventor or patent attorney. Therefore, by outputting patent application related information by the invention document analysis system 20 and referring to the patent application related information, it may be possible to appropriately or efficiently translate the invention document. When it is assumed that a translator uses the invention document analysis system 20, or when it is assumed that patent application related information is provided to the translator, the translator who receives and translates the invention document By viewing application-related information, it may be possible to translate invention documents appropriately or efficiently, and by checking warning information in patent application-related information, inventors, etc. or patent attorneys can translate the warning information. It may be possible to know that the situation is judged to be negligible. The inventor, etc. or patent attorney may provide the translator with information that specifies corresponding translation words for important words and phrases in the invention document (for example, a dictionary for setting important words), as necessary, and the translator will be able to By referring to information related to corresponding translation words for a word or phrase, it may be possible to appropriately or efficiently translate the important word or phrase. In addition, the inventor, etc. or a patent attorney may use the information specifying the corresponding translation words for important words and phrases in the invention document as important words and phrases in the patent application related information output as a result of having the invention document analyzed by the invention document analysis system 20. By referring to, editing, etc. a list (for example, a list listing pairs of important words and translation word candidate sets), it is possible to generate the list relatively easily. In addition, translators, etc. can view patent application related information as a display document displayed on a browser, etc., and perform operations on important words, etc. displayed as operation-compatible GUI elements in the patent application related information. The content of important words, etc. in one or more other invention documents related to the invention document that is the basis of the patent application related information, or the translation assistance information, etc. in one or more other invention documents. It is possible to check the parts related to important words, etc. in the drawing document corresponding to the invention document that is the basis of the patent application related information, and the definition of important words, etc. displayed on search sites etc. You can check translated words, etc. As a result, translators and the like can sufficiently understand the important words and phrases of the invention document and translate it appropriately and efficiently.

Patent application related information as display documents generated based on the analysis results of invention documents in the invention document analysis system 20 includes, for example, claim information, revised claim information, and specification information with important words and phrases highlighted. , a specification heading list, revised specification information, a claim tree, a revised claim tree, a constituent claim tree, and a revised constituent claim tree, so translators and others who view patent application related information will find the invention document useful. It may be possible to quickly grasp the contents of the document. For example, by viewing the specification heading list in the patent application related information, an inventor or a patent attorney can easily determine whether or not the specification heading of an invention document is appropriate. For example, an inventor or a patent attorney may specify, in setting information, information specifying a region (for example, the United States) where a claim can be filed, and then use the patent application-related information output by the invention document analysis system 20. By simply informing the translator to refer to the revised claim tree or the revised structured claim tree and adopt the revised claim information, the translator can, for example, Therefore, it is possible to clearly understand how to modify the description of cited items in each claim section of an invention document based on the modified claim information, the modified claim tree, or the modified constituent claim tree. It could be possible.

The patent application related information generated and output based on the analysis results of invention documents in the invention document analysis system 20 includes, for example, an important word list, important word related information (for example, inclusion related information, record related information, program implementation related information). information, equivalence relation information, and superordinate concept relation information), and important word/phrase relation translation information (e.g., inclusion relation translation information, record relation translation information, program realization relation translation information, equivalence relation translation information, and superordinate relation relation) Translation information), it may be possible for a translator or the like to appropriately or efficiently determine a corresponding translated word corresponding to an important word or phrase. The patent application related information as a display document generated based on the analysis results of invention documents in the invention document analysis system 20 includes, for example, an important word list in which important words are associated with a set of translation word candidates. Display control information is added to the display element of the set to provide a GUI function to determine one element from the set of translation word candidates as a corresponding translation word according to the operation, so translators etc. can easily use the browser etc. By selecting the elements of the translation word candidate set displayed by , it may be possible to efficiently determine the corresponding translation word corresponding to the important word/phrase.

The patent application related information generated and outputted by the invention document analysis system 20 based on the analysis result of the invention document includes, for example, a code list and a drawing code list, so that for translators etc., important words and phrases corresponding to the codes are useful. It may be possible to efficiently understand the information and efficiently refer to the contents of the drawings corresponding to the reference numerals.

The patent application related information generated based on the analysis results of invention documents in the invention document analysis system 20 includes, for example, claim translation information, revised claim translation information, specification translation information, and revised specification translation information. By referring to this information, a translator may be able to translate an invention document appropriately or efficiently.

The patent application related information generated and output based on the analysis results of invention documents in the invention document analysis system 20 may include, for example, claim-based drawing information, claim-based specification information, and claim-based abstract information. . As a result, for example, an inventor or a patent attorney who has created an invention document that includes claims but does not include a specification and has it analyzed by the invention document analysis system 20 can obtain claim-based drawing information, claim-based specification information, and claim-based drawing information. By referring to, modifying, and using the base abstract information, it may be possible to efficiently generate drawing documents and invention documents including claims, specifications, and the like. For example, an inventor or a patent attorney may be able to efficiently generate a specification in an invention document by modifying the unfinished message part for writing support written in a predetermined format in claim-based specification information. There is. The inventor, etc. or a patent attorney causes the invention document analysis system 20 to analyze the invention document generated by referring to the claim-based drawing information, claim-based specification information, and claim-based summary information, so that the invention document is output. It may be possible to more appropriately modify the invention document by referring to the patent application related information.

The patent application related information generated and outputted by the invention document analysis system 20 based on the analysis results of the invention document includes, for example, technical field information, fee information, and aggregate information, so that the inventor or patent attorney can By referring to the patent application related information, the invention document can be modified as necessary.

The invention document analysis system 20 includes, for example, warning information of patent application related information generated and output about invention documents based on the acquired standard provisions for translation of invention documents, based on a web site that publishes the standard provisions for invention documents for translation. It may also include regulatory access information necessary to access the standard provisions for translation of invention documents, such as the network address of the site. Thereby, a translator or the like can access the standard regulations for invention documents for translation and view their contents by referring to the regulation access information in the warning information.
For example, the invention document analysis system 20 acquires a new version of the standard provisions for invention documents for translation, if available, and generates and outputs patent application-related invention documents based on the acquired standard provisions for invention documents for translation. The warning information of the information includes, for example, version information indicating the version of the standard specification of invention documents for translation used. As a result, it is possible to respond to the fact that the standard provisions for invention documents for translation are refined and revised over time, and translators can check the version by referring to the version information of the standard provisions for invention documents for translation in the warning information. It will be possible to check whether the invention document standard provisions for translation have been used.

The invention document analysis system 20 includes, for example, a program module responsible for processing for determining whether an invention document conforms to each of a plurality of items of the invention document standard provisions for translation, and a program module responsible for extracting important words. Each plug-in module, which is a program module, is acquired and used, the version information of the plug-in module is managed, and if a new version of the plug-in module is available, it is acquired and used. This allows new plug-in modules that have been refined over time to be used, making it possible to perform various processes more appropriately.

A learning model output unit 2081 that outputs various learning models generated by the machine learning unit 207 of the invention document analysis system 20 outputs, for example, a technical field identification learning model. By using this technical field identification learning model, it may be possible to identify the technical field to which an invention document belongs by performing predetermined arithmetic processing using a group of important words extracted from an invention document as input. .

The learning model output unit 2081 of the invention document analysis system 20 outputs, for example, the learning model for important phrase translation generated by the machine learning unit 207. By using this learning model for important word translation, it is possible to obtain machine translation results for important words and phrases by inputting each important word extracted from an invention document and performing predetermined arithmetic processing. obtain.

The learning model output unit 2081 of the invention document analysis system 20 outputs the translation learning model generated by the machine learning unit 207, for example. By using this learning model for translation, it may be possible to obtain a machine translation result of an invention document by performing predetermined arithmetic processing using the invention document as input.

The invention document analysis system 20 classifies invention documents into groups by a classification unit 2053 based on, for example, the correlation between a group of important words extracted by the important word extraction unit 2051 from each of a plurality of invention documents, or based on a user's designation. is executed, and the machine learning unit 207 generates a learning model for important word/phrase translation for each group by machine learning using a plurality of pairs of important words/phrases extracted from the invention document and corresponding translated words as training data. As a result, it may be possible to machine translate each important term in an invention document through arithmetic processing using a learning model for important term translation that is suitable for the group to which the invention document belongs. In addition, the invention document analysis system 20, for example, machine-translates important words and phrases extracted from invention documents through arithmetic processing using a learning model for important words and phrases translation that is suitable for the group to which the invention document belongs. It is also possible to output patent application related information including a list of important words that enumerate pairs with a set of translation word candidates that include machine translation results. By referring to this list of important words, the translator may be able to efficiently or appropriately translate the important words and phrases of the invention document. The invention document analysis system 20 may generate a learning model for important word/phrase translation without performing group classification. Even if a learning model for translation of important words that is not classified into groups is used, there is a possibility that appropriate machine translation of important words in invention documents can be achieved.

The invention document analysis system 20 classifies invention documents into groups by a classification unit 2053 based on, for example, the correlation between a group of important words extracted by the important word extraction unit 2051 from each of a plurality of invention documents, or based on a user's designation. The machine learning unit 207 generates a learning model for translation for each group by machine learning using a set of an invention document and a corresponding translation result document as training data. As a result, in a machine translation system, for example, it becomes possible to machine translate an invention document through arithmetic processing using a translation learning model suitable for the group to which the invention document belongs, among translation learning models for each group. In addition, the invention document analysis system 20 may machine translate at least a portion of the invention document through arithmetic processing using, for example, a translation learning model suitable for the group to which the invention document belongs among the translation learning models for each group. , patent application related information including the machine translation results may be output. By referring to the results of machine translation of such an invention document, a translator may be able to efficiently or appropriately translate the invention document. The invention document analysis system 20 may generate a learning model for translation without performing group classification. Even if a translation learning model that is not classified into groups is used, it is possible to achieve appropriate machine translation of invention documents.

For example, the invention document analysis system 20 analyzes invention documents used for accumulating training data for machine learning to generate a learning model for translation or a learning model for important word translation, according to the evaluation result by the document evaluation unit 2071. This makes it possible to generate an appropriate translation learning model or important word/phrase translation learning model. Being able to generate an appropriate translation learning model or important term translation learning model can be useful for improving the accuracy of machine translation using these learning models.

The invention document analysis system 20 may, for example, generate a set of a modified invention document generated based on an invention document and a translation result document, or analyze each sentence of the modified invention document and its corresponding translation. The combination of the resulting document and the sentence may be used as training data for machine learning, which may generate a more appropriate translation learning model. The invention document analysis system 20, for example, removes codes from the invention document and the translation result document used as the teacher data and uses it as the teacher data, so that it corrects the omission of codes in the invention document at the time of translation and improves the translation result. Discrepancies between the content of both documents that occur when two documents are created can be suppressed, making it possible to generate an appropriate translation learning model, and using the appropriate translation learning model can improve the accuracy of machine translation. can be useful for improvement.

The invention document analysis system 20 is configured, for example, by installing an invention document analysis processing program for executing an invention document analysis process as illustrated in FIG. 40 into a device including a microprocessor. When the invention document analysis processing program is started in the invention document analysis system 20, invention document analysis processing is executed. The invention document analysis processing program may include data used for invention document analysis processing. For translators and the like, understanding of the invention document may be facilitated by referring to the information output by the invention document analysis process (for example, patent application related information as a display document). Facilitating the understanding of invention documents will be useful for translators to appropriately or efficiently translate invention documents.

(Other embodiments, etc.)
As described above, the invention document analysis systems 10, 20, etc. have been described in Embodiments 1 and 2 as examples of the technology in the present disclosure. However, the technology in the present disclosure is not limited to this, and can be embodied as an embodiment in which modifications, substitutions, additions, omissions, etc. are made to the first or second embodiment described above as appropriate.

In the embodiments described above, an example was given in which the invention document analysis systems 10 and 20 acquire setting information, invention document standard provisions for translation, important word/phrase setting dictionary, invention documents, drawing documents, translation result documents, etc. Acquisition of such various information may be realized by any method. Acquisition of various types of information may be realized, for example, by receiving information signals from an external communicable device, or by reading information from a removable recording medium such as a memory card. , may be the acquisition of information input from a user (e.g., inventor, patent attorney, translator, etc.) via a keyboard, mouse, voice input device, or other input device, and includes imaging by an image sensor. Information may be acquired as sensing results by various sensors. Further, in the above-described embodiments, an example was shown in which the invention document analysis systems 10 and 20 accept operations on operation-compatible GUI elements in display documents, but acceptance of such operations may be realized by any method. good. For example, the operation can be accepted by obtaining some information input from the user (e.g., inventor, patent attorney, translator, etc.) via a keyboard, pointing device, touch pad, voice input device, or other input device. It may be possible to obtain some information (for example, a predetermined gesture) obtained from the user as a result of sensing by various sensors including imaging by an image sensor. Furthermore, the user described in the above-described embodiments is not necessarily limited to a human, and may be a robot or the like having some functions similar to those of a human.

Further, the division of functions among the above-mentioned parts in the invention document analysis systems 10 and 20 described above is only an example, and can be arbitrarily changed, and a plurality of parts may be integrated. In addition, the invention document analysis systems 10 and 20 perform part of the invention document analysis processing, the processing related to the generation of various learning models, etc. to an external device that can communicate with the devices that constitute the invention document analysis systems 10 and 20. You may share the work. In addition, all or part of the functions of each part in the invention document analysis systems 10 and 20 may be implemented by IC (Integrated Circuit), ASIC (Application Specific Integrated Circuit), FPGA (Field Programmable Gate Array), LSI (Large Scale Integration), etc. It may be realized by a semiconductor device, and the semiconductor device may include a memory and a microprocessor in which a program for realizing all or part of the functions of each part is recorded. All or part of the various processes in the invention document analysis systems 10 and 20 may be realized using hardware such as a dedicated electronic circuit, or may be realized using a microprocessor and software (for example, programs and information used in the programs). It's okay. A program for realizing all or part of the processing of the invention document analysis systems 10 and 20 (for example, an invention document analysis processing program) is recorded on a computer-readable non-temporary recording medium such as a ROM, an optical disk, or a hard disk. It can be distributed or distributed. When a program for realizing all or part of the processing of the invention document analysis systems 10 and 20 is distributed, the program is installed in a device equipped with a microprocessor and the microprocessor of the device to which it is installed is installed. By executing the process, it becomes possible to cause the device to perform all or part of the process. Note that the program may be stored in advance in a recording medium included in the invention document analysis systems 10 and 20, or, for example, may be distributed and supplied to the recording medium via a wide area communication network including the Internet. It's okay. Further, regardless of the method by which the computer performs processing, the computer may be regarded as a device equipped with a microprocessor. Further, a program for realizing all or part of the processing of the invention document analysis systems 10, 20, such as the invention document analysis processing program, may be a program written in any language, and may be a program written in an interpreter or an execution time program. The program may be a program executed via a compiler (Just-In-Time Compiler) or the like, or a program that operates on an OS or various existing execution environments. Further, the invention document analysis processing program may be composed of a plurality of programs, and each of the plurality of programs may be distributed on a plurality of computers and run on a plurality of mutually different execution environments. It may be. In order to realize all or part of the processing of the invention document analysis systems 10 and 20, a plurality of programs written in different languages may work together.

In addition, the above-described invention document analysis system 20 extracts a plurality of important words from the claims of an invention document by the important word extraction unit 2051, and based on the structural inclusion relationship between the elements represented by each of the extracted plurality of important words. Then, a code determination process is executed to determine the codes corresponding to some of the plurality of important words and phrases, and various relationships between elements expressed by each of the plurality of extracted important words and phrases (for example, relationships expressed by important word and phrase relationship information) are executed. Based on this, a drawing content determination process is executed to determine important words and phrases related to the contents to be written in each drawing that can be identified by the drawing number, and each important word and phrase extracted from the claims of the invention document and A specification content determination process that determines the content to be included in the specification based on the results of the code determination process and the drawing content determination process, and outputs the specification with the content determined in the specification content determination process. It is also possible to perform specification content output processing. In the specification content determination process, the invention document analysis system 20 generates an invention name character string to be the content of the "Name of the Invention" column based on the important phrases that are claimed in the claims of the invention document. Based on the results of the determination process and the drawing content determination process, for each important term related to the content to be described in each figure in the order of the drawing number, a code is attached to the important term and the attributes of the important term (e.g. A sentence that explains the important word/phrase according to the attribute (attribute represented by the attribute information shown in FIG. 19) (for example, a sentence added to the important word in the important word/phrase setting dictionary, or a sentence including an incomplete message for writing support) may be generated as the contents of the "Detailed Description of the Invention" column. In addition, the invention document analysis system 20 extracts a plurality of important words from the invention document by the important word extraction unit 2051, and when there is a structural inclusion relationship between the elements represented by each of the extracted plural important words, For example, inclusion relationship information (see FIG. 20), such as a diagram showing the inclusion relationship between elements, may be output. The invention document analysis system 20 also generates a claim tree including lines or symbols expressing citation relationships between claim terms based on the claims of the invention document, and generates a claim tree including lines or symbols expressing citation relationships between claim terms. or the claim tree is displayed using a symbol as a GUI element, and in response to the user's operation on the GUI element, the claim tree is disconnected from the citation relationship between the clauses of the claim expressed by lines or symbols related to the GUI element. Modified claim information including each clause may be generated and the modified claim information may be output. In addition, the invention document analysis system 20 extracts a plurality of important words from the claims of the invention document by the important word extraction unit 2051, and displays each of the extracted plurality of important words in different display modes (for example, A display mode in which a left border line of a different color and width is added to an important word or phrase) is displayed as a GUI element, and in response to a user's operation on the GUI element, the invention document containing the important word or phrase related to the GUI element is displayed as a GUI element. Alternatively, depending on the user's operation on the GUI element, a list of each line or sentence of the GUI element may be displayed, and in response to the user's operation on the GUI element, an invention document other than the invention document (e.g. It may also be possible to display a list that enumerates each line or each sentence in the invention document (invention document specified by the user), or it may display a list that enumerates each line or sentence in the invention document specified by the user, or it may display a list of important words and phrases included in the important phrase management information (see Figure 18) regarding the important phrases related to the GUI element. Information may also be displayed. The invention document analysis system 20 also displays specification information representing the specification of the invention document, displays codes in the specification information as GUI elements, and displays the GUI elements in response to user operations on the GUI elements of the codes. It may also be possible to display the drawing in which the code in the drawing document containing the code related to is written or the drawing number of the drawing.

Furthermore, when the translation result document acquisition unit 2023 acquires a translation result document generated as a result of translation based on an invention document, the invention document analysis system 20 described above outputs translation result information representing the translation result document. It may be included in the patent application related information as a display document such as an HTML document output by the unit 208. In this case, the invention document analysis system 20 uses the important word translation model to calculate the translated words in the translation result document corresponding to each important word extracted from the invention document by the important word extraction unit 2051. The translation results of the document for display are identified by machine translation results based on the key words, arithmetic processing based on the important word setting dictionary, translation auxiliary information, etc., or processing to search for descriptions with the same symbol as the important words. The specified translated word/phrase in the information part is used as a display element, and the display element of the translated word/phrase is combined with display control information added to the display element of the important word/phrase in the claim information part, specification information part, etc. of the display document. It is also possible to add display control information that performs similar display mode control (for example, highlighting by adding a colored left border line with the same display color, converting display elements into GUI elements corresponding to operations, etc.). By converting the display elements of translated words corresponding to important words into GUI elements that support operations, a function that is activated in response to the operation can replace all translated words corresponding to the same important word with the word specified by the user. It may be possible to realize processing such as When the invention document analysis system 20 includes translation result information in patent application related information as a display document such as an HTML document outputted by the output unit 208, a heading whose description format is determined in the translation result information. (For example, the heading of each claim section, the paragraph number of the specification) as a display element, and depending on the operation, the display element is changed to the corresponding heading part of the corresponding claim information, specification information, etc. (For example, the content of the corresponding claim section, the content of the corresponding paragraph of the specification, etc.) may be added with display control information for making it into a GUI element that displays the content. For example, when translation result information is displayed in the "content display column of patent application related information items according to menu selection" shown in FIG. 17, the display element of the heading that is the paragraph number of the specification in the displayed content is operated (for example, by a mouse click operation), the corresponding paragraph number in the specification information is displayed in the "display column for items corresponding to the operation of the operation-compatible GUI element" shown in FIG. 17. Script functions related to display control information added to display elements can be defined in advance. For example, the translator can check the validity of the translated document and respond to important words by viewing and operating the displayed content in the translated result information section of patent application related information as a display document displayed on a browser, etc. It may be possible to efficiently modify translated words and phrases. The invention document analysis system 20 also calculates the number of claims in the translation result document, the number of lines, sentences, and paragraphs in the specification of the translation result document, or the number of lines and sentences in the translation result document. The total number of words, the number of translated words corresponding to important words in the embodiment description part of the specification in the translated document, the number of codes written in the translated document, etc. Then, each count result may be expressed in a distinguishable manner and included in the total information in the patent application related information output by the output unit 208. The translator can, for example, refer to aggregate information on patent application related information and compare the number of lines of the specification in the invention document with the number of lines of the specification in the translation result document. It may become possible to efficiently check validity, etc.

Furthermore, the invention document analysis system 20 described above executes invention document analysis processing, and in the invention document analysis processing, for example, acquires setting information including condition information indicating conditions for extracting important words and phrases, and analyzes technical ideas. Acquires an invention document that is a document that describes claims or specifications for a patent application expressed in Japanese, performs an analysis process to analyze the invention document based on the setting information, and generates a file based on the result of the analysis process. The analysis process includes an extraction process that extracts important words and phrases from the invention document based on the condition information. Information generated based on each word may be output.

Furthermore, embodiments realized by arbitrarily combining the components and functions shown in the embodiments described above are also included within the scope of the present disclosure.

(Modified example group 1)
The invention document analysis system in the present disclosure includes a setting information acquisition unit that acquires setting information including condition information indicating conditions for extracting important words, and a claim or specification for a patent application that expresses a technical idea in Japanese. a document acquisition unit that acquires an invention document that is a written document; an analysis unit that analyzes the invention document acquired by the document acquisition unit based on the setting information; and an invention document acquired by the document acquisition unit. an output unit that outputs information generated based on the analysis result by the analysis unit, and the analysis unit extracts important words from the invention document acquired by the document acquisition unit based on the setting information. an invention document analysis unit that includes an important term extraction unit; the output unit outputs information generated based on each of the plurality of important terms extracted by the important term extraction unit from the invention document acquired by the document acquisition unit; It is a system.

Here, for example, the condition information includes important term candidate character string pattern information indicating an important term candidate character string pattern representing a pattern of character strings that are candidates for important terms, and important term candidate character string pattern information that is used for reference to previously described character strings. and reference character string pattern information indicating a reference character string pattern representing a pattern of reference character strings, and the important word/phrase extracting unit extracts information corresponding to the reference character string pattern from the invention document acquired by the document acquisition unit. a first extraction unit that extracts a character string identical to the second character string as an important word when a second character string corresponding to the important word/phrase candidate character string pattern exists following the first character string; , when the invention document acquired by the document acquisition unit includes a claim, and in the claim there is a period character following the third character string that corresponds to the important word candidate character string pattern; , and a second extraction unit that extracts the same character string as the third character string as an important phrase.

As a result, for example, when there is an expression that refers to a word in an invention document, the important word extraction unit treats the referenced word as an important word under certain conditions, and treats the referenced word as an important word under certain conditions. Since words and phrases are treated as important words under certain conditions, there is a possibility that words and phrases that are noted enough to be referenced and words that represent the subject matter of a claim will be treated as important words. It is assumed that there is a strong tendency for a period character to follow the claimed subject matter in the claims of ordinary invention documents. Therefore, since the invention document analysis system can output information by the output unit based on important words and phrases that are estimated to be important in the invention document, the output information can be useful.

For example, the condition information may include code string pattern information indicating a code string pattern representing a pattern of a code string appended to a string representing an element name, and a character that is a predetermined particle. and specific particle string pattern information indicating a specific particle string pattern representing a string pattern, and the important term candidate string pattern information includes at least a predetermined character string whose first character is a kanji, and A predetermined character string in which is a katakana is determined to correspond to the relevant important word candidate character string pattern, and the reference character string pattern information is such that at least the character string "said" corresponds to the relevant reference character string pattern. The code string pattern information is defined such that at least a predetermined character string whose first character is a number corresponds to the code string pattern, and the code string pattern information is defined such that at least a predetermined character string whose first character is a number corresponds to the code string pattern The character string pattern information indicates that at least one of the characters "wo", "ni", "ga", "to", "kara", "he", and "wa" corresponds to the specific particle string pattern. The important word/phrase extraction section extracts the code string pattern into the invention document acquired by the document acquisition section after the fourth character string corresponding to the important word/phrase candidate character string pattern. If there is a fifth character string corresponding to the fourth character string, a third extraction unit extracts the same character string as the fourth character string as an important term, and the important term is added to the invention document acquired by the document acquisition unit. If a seventh character string that corresponds to the specific particle string pattern follows a sixth character string that corresponds to the candidate character string pattern, a character string that is the same as the sixth character string is extracted as an important phrase. It is also possible to include a fourth extraction section.

As a result, the important word/phrase extraction unit treats, for example, a word (for example, a noun phrase, etc.) that precedes a predetermined particle in the invention document as an important word under certain conditions, and treats the word or phrase written with a symbol in the invention document as an important word. Since they are treated as important words under certain conditions, relatively important words such as subject and object, and words representing elements presumed to be illustrated in the drawings related to the invention document can be treated as important words. Since the invention document is written in Japanese, there is a good chance that words starting with kanji or katakana are important words. Therefore, since the invention document analysis system can output information through the output unit based on important words and phrases that are estimated to be important in the invention document, the output information can be useful.

Further, for example, the output unit outputs patent application related information generated based on the analysis result by the analysis unit on the invention document acquired by the document acquisition unit, and the patent application related information is For each of the plurality of important phrases extracted by the important phrase extraction section from the invention document acquired by the acquisition section, the information may include information in which additional information related to the important phrase is added to the important phrase. Examples of additional information include highlighting, information addition display, display control information for executing display transitions and other processes related to important words, corresponding translation words corresponding to important words, and candidates for corresponding translation words for important words. Examples include a set of translation word candidates as a set, attribute information of important words and phrases, and the like.

As a result, translators can grasp multiple important words and phrases in invention documents by obtaining patent application related information output by the invention document analysis system, and translators can use additional information to translate invention documents. Therefore, it may be possible to efficiently or appropriately determine a translated phrase that corresponds to the important phrase and has the same meaning as the important phrase in the destination language.

Further, for example, in the patent application related information, the additional information related to the important words added for each of the plurality of important words extracted from the invention document acquired by the document acquisition unit is The translated word information may be information regarding a corresponding translated word that is a translated word or phrase that expresses the meaning of . The translated word/phrase is, for example, a word/phrase expressed in one type of language other than Japanese (eg, English, Chinese, etc.). The translated word information may be, for example, information representing the corresponding translated word itself, or information representing a set of translated word candidates as a set of candidates for the corresponding translated word.

As a result, by referring to the translated word information included in the patent application related information output by the invention document analysis system, translators can efficiently and appropriately determine translated words that correspond to important words. obtain.

Further, for example, the analysis unit may, for each important term extracted by the important term extraction unit from the invention document acquired by the document acquisition unit, translate a translation word that is a set of corresponding translation word candidates for the important term. The translation word candidate set specifying unit includes a translation word candidate set specifying unit for specifying a candidate set, and the translated word information as the additional information related to the important words of the invention document acquired by the document obtaining unit is It is also possible to include translation word candidate information representing a set of translation word candidates for the important word/phrase identified by.

As a result, by referring to the translation word candidate information included in the patent application related information output by the invention document analysis system, the translator can select one translation word from the set of translation word candidates corresponding to the important word. , it may be possible to efficiently or appropriately determine translated words corresponding to important words.

Further, for example, the document acquisition unit acquires a plurality of invention documents including a first invention document, and the analysis unit analyzes a plurality of important words and phrases using the important word extraction unit for each of the plurality of invention documents. including a classification unit that extracts a group of words and performs group classification for specifying which of the plurality of groups each of the plurality of invention documents belongs to based on the correlation between each extracted important word group; , the translation word candidate set identification unit selects, for each of the plurality of groups, bilingual dictionary information for each group, in which Japanese words and phrases in invention documents belonging to the group are associated with words and phrases in the translation target language. Translation of each of the important words extracted from the first invention document by the important word extraction section using the bilingual dictionary information of the group to which the first invention document is specified to belong according to the group classification of the classification section. It is also possible to specify a word candidate set.

Thereby, in the invention document analysis system, bilingual dictionary information of a group suitable for the invention document is used, so that an appropriate set of translation word candidates can be specified for the important phrase.

Further, for example, the classification unit may utilize the result of extracting a group of important words by the important word extraction unit from each of one or more invention documents known to be related to the technical field for each of a plurality of technical fields. Then, among the plurality of technical fields, any one of the technical fields in which the correlation with the important word group extracted from the first invention document by the important word extraction section is at least a predetermined degree is selected as The translation word candidate set identification unit includes a technical field identification unit that identifies a first technical field corresponding to the first invention document, and the translation word candidate set identification unit identifies Japanese words and phrases in the technical field and the language of the translation target for each of the plurality of technical fields. From among the bilingual dictionary information for each technical field, which is associated with the words and phrases, the bilingual dictionary information for the first technical field specified by the technical field specifying unit is used to identify the important information from the first invention document. It is also possible to specify a set of translation word candidates for each of the important words extracted by the word extraction section.

Thereby, in the invention document analysis system, bilingual dictionary information corresponding to the technical field to which the invention document belongs is used, so that an appropriate set of translation word candidates can be specified for the important words.

Further, for example, the patent application related information may include a plurality of important words and phrases extracted by the important words and phrases extraction unit from the invention document acquired by the document acquisition unit, as content to be displayed. The additional information related to the important word/phrase extracted from the invention document is output as a document (for example, an HTML document), and the additional information is outputted as a document (for example, an HTML document), and the additional information related to the important word/phrase extracted from the invention document is outputted as a document (for example, an HTML document). The display control information included in the additional information related to the important term is displayed as the GUI element so as to be selectable. a script for executing a determination process for determining an element selected according to an operation from the set of translation word candidates for the important word/phrase in a state in which the important word/phrase is translated as a corresponding translated word corresponding to the important word/phrase; A script for executing processing for generating and outputting translation auxiliary information, which is information associating the important word and phrase with the corresponding translated word, for each important word and phrase for which a corresponding translated word has been determined as a result of processing. , the translation word candidate set specifying unit is configured to determine, for each of the one or more invention documents other than the first invention document acquired by the document acquisition unit, based on the analysis result of the invention document by the analysis unit. The generated display document related to the patent application related information is outputted to the output unit and used for display, and the generated display document is generated for each important term in each of the one or more invention documents. The translation word candidate set may be specified for each important word/phrase extracted from the first invention document by the important word/phrase extraction unit based on the translation auxiliary information.

Thereby, the translator can easily determine the corresponding translated word corresponding to the important word/phrase from the set of translated word candidates by selecting the GUI element of the display document. Furthermore, the translator can easily obtain the determined corresponding translated words from the display content of the display document using a browser or the like, for example, by so-called copy and paste, so-called cut and paste, or the like. Copy and paste or cut and paste generally involves selecting data to be copied (that is, duplicating) or cutting (that is, cutting), specifying another location, and pasting that data to that location (that is, cutting). This operation causes the data to be copied or cut to be held in a predetermined memory area under the management of the OS, etc., and the paste operation causes the data to be transferred from that predetermined memory area to the paste destination. will be pasted at the specified location. In addition, the results of the translator's selections on the GUI elements of the display document are used to improve the validity of the invention document analysis system's identification of a set of translation word candidates corresponding to important words and phrases through translation auxiliary information. there is a possibility.

Further, for example, the patent application related information may include a plurality of important words and phrases extracted by the important words and phrases extraction unit from the invention document acquired by the document acquisition unit, as content to be displayed. The additional information related to the important words and phrases that are output as a document and extracted from the invention document are configured such that any element of the set of translation word candidates for the important words can be selected according to an operation. , including display control information for displaying the set of translation word candidates as a GUI element, and the display control information included in the additional information related to the important word/phrase, A script for executing a determination process that determines an element selected according to an operation from the set of translation word candidates for a word as a corresponding translation word corresponding to the important word, and a corresponding result of the determination process. It may also include a script for executing processing for generating and outputting translation auxiliary information, which is information associating each important word and phrase with the corresponding translated word for which a translated word has been determined. .

As a result, it is possible to utilize the translation auxiliary information output by the invention document analysis system, and by using this, it is possible to improve the validity of subsequent identification of a set of translation word candidates corresponding to important words by the invention document analysis system. It could be possible.

Further, for example, the document acquisition unit acquires a plurality of invention documents, and the output unit displays each of the display documents output for each of the plurality of invention documents acquired by the document acquisition unit. A set of important words and corresponding translated words in each of the plurality of translation auxiliary information generated as a result is used as training data for machine learning for important word phrase translation, and the information is generated by machine learning. It is also possible to include a learning model output unit that outputs a learning model for important word/phrase translation, which is a learning model.

Thereby, machine translation of important words and phrases in an invention document can be realized through arithmetic processing using the important words and phrases translation learning model and inputting important words and phrases. By referring to the results of this machine translation, the translator may be able to efficiently or appropriately determine translated words and phrases that are translated important words and phrases in the invention document.

Further, for example, the patent application related information may include a plurality of important words and phrases extracted by the important words and phrases extraction unit from the invention document acquired by the document acquisition unit, as content to be displayed. The setting information is output as a document, and the setting information includes information for specifying an important word setting dictionary that lists important words to be extracted. The invention document analysis system includes an important word/phrase setting dictionary acquisition unit that obtains the important word/phrase setting dictionary based on the setting information, and the important word/phrase extraction unit is configured to correspond to the important word/phrase setting dictionary. If the important words listed in the word setting dictionary are included in the invention document acquired by the document acquisition section, the important words and phrases are extracted from the invention document, and the important word extraction section extracts the important words from the invention document. When an important phrase to which a corresponding translation word is associated in the phrase setting dictionary is extracted from the invention document acquired by the document acquisition unit, the important phrase extracted from the invention document in the patent application related information The additional information related to the word/phrase represents the corresponding translated word, and the important word/phrase extraction unit extracts important words/phrases other than the important word/phrase to which the corresponding translated word is associated in the important word/phrase setting dictionary by the document acquisition unit. In the case of extraction from an acquired invention document, the additional information related to the important word/phrase extracted from the invention document in the patent application related information is one of the candidate translation words for the important word/phrase. The display control information includes display control information for displaying the set of translation word candidates as a GUI element so that the elements can be selected according to the operation, and the display control information includes: A script for executing a determination process that determines an element selected according to an operation from the set of translation word candidates for the important term as a corresponding translation word corresponding to the important term, and as the execution result of the determination process. and a script for executing an update process for updating the important word/phrase setting dictionary so as to include information associating each important word/phrase with the corresponding corresponding translation word for each important word/phrase for which a corresponding translation word has been determined. You can also use it as

As a result, for example, when an inventor or a patent attorney who has created an invention document wants to define a specific word as an important word, or when he or she wants to define a corresponding translation for the important word, the important word can be set. By generating a dictionary, the invention document analysis system can output appropriate patent application related information. Therefore, the translator may be able to efficiently or appropriately translate the invention document by referring to the patent application related information. In addition, when translators have defined corresponding translation words for important words, by creating an important word setting dictionary, invention documents provided by inventors or patent attorneys can be used in the invention document analysis system. When analyzed, appropriate patent application related information can be output.

Further, for example, the document acquisition unit acquires a first invention document and a plurality of invention documents other than the first invention document, and the document acquisition unit further acquires a first invention document and a plurality of invention documents other than the first invention document, and further, for each of the plurality of acquired invention documents, the document acquisition unit further acquires the first invention document and a plurality of invention documents other than the first invention document. Obtaining a translation result document that is a translation result of an invention document, the analysis unit extracts a group of important words that are a plurality of important words using the important word extraction unit for each of the first invention document and the plurality of invention documents. and a classification unit that performs group classification for specifying which of the plurality of groups each of the first invention document and the plurality of invention documents belongs to based on the correlation between each extracted important word group. , the translation word candidate set specifying unit extracts from the plurality of invention documents belonging to the group to which the first invention document is specified by the group classification of the classification unit and the corresponding translation result document. , a learning model for important phrase translation, which is a learning model generated by machine learning, using multiple sets of important phrases and corresponding translated words corresponding to the important phrases as training data for machine learning for important phrase translation. Elements of a set of translation word candidates for each of the important words extracted from the first invention document by the important word extraction unit may be specified by machine translation using a model.

Thereby, the invention document analysis system can appropriately identify a set of translation word candidates.

Further, for example, if the patent application related information is output as a display document, and the invention document acquired by the document acquisition unit includes a claim, the patent application related information is output as a display document. The claim information representing the claim content of the invention document acquired by the document acquisition unit is extracted from the invention document by the important phrase extraction unit. The additional information related to each of the plurality of important words and phrases in the claim may include display control information for controlling the display mode of the important word and phrase.

As a result, for example, a translator or the like may be able to identify important phrases whose display mode is controlled in claim information of patent application related information related to a display document displayed by a browser or the like.

Further, for example, for each of a plurality of important terms in the claim extracted by the important term extraction unit, the claim information representing the contents of the claim of the invention document acquired by the document acquisition unit The display control information as the related additional information further includes information added that displays the importance level, attribute information, corresponding translation words, a set of translation word candidates, or search results for the important word/phrase. It may include information for display and information for displaying the important word/phrase as a GUI element for executing the information addition display related to the important word/phrase in response to an operation. The search for important terms may be, for example, a search within the invention document, a search within another invention document, or a search on a search site on a network such as the Internet.

As a result, translators and others who view claim information in patent application related information related to display documents displayed on a browser etc. can check information such as importance, attribute information, search results, etc. by operating on the displayed important words. Since it may be possible to know the important words and phrases in the invention document, it may be possible to efficiently or appropriately understand the important words and phrases in the invention document. Therefore, the translator may be able to efficiently or appropriately determine translation terms for important terms.

Further, for example, if the display document is an HTML document including an HTML file written in HTML (Hypertext Markup Language), and the invention document acquired by the document acquisition unit includes a specification, , the patent application related information includes specification information representing the content of the specification to be displayed in the display document, and the specification information representing the content of the specification of the invention document acquired by the document acquisition unit. The information is the additional information related to each of the plurality of important words and phrases in the specification extracted from the invention document by the important word and phrase extraction unit, and the information is a display control that controls the display mode of the important word and phrase. The display control information for controlling the display mode of each of the plurality of important words extracted by the important word extraction unit in the claim information and the specification information, Includes highlighting information for differentiating the display mode of the important phrases corresponding to the control information from the display mode of words other than the plurality of important phrases in the claim information and the specification information, Each of the display control information for controlling a display mode corresponding to each of a plurality of important words and phrases having different character strings constituting the word/phrase is configured to display each of the plurality of important words and phrases in a mutually different display manner. The display control includes mutually different identification information for controlling, and controls the display mode of each of a plurality of important words and phrases extracted by the important word and phrase extraction unit in the specification information. If the key phrase is the same as any of a plurality of key phrases in the claim extracted by the key phrase extraction unit in the claim information, the key phrase is extracted from the same key phrase in the claim information. If the important phrase is displayed in the same display mode as the important phrase, and the important phrase is different from any of the plurality of important phrases in the claim extracted by the important phrase extracting section in the claim information, the important phrase is displayed. , the information may include information for controlling display in a display mode different from any of the plurality of important words/phrases in the claim information.

As a result, a translator or the like can view patent application related information as a display document, which is an HTML document, using a PC, smartphone, or other computer equipped with a browser compatible with HTML. Translators who view the patent application related information as a display document displayed by a browser etc. can check whether each important word or phrase described in the specification information in the patent application related information is an important word or phrase described in the claim. It becomes possible to quickly determine whether a plurality of important words and phrases are the same or different depending on the display mode of each important word and phrase. In addition, if the invention document analysis system controls the display mode of important words and phrases, for example, by highlighting them (for example, by displaying them in bold), translators, etc. can quickly grasp the important words and phrases in claim information. It can be.

Further, for example, the claim information representing the content of the claim of the invention document acquired by the document acquisition unit may include the important words and phrases in the content of the claim in addition to the content of the claim to be displayed in the display document. For each of the plurality of important words and phrases extracted by the extraction unit, the invention document includes content in which the part of the important word in the content of the claim is replaced with a corresponding translation word or a set of translation word candidates corresponding to the important word and phrase, and the invention document The specification information representing the contents of the specification includes, in addition to the contents of the specification to be displayed in the display document, each of a plurality of important words extracted by the important word extraction section from the contents of the specification. , the corresponding translation word or translation word candidate set corresponding to the important word/phrase may include content in which the important word/phrase in the content of the specification is replaced.

As a result, translators can efficiently translate invention documents by editing the parts of the content in which important words and phrases in claim information and specification information are replaced with corresponding translation words or a set of translation word candidates. It can be.

Further, for example, the document acquisition unit further acquires a drawing document corresponding to the invention document acquired by the document acquisition unit, and generates a drawing document based on an analysis result of the invention document acquired by the document acquisition unit. Among the plurality of important phrases extracted from the invention document by the important phrase extraction unit in the claim information or the specification information of the patent application related information, the element name described with a code added in the invention document. The display control information as the additional information related to the important word or phrase that is the same as the word representing the important word or phrase in the drawing document corresponding to the invention document It may also include information for displaying as a GUI element for displaying the part in which the code related to the phrase is written.

As a result, translators who have viewed patent application related information related to display documents displayed on a browser, etc. can manipulate the GUI elements of important terms displayed in claim information or specification information to edit the information in drawing documents. , it becomes possible to display and confirm the part corresponding to the element expressed by the important word or phrase, and it may become possible to understand the important word or phrase efficiently or appropriately.

Further, for example, when the output unit outputs the patent application related information by displaying the patent application related information, and the invention document acquired by the document acquisition unit includes a claim, The patent application related information includes claim information representing the content of the claim, and the patent application related information includes claim information representing the content of the claim of the invention document acquired by the document acquisition unit, and includes the relevant claim information extracted by the important phrase extraction unit. The additional information related to each of the plurality of important words in the claim indicates a color to be added to the important word and phrase, and the specification is included in the invention document acquired by the document acquisition unit. In this case, the patent application related information includes specification information representing the contents of the specification, and the patent application related information includes the important information that the specification information representing the contents of the specification of the invention document acquired by the document acquisition unit. The additional information related to each of the plurality of important words in the specification extracted by the word/phrase extraction unit indicates a color to be added to the important word/phrase, and the output unit extracts the claim information and Displaying the plurality of important words extracted by the important word extraction section in the specification information in a different manner from the display manner of words other than the plurality of important words; , displaying each of a plurality of important words extracted by the important word extraction unit and having different character strings constituting the important words by adding different colors to each other; and the output unit displaying the statement information Regarding each of the plurality of important words and phrases extracted by the important word and phrase extraction section in , if the important word and phrase are the same as any of the plurality of important words and phrases extracted by the important word and phrase extraction section in the claim information, The important words and phrases in the information are displayed with the same display color as the same important words and phrases in the claim information, and the important words and phrases in the specification information are extracted by the important words and phrases in the claim information. If the key word/phrase is different from any of the plurality of important words/phrases in the specification information, the output unit displays the key word/phrase in the specification information in a display mode that is different from any of the plurality of key words/phrases in the claim information, and the output unit For each of the plurality of important words and phrases extracted by the important word and phrase extracting unit in the information, the important word and phrase are displayed as a GUI element for executing information addition display related to the important word and phrase in accordance with the operation, and The additional information display may include displaying the importance level, attribute information, corresponding translation words, a set of translation word candidates, or search results for the important word/phrase. Adding color to important words includes, for example, adding font color to important words, adding text background color to important words, adding colored line width to all or one side of a rectangular frame surrounding important words. For example, adding a border line with .

As a result, translators, etc. who view patent application related information as a display document displayed on a browser etc. can check whether each important term indicated in the patent application related information is an important term stated in the claim. It becomes possible to quickly identify whether a plurality of important words and phrases are the same or different based on the color added to each important word and phrase.

Further, for example, the invention document analysis system includes a translation invention document standard provision acquisition unit that obtains translation invention document standard provisions that define a plurality of conditions that an invention document to be translated should satisfy; , determines whether the invention document acquired by the document acquisition unit satisfies at least some of the plurality of conditions defined in the standard regulations for invention documents for translation, and Based on the results, warning information is generated to point out potentially inappropriate parts in the invention document, and warning information related to the patent application generated based on the analysis results of the invention document by the analysis unit is generated. The information may include the warning information.

Thereby, by referring to the patent application related information outputted by the invention document analysis system, the inventor or patent attorney who created the invention document can modify the invention document if necessary. In addition, by receiving the invention documents to be translated and the patent application related information from the inventor or patent attorney, the translator may be able to receive information from the patent application related information even if the warning information in the patent application related information remains. It can be assumed that the inventor or patent attorney has determined that there is no need to take particular action on the indication, so it is possible to obtain an appropriate invention document to be translated, and as a result, the translation of the invention document can be performed efficiently or appropriately.

Further, for example, the warning information in the patent application related information generated based on the analysis result of the invention document may be invention document evaluation information indicating the conformity of the invention document to the invention document standard provisions for translation. It may also include. The suitability is expressed, for example, by an evaluation value as a result of suitability evaluation.

As a result, by referring to the invention document evaluation information included in the patent application related information outputted by the invention document analysis system, an inventor or a patent attorney may be able to appropriately modify the invention document. Furthermore, by checking the invention document evaluation information, the translator may be able to recognize whether the invention document is suitable for translation. Invention documents that contain inappropriate descriptions, for example, can easily lead to translation errors, and there may be situations where the translator has to modify the invention document before translating it in order to obtain appropriate rights (for example, to obtain patent rights). . Therefore, if the warning information in the patent application related information generated based on the invention document includes the invention document evaluation information indicating that the conformity with the invention document standard provisions for translation is relatively low, It can be determined that there is a good possibility that there is a discrepancy between the invention document and the translation result document that is the result of translation based on the invention document, in addition to the difference in language. It may be possible to determine that it is not useful to use the combination of the invention document and the translation result document as training data for machine learning in order to generate a learning model for translation to be used.

Furthermore, for example, the above-mentioned standard provisions for invention documents for translation represent character string patterns for potentially inappropriate descriptions that do not meet the conditions, as information that defines the conditions that an invention document to be translated should satisfy. includes inappropriate character string pattern information indicating an inappropriate character string pattern that has been written in correspondence with correction information regarding a method for correcting the written portion, and satisfies the conditions set forth in the standard provisions for invention documents for translation by the analysis department. The determination as to whether or not the above is made based on whether a character string corresponding to the inappropriate character string pattern indicated by the inappropriate character string pattern information exists, and is determined by the standard provisions for invention documents for translation. If a character string corresponding to the inappropriate character string pattern indicated by the inappropriate character string pattern information related to the condition exists in a claim in the invention document acquired by the document acquisition unit, the patent application related information includes modified claim information generated by modifying the claim to satisfy the condition based on the claim and the modification information associated with the inappropriate character string pattern information, and If the character string corresponding to the inappropriate character string pattern indicated by the inappropriate character string pattern information related to the conditions specified in the standard provisions for invention documents for translation is included in the specification of the invention document acquired by the document acquisition unit, If the patent application related information exists, the patent application related information corrects the specification to satisfy the condition based on the specification and the modification information associated with the inappropriate character string pattern information. It is also possible to include revised version specification information generated by this process.

As a result, inventors or patent attorneys who have created invention documents can efficiently refer to the revised claim information and revised specification information included in the patent application related information output by the invention document analysis system. It is possible to amend the invention document. In addition, the translator is provided with a revised invention document by referring to the revised claim information and the revised specification information, so that the translator can efficiently or appropriately translate the invention document. In addition, a translator who receives an invention document and patent application related information from an inventor, etc. or a patent attorney, and is further instructed to translate based on the revised claim information and revised specification information, It may be possible to perform translation work efficiently or appropriately by referring to the revised claim information and revised specification information included in the application-related information.

Furthermore, for example, when the invention document acquired by the document acquisition unit includes a claim, if a given clause of the claim cites multiple clauses, the analysis result of the invention document The patent application related information generated based on the above information includes the content in which the predetermined paragraph is changed to cite only the first paragraph among the plurality of paragraphs. The revised claim information representing the result of modifying the claim and the citation relationship between each clause of the claim represented by the revised claim information are correlated using lines or symbols between the item numbers of the terms that have a mutual citation relationship. It is also possible to include a modified version of the claim tree as the expressed information.

As a result, a translator who has been provided with patent application-related information by an inventor or a patent attorney, and who has been informed that the claim portion of an invention document should be translated based on the revised claim information, can It may be possible to translate efficiently without having to worry about whether or not to revise the section that cites it, and if so, how to revise it. In addition, a translator or the like can quickly grasp the citation relationship between each clause of a claim expressed in the revised claim information using the revised claim tree.

Further, for example, the patent application related information includes important word/phrase relationship information indicating a relationship between each of the plurality of important words extracted by the important word/phrase extraction unit from the invention document acquired by the document acquisition unit, The important term/phrase relationship information may include inclusion relationship information indicating a structural inclusion relationship between elements expressed by each of a plurality of important terms/phrases.

As a result, translators who have obtained the patent application related information output by the invention document analysis system can relatively quickly grasp the relationship between multiple important words in the invention document using the important word related information in the patent application related information. Therefore, by understanding the relationship, it is possible to efficiently or appropriately determine the translation words corresponding to each important word and phrase, and as a result, it may be possible to efficiently or appropriately translate the invention document.

Further, for example, the document acquisition section acquires a plurality of invention documents and a plurality of translation result documents that are translation results of each of the plurality of invention documents, and the analysis section acquires a plurality of translation result documents for each of the plurality of invention documents. , the important word extraction unit extracts a group of important words, which are a plurality of important words, and based on the correlation between the extracted groups of important words, to which group each of the plurality of invention documents belongs. For each of the plurality of groups, the invention document analysis system includes a classification unit that performs group classification to identify the invention documents classified into the group among the plurality of invention documents. Includes a machine learning unit that generates a learning model for translation, which is a learning model for machine translation, by machine learning using each set of the invention document and a translation result document that is a translation result of the invention document as training data. You can also use it as

With this, for example, in a machine translation system, an invention document analysis system, etc., an invention document can be machine translated by arithmetic processing using a translation learning model suitable for the group to which the invention document belongs among the translation learning models for each group. becomes possible.

Further, for example, the document acquisition unit acquires a plurality of invention documents and a plurality of translation result documents that are translation results of each of the plurality of invention documents, and the invention document analysis system analyzes the inventions to be translated. The invention document standard provision acquisition unit for translation acquires the invention document standard provision for translation that defines a plurality of conditions that the document should satisfy, and the invention document is A determination is made as to whether at least some of the plurality of conditions specified in the standard provisions for invention documents for translation are satisfied, and among the plurality of invention documents, the conditions for which the determination has been made are A document evaluation unit that selects invention documents whose satisfaction exceeds a predetermined degree, and a set of the invention documents selected by the document evaluation unit and the translation result document corresponding to the invention documents as training data. a machine learning unit that generates a translation learning model, which is a learning model for machine translation, by using machine learning, and the output unit outputs the translation learning model generated by the machine learning unit. It may also include a learning model output section.

As a result, a combination of an invention document containing inappropriate descriptions and a translation result document whose content is inconsistent with the invention document due to translation errors etc. generated based on the invention document is difficult to be used as training data. An appropriate learning model for translation can be output from the invention document analysis system. Therefore, for example, in a machine translation system or the like, it becomes possible to machine translate an invention document through arithmetic processing using an appropriate translation learning model output from an invention document analysis system. By referring to the results of machine translation of the invention document, the translator may be able to efficiently or appropriately translate the invention document.

Further, for example, the invention document analysis system includes a translation invention document standard provision acquisition unit that obtains a translation invention document standard provision that defines a plurality of conditions that an invention document to be translated should satisfy; Document standard provisions indicate inappropriate character string patterns that represent character string patterns for potentially inappropriate descriptions that do not meet the conditions, as information that determines the conditions that invention documents to be translated must satisfy. The invention document analysis system includes inappropriate character string pattern information and correction information regarding a correction method for the written portion in association with each other, and the invention document analysis system includes the invention document standard for translation in the invention document acquired by the document acquisition unit. If there is a character string that corresponds to the inappropriate character string pattern indicated by the inappropriate character string pattern information related to the conditions specified in the regulations, the content of the invention document and the inappropriate character string pattern information A set of a modified version of the invention document generated by modifying the contents of the invention document to satisfy the conditions based on the associated modification information and a translation result document corresponding to the invention document is created. The output unit includes a machine learning unit that generates a translation learning model, which is a learning model for machine translation, by machine learning used as training data, and the output unit generates the translation learning model generated by the machine learning unit. It is also possible to include a learning model output unit that outputs the learning model.

As a result, the revised invention document used as training data for machine learning and the translation result document may have less discrepancies in content, so the invention document analysis system A learning model may be output. Therefore, for example, in a machine translation system or the like, it becomes possible to machine translate an invention document through arithmetic processing using an appropriate translation learning model output from an invention document analysis system. By referring to the results of machine translation of the invention document, the translator may be able to efficiently or appropriately translate the invention document.

Further, for example, the analysis section includes a translation section that performs machine translation, and extracts a group of important words, which are a plurality of important words, from each of the plurality of invention documents in the same manner as the important word extraction section. Based on the correlation between the extracted important word groups, group classification is performed to identify which group each of the plurality of invention documents belongs to among the plurality of groups. For each invention document classified into the group among the plurality of invention documents, machine learning uses each set of the invention document and the translation result document that is the translation result of the invention document as training data. In a state where a learning model for translation, which is a learning model for machine translation, is being generated, the document acquisition unit acquires a predetermined invention document, and the translation unit extracts the important word/phrase from the predetermined invention document. According to the result of specifying which of the plurality of groups the predetermined invention document belongs to based on the key word group, which is a plurality of key words extracted by Machine translation of at least a portion of the predetermined invention document is performed using the corresponding translation learning model, and the output unit is generated based on an analysis result of the analysis unit for the predetermined invention document. Patent application related information including the result of the machine translation by the translation unit may be output.

As a result, a translator can efficiently or appropriately translate an invention document by referring to the result of machine translation of at least a part of a given invention document in the patent application related information outputted by the invention document analysis system. It may become possible.

(Modification group 2)
The invention document analysis system in the present disclosure is an invention document analysis system that analyzes an invention document, which is a document that expresses a technical idea in Japanese and includes a statement of claims for a patent application. a settings information acquisition unit that acquires configuration information including condition information indicating conditions; a document acquisition unit that acquires an invention document; and an analysis unit that analyzes the invention document acquired by the document acquisition unit based on the configuration information. and an output unit that outputs information generated based on the analysis result of the analysis unit regarding the invention document acquired by the document acquisition unit, and the analysis unit outputs information generated based on the analysis result of the invention document acquired by the document acquisition unit. The output unit includes an important word extraction unit that extracts important words from a document based on the setting information, and the output unit outputs each of the plurality of important words extracted by the important word extraction unit from the invention document acquired by the document acquisition unit. This is an invention document analysis system that outputs information generated based on.

Thereby, the invention document analysis system outputs information based on the result of analyzing an invention document including claims prepared by, for example, an inventor or a patent attorney. For example, by referring to the output information, an inventor or a patent attorney may be able to efficiently complete an invention document into a more appropriate one. An invention document that has been completed into a more appropriate one is useful for a translator who translates the invention document in order to translate the invention document appropriately or efficiently.

Further, for example, the condition information may be used for reference to important word candidate character string pattern information indicating an important word candidate character string pattern representing a pattern of character strings that are candidates for important word phrases, and previously described character strings. including reference character string pattern information indicating a reference character string pattern representing a pattern of a reference character string, and specific particle string pattern information indicating a specific particle string pattern representing a pattern of a character string that is a predetermined particle; The important word/phrase candidate character string pattern information is determined such that at least a predetermined character string whose first character is a kanji and a predetermined character string whose first character is a katakana correspond to the important word/phrase candidate character string pattern. The reference character string pattern information is defined such that at least the character string "above" corresponds to the reference character string pattern, and the specific particle character string pattern information is defined such that at least the character string "above" corresponds to the reference character string pattern. '', ``ni'', ``ga'', ``to'', ``kara'', ``he'', and ``wa'' are defined to correspond to the specific particle character string pattern, and the above-mentioned important The word/phrase extraction unit is configured to detect, in the invention document acquired by the document acquisition unit, a second character string that corresponds to the important word/phrase candidate character string pattern following the first character string that corresponds to the reference character string pattern. In this case, a character string that is the same as the second character string is extracted as an important word, and a third character string that corresponds to the important word candidate character string pattern is included in the claim of the invention document acquired by the document acquisition unit. When there is a subsequent period character, the same character string as the third character string is extracted as an important word and phrase, and the character corresponding to the important word candidate character string pattern is added to the invention document acquired by the document acquisition unit. If there is a specific particle string that is a character string that corresponds to the specific particle string pattern following an important word/phrase candidate character string that is a string, the same character string as the important word/phrase candidate character string is set as an important word. In the extraction of important words, when a plurality of character strings correspond to the important word candidate character string pattern, one character string among the plurality of character strings is is included in the description of the invention document, the remaining character strings excluding the one character string among the plurality of character strings are selected as the selection target, and the important words are selected and extracted, and the analysis section generates claim-based specification information representing the content of the incomplete specification based on each important term extracted by the important term extraction unit from the claims of the invention document acquired by the document acquisition unit; The output unit outputs patent application related information generated based on an analysis result by the analysis unit of the invention document acquired by the document acquisition unit, and the patent application related information is Each of the plurality of important words and phrases extracted by the important words and phrases extraction unit from the invention document is output as a display document that is included as content to be displayed, and the patent application related information is displayed in the display document. including claim information representing the content of the claim of the invention document acquired by the document acquisition unit and the claim base specification information, and representing the content of the claim of the invention document acquired by the document acquisition unit. The claim information includes display control information for controlling the display mode of each of the plurality of important words and phrases in the claim extracted from the invention document by the important word and phrase extraction unit, and , the display control information for controlling the display mode of each of the plurality of key words extracted by the key word/phrase extracting unit controls the display mode of the key word/phrase corresponding to the display control information, Contains highlighting information for differentiating the display mode of words other than the plurality of important words and phrases in the claim information, and corresponds to each of the plurality of important words and phrases in which the character strings constituting the important words and phrases are different from each other in the claim information. The display control information for controlling the display mode may include different identification information for controlling the plurality of important words to be displayed in different display modes.

As a result, inventors, etc. or patent attorneys who view patent application-related information can quickly identify important words and phrases with different character strings extracted from the claims of invention documents in the claim information, so they can easily identify the content of the claims. It may be possible to efficiently generate or update the specification of an invention document by referring to claim-based specification information that can be efficiently confirmed and generated based on important words extracted from the claims.

Further, for example, the analysis unit may cause the generated claim-based specification information to include a heading similar to the heading of the specification of a patent application in Japan, and the analysis unit may cause each clause in the claim of the invention document to include a heading similar to the heading of the specification of a patent application in Japan. Extract all the claimable originating terms from among them, identify the important words and phrases representing the claimed subject matter of each of the extracted claimable originating terms, and if the identified important terms include multiple same important terms, the relevant From a set of important words representing mutually different claim subjects that remains as a result of removing all but one important word from a plurality of important words, the elements of the set are listed in a predetermined order in a predetermined format. The arranged character string is determined as the invention name character string that follows the heading of "Title of the Invention" in the claim-based specification information, and the claimed starting claim is a claim that does not cite any other claim, or This is a claim that cites another claim but has a different claimed subject matter from the cited claim, and the analysis department has defined it based on important words expressing the claimed subject matter for each of the claim originating claims. Using the set of important words representing mutually different billing subjects, each important word in the set and the incomplete message for description support specified corresponding to each important word are connected in a predetermined format. one or more sentences are generated, and the generated sentence is determined as the written content that follows the heading of "Problem to be Solved by the Invention" in the claim-based specification information, and the incomplete message for writing support is , a character string sandwiched between an unfinished part start symbol and an unfinished part end symbol indicating that it is an unfinished stated part, and the analysis unit includes the item number of each of all the claimed starting points, and the Each important word in the set of important words representing mutually different claim objects determined based on the important words representing the claim object for each of all the claim object originating items and the description support specified corresponding to each important word. Generate one or more sentences by concatenating the unfinished messages for use in a predetermined format, and add the generated sentences to the heading of "Means for solving the problem" in the claim-based specification information. The analysis unit determines the content of the claim based on the item number of each claim of the invention document, the important phrase representing the claimed subject of the claim, and the incomplete message for description support. The analysis unit determines a part of the description following the heading “Detailed Description of the Invention” in the base specification information, and the analysis unit extracts the plurality of words extracted by the important phrase extraction unit from the claims of the invention document. Regarding each important term, the structural inclusion relationship between the element expressed by the important term and the element expressed by other important terms, the recorded content of the element expressed by the important term, or the realization content of the element expressed by the important term Based on the unfinished message for writing support to explain the relevant important phrases, which is determined based on The analysis unit determines, for each of the plurality of important words and phrases extracted by the important words and phrases extraction unit from the claims of the invention document, that the element represented by the important word and phrase structurally includes the element represented by other important words. or, when the important word/phrase is structurally included in the element represented by another important word/phrase, a code determination for determining the sign for the important word/phrase based on the structural inclusion relationship of the important word/phrase and the other important word/phrase. Among the important words and phrases included in at least a part of the content following the heading of "Detailed Description of the Invention" in the claim-based specification information, the code is determined for the important words and phrases in the code determination process. Regarding an important phrase for which a has been determined, the determined code may be inserted immediately after the important phrase.

As a result, the inventor, etc. or patent attorney can modify the incomplete message for description support in the claim-based specification information (for example, modify the part of the character string sandwiched between the incomplete part start symbol and the incomplete part end symbol, etc.) ), it is possible to efficiently and appropriately generate or update the specification of an invention document.

In addition, for example, in the code determination process that determines a sign for an important phrase, when determining a sign for the important phrase that represents an element that structurally includes an element represented by another important phrase, The code may be determined depending on the number of elements represented by other important words that the element represented by the word contains in its structure.

As a result, through the code determination process, for example, the code for an important word that represents an element that includes n elements (n is a natural number) represented by another important word is determined by k digits (k is a natural number) necessary to represent n elements. It may be possible to determine the number by adding zeros "0" equal to the number of digits of a natural number) to the end. For example, if n is 10 or more and 99 or less, k is 2, and the code for an important word that represents an element that includes n elements (n is a natural number) that other important words represent is a two-digit zero added to the end. For example, "500" or the like may be determined. For example, the code for each important phrase representing an element included in the element represented by the important phrase to which the code "500" is added is, for example, "501", "502", ..., "510", ..., "599". etc. may be determined. Therefore, the sign of each important word can be determined so that the inclusion relationship between the elements represented by each important word can be easily recognized at a glance.

Further, for example, the analysis unit adds display control information for controlling the display mode of the display element to the display element of the important phrase included in the claim-based specification information in the display document, and adds display control information for controlling the display mode of the display element. The control information is a graphical user interface (GUI) for displaying the display element of the important word/phrase according to the operation of the display element, and the display portion of the claim information in the display document that includes the important word/phrase. In the case where at least one incomplete description support message included in the claim-based specification information in the display document is a message listing a plurality of options, Display control information for controlling the display mode of the display element is added to the display element of the incomplete message for writing support, and the display control information adds the display element of the incomplete message for writing support to the display element of the incomplete message for writing support. Functions as a GUI for enabling a selection operation of the plurality of options of the incomplete message for writing support, and when the selection operation of any one of the plurality of options is performed, the display element is displayed. It may be for changing the content of the incomplete message for writing support to another predetermined incomplete message for writing support that is suitable for the option for which the selection operation has been performed.

As a result, an inventor or patent attorney who viewed the claim-based specification information in the patent application related information as a display document can select one of the multiple options included in the incomplete message for writing support in the claim-based specification information. , by selecting one option related to the item you want to include in the statement, the unfinished message for description support is changed, and as a result, another incomplete message for description support that is suitable for the selected option is displayed. You can view the completed message. Therefore, the inventor, etc. or patent attorney may, for example, make use of the claim-based specification information displayed after the selection operation (for example, copy and paste the claim-based specification information into a text editor, etc., and modify it as appropriate). ), it is possible to efficiently and appropriately generate or update the specification of an invention document.

Further, for example, the analysis unit generates claim-based drawing information based on the claims of the invention document acquired by the document acquisition unit, and the patent application related information The claim-based drawing information includes the drawing number of one or more drawings, important words and phrases to be written in each drawing, relationships between the important words, and codes corresponding to the important words. If the key word extraction unit extracts the key words from the claim of the invention document acquired by the document acquisition unit, the analysis unit includes information indicating the code. If there is an inclusion relationship, based on the inclusion relationship, enclose the important word representing the included element, and display the containing element outside the frame line connecting the code corresponding to the important word with a leader line. Arrange important words and phrases to surround the important words on the containing side and the frame line related to the included element, and connect the code determined in the code determination process to the important words on the containing side with a leader line. The analysis unit generates a configuration diagram depicting a frame line as a drawing in the claim-based drawing information, and the analysis unit adds the display element to the display element of the code included in the claim-based specification information in the display document. Display control information for controlling the display mode of the code is added, and the display control information changes the display element of the code to the code in the claim-based drawing information in the display document according to the operation of the display element. This is to function as a GUI for displaying the part that includes the same code, and the display control information related to the display element of the code included in the claim-based specification information is controlled according to the operation of the code. The function name or the function itself of a script that is activated with the code as an argument and identifies a portion of the claim-based drawing information that includes the code of the argument by searching for the code.
As a result, claim-based drawing information including drawings is generated based on the inclusion relationship between important words extracted from the claims of the invention document, so inventors, etc. or patent attorneys can refer to the claim-based drawing information. It is possible to generate appropriate drawings efficiently. In addition, by operating the GUI related to the code included in the claim-based specification information of patent application related information as a display document, the inventor, etc. or patent attorney can obtain claim-based drawing information containing the code. Since the internal parts can be quickly confirmed, it may be possible to efficiently generate or update the specification of the invention document appropriately by utilizing claim-based specification information and the like.

(Modification group 3)
The invention document analysis system in the present disclosure is an invention document analysis system that analyzes an invention document, which is a document that describes a claim or specification for a patent application, expressing a technical idea in Japanese. a settings information acquisition unit that acquires configuration information including condition information indicating extraction conditions; a document acquisition unit that acquires the invention document; and an analysis of the invention document acquired by the document acquisition unit based on the configuration information. an analysis unit; and an output unit that outputs information generated based on an analysis result of the analysis unit regarding the invention document acquired by the document acquisition unit, and the analysis unit includes an invention document acquired by the document acquisition unit. the output unit includes a plurality of important words and phrases extracted by the important words and phrases from the invention document acquired by the document acquisition unit; This is an invention document analysis system that outputs information generated based on each word.

As a result, the invention document analysis system outputs information based on the results of analyzing an invention document created by, for example, an inventor (for example, an inventor or a person related to the inventor) or a patent attorney, so that the invention document can be translated, for example. For a translator or the like (for example, a translator or other person using an invention document), referring to the output information may facilitate understanding of the invention document. Facilitating the understanding of invention documents will be useful for translators to appropriately or efficiently translate invention documents. Information generated based on key words in an invention document can be useful in cases where a translator determines corresponding translated words corresponding to key words in an invention document before translating the entire invention document.

Here, for example, the output unit outputs patent application related information generated based on the analysis result by the analysis unit of the invention document acquired by the document acquisition unit, and the patent application related information is For each of the plurality of important phrases extracted by the important phrase extraction section from the invention document acquired by the document acquisition section, the information includes information in which additional information related to the important phrase is added to the important phrase, and information related to the patent application is included. The additional information related to each of the plurality of important words and phrases extracted from the invention document acquired by the document acquisition unit in the information is a translated phrase that represents the meaning of the important word and phrase. The analysis unit includes translated word information as information regarding corresponding translated words, and the analysis unit calculates a corresponding translation of each important word extracted by the important word extraction unit from the invention document acquired by the document acquisition unit. a translation word candidate set specifying unit for specifying a translation word candidate set that is a set of word candidates; The translated word information as information regarding translated words includes translated word candidate information representing a set of translated word candidates for the important word or phrase specified by the translated word candidate set specifying unit, or includes the corresponding translated word. Good too. The translated word/phrase is, for example, a word/phrase expressed in one type of language other than Japanese (eg, English, Chinese, etc.).

As a result, translators can grasp multiple important words and phrases in invention documents by obtaining patent application related information output by the invention document analysis system, and translators can use additional information to translate invention documents. Therefore, it may be possible to efficiently or appropriately determine a translated phrase that corresponds to an important phrase and has the same meaning as the important phrase in the destination language. In addition, by referring to the translated word information included in the patent application related information output by the invention document analysis system, the translator may be able to efficiently or appropriately determine the translated word/phrase that corresponds to the important word/phrase. . In addition, by referring to the translation word candidate information or the corresponding translation words included in the patent application related information output by the invention document analysis system, the translator can, for example, obtain the translation word candidate set represented by the translation word candidate information corresponding to the important word. By selecting one translated word/phrase from the list, it may be possible to efficiently or appropriately determine the translated word/phrase corresponding to the important word/phrase.

For example, the patent application related information may include a display in which each of the plurality of important words extracted by the important word extraction unit from the invention document acquired by the document acquisition unit is included as content to be displayed. The additional information related to the important word/phrase extracted from the invention document is output as a business document, and the additional information is such that any element of the set of translation word candidates for the important word/phrase can be selected according to an operation. includes display control information for displaying the set of translation word candidates as a graphical user interface (GUI) element, and the display control information included in the additional information related to the important term is A script for executing a determination process that determines an element selected according to an operation from the set of translation word candidates for the important word/phrase displayed as an element as a corresponding translated word corresponding to the important word/phrase. The determining process of determining the element selected from the set of translation word candidates as a corresponding translation word includes a process of replacing the displayed set of translation word candidates with the corresponding translation word. It may also be a thing.

Thereby, the translator can easily determine the corresponding translated word corresponding to the important word/phrase from the set of translated word candidates by selecting the GUI element of the display document. Further, the translator can easily obtain the corresponding translated words displayed by the determination process from the display content of the display document using a browser or the like, for example, by copying and pasting, cutting and pasting, or the like. Copy and paste or cut and paste generally involves selecting data to be copied (that is, duplicating) or cutting (that is, cutting), specifying another location, and pasting that data to that location (that is, cutting). This operation causes the data to be copied or cut to be held in a predetermined memory area under the control of the operating system (OS), etc., and the paste operation causes the data to be copied or cut. The image is pasted from the specified memory area to the location specified as the paste destination.

Further, for example, the condition information may be used for reference to important term candidate character string pattern information indicating an important term candidate character string pattern representing a pattern of character strings that are candidates for important terms, and to previously described character strings. Reference string pattern information indicating a reference string pattern representing a reference string pattern, and a code indicating a string pattern representing a code character string pattern appended to a character string representing an element name. The key term candidate character string pattern information includes character string pattern information and specific particle string pattern information indicating a specific particle string pattern representing a pattern of character strings that are predetermined particles. A predetermined character string that is a kanji character and a predetermined character string whose first character is a katakana character are determined to correspond to the relevant important word candidate character string pattern, and the reference character string pattern information includes at least " The character string "aforesaid" is determined to correspond to the reference character string pattern, and the code string pattern information is such that at least a predetermined character string whose first character is a number corresponds to the reference character string pattern. The specific particle string pattern information is at least one of "wo", "ni", "ga", "to", "kara", "he" and "wa". is determined to correspond to the specific particle character string pattern, and the important word extraction section extracts the first character corresponding to the reference character string pattern from the invention document acquired by the document acquisition section. a first extraction unit that extracts a character string that is the same as the second character string as an important word when a second character string that corresponds to the important word candidate character string pattern exists subsequent to the string; and the document acquisition unit. In the case where the invention document obtained by the department contains a claim, if there is a period character following the third character string corresponding to the important word candidate character string pattern in the claim, the third character string a second extraction unit that extracts a character string that is the same as the character string as an important word; and a second extraction unit that extracts a character string that is the same as the character string as an important word; a third extraction unit that extracts a character string that is the same as the fourth character string as an important phrase when there is a fifth character string that corresponds to the coded character string pattern; , if there is a seventh character string that corresponds to the specific particle character string pattern following the sixth character string that corresponds to the important word candidate character string pattern, the same character string as the sixth character string is designated as important. and a fourth extraction unit that extracts the important word/phrase as a word, and in the extraction of the important word/phrase, the important word/phrase extraction unit extracts the plurality of character strings when a plurality of character strings correspond to the key word/phrase candidate character string pattern. If one of the character strings is included in another character string in the description of the invention document, the remaining character strings after excluding the one character string from among the plurality of character strings are selected as important words. The setting information includes information for specifying an important word/phrase setting dictionary that lists important words to be extracted, and in the important word/phrase setting dictionary, at least some of the listed important words are included. The invention document analysis system includes an important word/phrase setting dictionary acquisition unit that obtains the important word/phrase setting dictionary based on the setting information, and the important word/phrase extraction unit is configured to correspond to the important word/phrase setting dictionary. If the important words listed in the important word phrase setting dictionary are included in the invention document acquired by the document acquisition unit, the important words and phrases are extracted from the invention document, and the important word extraction unit extracts the important words from the invention document. When an important phrase that is associated with a corresponding translation word in the important phrase setting dictionary is extracted from the invention document acquired by the document acquisition unit, the relevant phrase extracted from the invention document in the patent application related information The additional information related to the important word/phrase represents the corresponding translated word, and the important word/phrase extraction section extracts the important word/phrase other than the important word/phrase to which the corresponding translated word is associated in the important word/phrase setting dictionary to the document acquisition section. When extracted from an invention document obtained by Contains information for displaying the set of translation word candidates as a GUI element so that any element of the set of translation word candidates for the important word/phrase identified by the identification unit can be selected in response to an operation; The display control information determines an element selected according to an operation from the translation word candidate set of the important word/phrase displayed as the GUI element as a corresponding translated word corresponding to the important word/phrase. It may also include a script for executing the determination process.

As a result, for example, when there is an expression that refers to a word in an invention document, the important word extraction unit treats the referenced word as an important word under certain conditions, and treats the referenced word as an important word under certain conditions. Since words and phrases are treated as important words under certain conditions, there is a possibility that words and phrases that are noted enough to be referenced and words that represent the subject matter of a claim will be treated as important words. It is assumed that there is a strong tendency for a period character to follow the claimed subject matter in the claims of ordinary invention documents. Therefore, since the invention document analysis system can output information by the output unit based on important words and phrases that are estimated to be important in the invention document, the output information can be useful. In addition, the important word extraction unit may, for example, treat words that form part of the description of other important words as not to be extracted as important words, and may keep certain words (for example, noun phrases, etc.) that precede a predetermined particle in the invention document. Under certain conditions, words and phrases that are marked and written in invention documents are treated as important words and phrases, so relatively important words and phrases such as subjects and objects and in drawings related to invention documents are treated as important words and phrases under certain conditions. Words that represent elements that are presumed to be illustrated can be treated as important words. Since the invention document is written in Japanese, there is a good chance that words starting with kanji or katakana are important words. Therefore, since the invention document analysis system can output information through the output unit based on important words and phrases that are estimated to be important in the invention document, the output information can be useful. In addition, this allows, for example, an inventor or a patent attorney who has created an invention document to specify a specific word or phrase as an important term, or to specify a corresponding translation for an important term. By generating a word/phrase setting dictionary, the invention document analysis system can output appropriate patent application related information. Therefore, the translator may be able to efficiently or appropriately translate the invention document by referring to the patent application related information. In addition, when translators have defined corresponding translations for important words, by creating an important word setting dictionary, invention documents provided by inventors or patent attorneys can be used in the invention document analysis system. When analyzed, appropriate patent application related information can be output.

Further, for example, when the important word extraction unit extracts important words other than the important words with which the corresponding translation words are associated in the important word phrase setting dictionary from the invention document acquired by the document acquisition unit, the The display control information included in the additional information related to the important word/phrase extracted from the invention document in the patent application related information is the display control information included in the additional information related to the important word/phrase extracted from the invention document. It is also possible to include a script for executing an update process for updating the important word/phrase setting dictionary so as to include information associating words and corresponding translated words.

As a result, the contents of the important word/phrase setting dictionary can be updated to contents that can be more appropriately utilized in the subsequent analysis of the invention document by performing a selection operation on the set of translation word candidates.

For example, the patent application related information may include, for each important phrase extracted by the important phrase extraction section from the invention document acquired by the document acquisition section, the important phrase and the corresponding translation word for the important phrase; Or, it includes an important word list configured by associating a set of translation word candidates for the important word/phrase, and the important word/phrase in the important word list in the patent application related information as the display document, the corresponding translation word. , and display control information for controlling the display mode of the display element, which is the additional information related to the important term, may be added to each display element of the translation word candidate set.

As a result, the translator can quickly grasp the important words and phrases included in the invention document by referring to the important word list, and can perform efficient translation. In addition, inventors, patent attorneys, or translators can efficiently edit the list of important words in patent application-related information that is output as a result of having the invention document analysis system analyze an invention document. , it may be possible to generate and update an important word/phrase setting dictionary.

Further, for example, if the invention document acquired by the document acquisition unit includes a claim, the patent application related information may include claim information representing the content of the claim to be displayed in the display document. , and claim translation information representing the result of machine translation for the claim expressed in the claim information, and represents the result of machine translation in the claim translation information that targets one line or paragraph of the claim information. A line or paragraph is a complete translation line that expresses the entire line or section of the claim information in the target language, or a corresponding translation of an important word or phrase included in the line or section of the claim information that corresponds to the important word or phrase. or a partial translation line that is the result of substitution in the translation word candidate set for the important word or phrase, and if the specification is included in the invention document acquired by the document acquisition unit, the patent The application-related information includes specification information representing the contents of the specification to be displayed in the display document, and specification translation information representing the result of machine translation of the specification represented in the specification information. In the specification translation information, a line representing the result of machine translation for one line of the specification information is a complete translation line expressing the entire line of the specification information in the target language, or , is a partial translation line that is the result of replacing an important term included in the specification information line with a corresponding translation word corresponding to the important term or a set of translation word candidates for the important term, and the document The claim information representing the content of the claim of the invention document acquired by the acquisition unit includes information on the display mode of each of the plurality of important words in the claim extracted from the invention document by the important word extraction unit. The specification information, which includes display control information for control and represents the contents of the specification of the invention document acquired by the document acquisition unit, is information in the specification extracted from the invention document by the important word extraction unit. includes display control information for controlling the display mode of the important words and phrases for each of the plurality of important words and phrases, and for each of the plurality of important words and phrases extracted by the important word and phrase extraction unit in the claim information and the specification information. The display control information for controlling the display mode of the important words/phrases is configured to control the display mode of the important words/phrases corresponding to the display control information in a manner other than the plurality of important words/phrases in the claim information and the specification information. The display control includes highlighting information for differentiating the display mode of words and phrases, and controls the display mode corresponding to each of a plurality of important words and phrases in the claim information in which the character strings forming the important words are different from each other. Each of the pieces of information includes different identification information for controlling each of the plurality of important words and phrases to be displayed in different display modes, and the plurality of pieces of information extracted by the important word and phrase extraction unit in the statement information The display control information for controlling the display mode of each of the important words and phrases includes the display control information for controlling the display mode of the important words and phrases, such that the important words and phrases are the plurality of important words and phrases in which the character strings constituting the important words and phrases are different from each other in the claim information. If the key words and phrases are the same as any of the key words and phrases in the claim information, the key words and phrases are displayed in the same display mode as the same key words and phrases in the claim information, and the key words and phrases in the claim information are displayed so that the character strings forming the key words in the claim information The claim translation information includes information for controlling the important word to be displayed in a display mode different from any of the plurality of important words in the claim information if the important word is different from any of the plurality of important words. and the specification translation information, for important words and phrases that are the same as the plurality of important words and phrases in which the character strings constituting the important words in the claim information are different from each other, the result of machine translation using the important words as the subject of machine translation. As information to be added to the display element of the machine translation, the machine translation is performed so that at least a part of the display form of the result of the machine translation is the same as at least a part of the display form of the important words and phrases in the claim information and the specification information. It may also include display control information for controlling the display mode of the translation results.

As a result, for example, a translator etc. may be able to identify important words and phrases whose display mode is controlled in claim information and specification information of patent application related information related to a display document displayed by a browser etc. , it may be possible to identify translation results of important phrases whose display modes are controlled in claim translation information and specification translation information, and quickly grasp which translation result corresponds to which important phrase. In addition, translators and others who view patent application-related information will be able to identify whether each important word or phrase described in the specification information in the patent application-related information is an important word or phrase described in the claim. , it may be possible to quickly understand whether a plurality of important words and phrases are the same or different depending on the display mode of each important word and phrase. In addition, if the invention document analysis system controls the display mode of important words, for example, by highlighting them (for example, displaying them in bold), translators, etc. can quickly grasp important words and phrases in claim information. It can be.

Further, for example, the display document is an HTML document that includes the patent application related information as content to be displayed and includes an HTML file written in HTML (Hypertext Markup Language), and is acquired by the document acquisition unit. The display control information related to each of the plurality of important words and phrases in the claim extracted by the important word and phrase extracting unit, which is included in the claim information representing the content of the claim of the invention document obtained by A plurality of important phrases in the specification extracted by the important phrase extraction unit and included in the specification information representing the contents of the specification of the invention document acquired by the document acquisition unit. The display control information related to each important phrase indicates a display color to be added to the important phrase, and the display control information for each important phrase indicates a display color to be added to the important phrase, and the display control information for each of the plural important phrases in the complaint information whose character strings forming the important phrases are different from each other. The display control information is for controlling each of the plurality of important words and phrases to be displayed in a display mode in which a different display color is added for each important word, and is for controlling the display of each of the plurality of important words and phrases in a display mode in which a different display color is added for each important word, and is used to control the display of each of the plurality of important words and phrases in a display mode that adds a different display color to each important word. Display control information relating to an important phrase in the claim information, which is composed of the same character string as any one of the plurality of important phrases whose character strings are different from each other, This is to control the display in a display mode in which the same display color as one important term is added, and for each of the plurality of important terms in the specification information, the important term is If the key word is the same as any one of the plurality of key words, the display control information related to the key word in the specification information makes the key word match the one key word in the claim information. This is to control the display in a display mode with the same display color as the display color added as the display mode, and the claim translation information and the specification translation information are used to display important words and phrases in the claim information. For important words and phrases that are the same as each of the plurality of important words and phrases whose constituent character strings are different from each other, the information added to the display element of the machine translation result that targeted the important word and phrase for machine translation is added to the display element of the machine translation result. Controlling the display mode of the result of the machine translation so that the display color of at least a part of the display mode is the same as the display color of at least a part of the display mode of the important words and phrases in the claim information and the specification information. It may also include display control information for.

As a result, a translator or the like can view patent application related information as a display document, which is an HTML document, using a PC, smartphone, or other computer equipped with a browser compatible with HTML. Translators, etc. who view patent application related information as a display document displayed by a browser etc., identify important words and phrases in the patent application related information and display colors added to the translation results of important words. By doing so, it becomes possible to identify whether a certain important phrase is an important phrase stated in a claim, it is possible to quickly understand whether multiple important phrases are the same or different, and it is possible to quickly determine whether a certain translation result is the same or different. You can quickly understand which important words correspond to the corresponding words.

Further, for example, the display control information relating to each of a plurality of important words and phrases in which the character strings constituting the important words are different from each other in the claim information may be such that each of the plurality of important words is displayed in a different display color for each important word. This is to control the display in a display mode in which a left border line having a predetermined width is added, and for each of the plurality of important words and phrases in the specification information, the important word and phrase are the same as those in the claim information. If the key word is the same as any one of the key words in the specification information, the display control information regarding the key word in the specification information may change the display of the key word to the display of the one key word in the claim information. control to display in a display mode in which a left border line having the predetermined width is added in the same display color as the display color of the left border line according to the claim information, and the important words in the specification information are If the key word/phrase is different from any of the plurality of important words/phrases in the specification information, the display control information regarding the important word/phrase in the specification information is configured to display the important word/phrase in a display mode without the left border line. The display control information relating to the important word/phrase in the claim information is for controlling the important word/phrase to be displayed in a display mode with a left border line having the predetermined width added. The display control information relating to the results of machine translation for the important words and phrases in the claim translation information and the specification translation information, which target important words and phrases in the It may be for controlling the display in a display mode in which a left border line having the predetermined width is added in the same display color as the display color of the left border line related to the display mode of the important word/phrase.

As a result, translators who view patent application related information as a display document displayed on a browser etc. can add different important words included in claim information to each important word in the patent application related information. By identifying the display color of the left border line, you can quickly distinguish between the two. In addition, there is a method of expressing different important words by using different display colors of the left border added to the important words. This has the advantage that the legibility of the important words themselves is not reduced regardless of whether the background color of the area is dark (for example, black) or light (for example, white). In addition, the left border that can be added to some display elements in HTML documents is not a character, so it will be ignored when copying and pasting the text of the display content that includes important words, so the display content It does not get in the way when copying and pasting into a text editor etc. to edit descriptions that include important words. In addition, translators who view patent application-related information can identify certain claims by identifying important words and phrases in the patent application-related information and the display color of the left border added to the translation result of the important words. It becomes possible to quickly understand which important words and phrases stated in the claim correspond to the translation results of the important words and phrases in the claim.

For example, the claim translation information includes each line of the claim information, and for each line or item of the included claim information, the content of the line or item is immediately followed by the content of the line or item. A complete translation line or a partial translation line representing the result of machine translation corresponding to the machine translation target is inserted, and the specification translation information includes each line of the specification information, and the specification translation information includes each line of the specification information. Immediately after each line of document information, a complete translation line or a partial translation line is inserted that represents the result of machine translation corresponding to the machine translation target, which is the line, and the claim translation information and the The specification translation information includes the display mode of display elements of at least a part of the machine translation target in the claim translation information and the specification translation information, and the display element of at least a part of the machine translation result corresponding to the machine translation target. Display control information may be added to each display element in order to control the display mode to be different at least in part.

As a result, translators who view patent application related information as a display document displayed on a browser, etc., can check the lines to be machine translated and the lines resulting from machine translation in the claim translation information and specification translation information. Becomes quickly identifiable.

In addition, the translator etc. visually confirms the machine translation target line in the claim translation information or specification translation information, and also checks the contents of the machine translation result line located after the machine translation target line. It may be possible to efficiently create a more appropriate translated text by making appropriate modifications.

For example, the claim translation information and the specification translation information may include the display mode of display elements of at least a portion of the machine translation target in the claim translation information and the specification translation information, and the display mode of display elements of at least a part of the machine translation target in the claim translation information and the specification translation information, To control the display mode of display elements of at least a portion of the translation results to be different in at least one of text background color, text color, text decoration, font size, font style, and font weight. , display control information may be added to each display element.

As a result, translators who view patent application related information as a display document displayed on a browser, etc., can check the lines to be machine translated and the lines resulting from machine translation in the claim translation information and specification translation information. , character background color, character color, character decoration, font size, font style, and font weight, and can be quickly identified.

For example, a display element for the result of machine translation that targets important words in the claim translation information and the specification translation information may include a machine translation result for the important words and phrases in accordance with the operation of the display element. Display control information including translation word candidate information is added in order to give the display element the function of a GUI that displays a set of translation word candidates different from the result, and the display control information is attached to the result of the operation. When a selection operation is performed on one element from the translation word candidate set on the GUI that displays the translation word candidate set displayed as , the content of the display element as a result of machine translation is selected by the selection operation. It may be determined that the one element that has been changed is changed.

As a result, translators who view patent application related information as a display document displayed on a browser etc. can select machine translation results for important terms in claim translation information and specification translation information as necessary. It can be easily changed by operation.

Further, for example, the document acquisition unit acquires a plurality of invention documents, and the corresponding translated words in the partial translation line of the claim translation information and the specification translation information are learning models for machine translation of important words. The invention document analysis system is generated as a result of machine translation using a learning model for important word translation, and the invention document analysis system is a translator for obtaining translation invention document standard provisions that define multiple conditions that an invention document to be translated must satisfy. The invention document standard provision acquisition unit and each of the plurality of invention documents acquired by the document acquisition unit, the invention document meets at least some of the plurality of conditions defined in the invention document standard provision for translation. a document evaluation unit that makes a determination as to whether or not each condition is satisfied, and selects, from among the plurality of invention documents, an invention document that satisfies the condition for which the determination was made to a degree that exceeds a predetermined degree; , a machine learning unit, the machine learning unit generates a display document by the invention document analysis system based on the invention document selected by the document evaluation unit, and generates a set of translation word candidates for important words. As a result of displaying the GUI, a set of the important word/phrase and one element selected from the translation candidate set is accumulated as training data for machine learning for generating the learning model for translation of the important word/phrase. The method may include a function of generating the learning model for machine translation of important words and phrases by machine learning using a plurality of previously accumulated training data.

As a result, invention documents with relatively high conformity to the standard provisions for invention documents for translation are selected, and important words and phrases in the invention documents and elements selected from the corresponding translation candidate set are used as a learning model for important word translation. Since inappropriate invention documents that are difficult to translate are excluded from the generation materials of the learning model for important word/phrase translation, the learning model for important word/phrase translation can be appropriately generated. Further, by using the generated learning model for important word/phrase translation later to analyze the invention document, more appropriate machine translation of the important words/phrases in the invention document can be realized. By referring to more appropriate machine translation results, the translator may be able to efficiently or appropriately determine the translated words and phrases that have been translated from the important words and phrases in the invention document.

Further, for example, the complete translation line in the claim translation information and the specification translation information is generated as a result of machine translation using a translation learning model that is a learning model for machine translation, and the document acquisition unit , acquires a plurality of invention documents and a plurality of translation result documents that are translation results of each of the plurality of invention documents, and the invention document analysis system defines a plurality of conditions that the invention document to be translated should satisfy. an invention document standard provision acquisition unit for translation that acquires the invention document standard provision for translation; and for each of the plurality of invention documents acquired by the document acquisition unit, the invention document is defined in the invention document standard provision for translation. A determination is made as to whether at least some of the plurality of conditions are satisfied, and the degree to which the conditions for which the determination was made is satisfied among the plurality of invention documents is a predetermined degree. The machine learning section includes a document evaluation section that selects an invention document that exceeds the limit, and a machine learning section. The method may include a function of generating the translation learning model by machine learning using a plurality of pairs of translation result documents as training data.

As a result, invention documents are selected based on the standard provisions for invention documents for translation, so that invention documents containing inappropriate descriptions and translation errors generated based on the invention documents may be affected by the invention documents and their content. Pairs with inconsistent translation result documents are less likely to be used as training data, and the invention document analysis system can generate an appropriate translation learning model. Further, in the invention document analysis system, it becomes possible to machine translate the invention document through arithmetic processing using the appropriate translation learning model. By referring to the complete translation lines in the claim translation information and specification translation information as a result of machine translation of this invention document and making appropriate corrections, the translator can efficiently or appropriately translate the invention document. It can be.

Further, for example, the analysis unit generates warning information under certain conditions based on the invention document acquired by the document acquisition unit, the patent application related information includes the warning information, and the warning information includes: Inspection information for inspecting whether the invention document acquired by the document acquisition unit satisfies a plurality of conditions specified in the standard provisions for invention documents for translation and pointing out if the conditions are not satisfied; A predetermined calculation based on the result of an inspection as to whether the invention document acquired by the document acquisition unit satisfies at least some of the plurality of conditions specified in the standard regulations for invention documents for translation. Invention document evaluation information indicating an evaluation value of the conformity of the invention document to the invention document standard provisions for translation, which is calculated by the above method, may also be included.

Thereby, by referring to the patent application related information outputted by the invention document analysis system, the inventor or patent attorney who created the invention document can modify the invention document if necessary. In addition, by receiving the invention documents to be translated and the patent application related information from the inventor or patent attorney, the translator can receive information from the patent application related information even if the warning information in the patent application related information remains. It can be assumed that the inventor, etc. or patent attorney has determined that there is no need to take particular action on the indication, so it is possible to obtain an appropriate invention document to be translated, and as a result, the translation of the invention document can be performed efficiently or appropriately.

Further, by referring to the invention document evaluation information included in the patent application related information outputted by the invention document analysis system, the inventor or the patent attorney may be able to appropriately modify the invention document. Furthermore, by checking the invention document evaluation information, the translator may be able to recognize whether the invention document is suitable for translation. Invention documents that contain inappropriate descriptions, for example, can easily lead to translation errors, and there may be situations where the translator has to modify the invention document before translating it in order to obtain appropriate rights (for example, to obtain patent rights). . Therefore, if the warning information in the patent application related information generated based on the invention document includes the invention document evaluation information indicating that the conformity with the invention document standard provisions for translation is relatively low, It can be determined that there is a good possibility that there is a discrepancy between the invention document and the translation result document that is the result of translation based on the invention document, in addition to the difference in language. It may be possible to determine that it is not useful to use the combination of the invention document and the translation result document as training data for machine learning in order to generate a learning model for translation to be used.

For example, the standard provisions for invention documents for translation include version information indicating the version of the standard provisions for invention documents for translation, and whether the invention document conforms to each of the plurality of conditions specified in the standard provisions for invention documents for translation. and an inspection program for a process of determining whether or not the invention document standard provisions for translation exist, and the said invention document standard provision acquisition unit accesses the said invention document standard provisions for translation as a resource on the network and acquires the invention document standard provisions for translation as a resource on the network. If it is possible to obtain a newer version of the invention document standard provisions for translation than the version indicated by the version information of the invention document standard provisions for translation, the new version of the invention document standard provisions for translation is obtained, and the analysis unit whether the invention document acquired by the document acquisition unit by executing the inspection program included in the standard provisions for invention documents for translation conforms to each of the plurality of conditions specified in the standard provisions for invention documents for translation. It is also possible to generate the warning information by determining whether or not the warning information is present, and using the result of the determination.

Standard provisions for invention documents for translation are provisions established by a specific organization or person, such as an organization for the purpose of standardizing invention documents for translation, and are provided on a network such as the Internet (for example, publicly available) In the case where the invention document analysis system According to the above, when necessary, a new version of the invention document standards for translation is obtained from the network and used for analysis of the invention document (i.e., generation of warning information, etc.). generation, etc.).

Further, for example, the complete translation line in the claim translation information and the specification translation information is generated as a result of machine translation using a translation learning model that is a learning model for machine translation, and the invention document analysis system includes an acquisition unit for obtaining standard provisions for invention documents for translation that defines a plurality of conditions that an invention document to be translated should satisfy, and the standard provisions for invention documents for translation are to be translated. As information defining the conditions that an invention document should satisfy, inappropriate character string pattern information indicating an inappropriate character string pattern representing a character string pattern for a potentially inappropriate written part that does not satisfy the conditions, and the relevant The invention document analysis system includes a machine learning unit, and the machine learning unit includes the invention document for translation in the invention document acquired by the document acquisition unit. If there is a character string that corresponds to the inappropriate character string pattern indicated by the inappropriate character string pattern information related to the conditions specified in the invention document standard regulations, the content of the invention document and the inappropriate character string. A modified invention document generated by modifying the content of the invention document to satisfy the condition based on the modification information associated with the pattern information, and a translation result document corresponding to the invention document. It is also possible to include a function of generating the translation learning model, which is a learning model for machine translation, by machine learning using the set of as training data.

As a result, the revised invention document used as training data for machine learning and the translation result document may have less discrepancies in content, so the invention document analysis system can provide appropriate translation A learning model may be generated. Further, in the invention document analysis system, it becomes possible to machine translate the invention document through arithmetic processing using the appropriate translation learning model. By referring to the complete translation lines in the claim translation information and specification translation information as a result of machine translation of this invention document and making appropriate corrections, the translator can efficiently or appropriately translate the invention document. It can be.

Further, for example, the invention document analysis system may group invention documents based on the correlation between important word groups extracted by the important word extraction unit from each of a plurality of invention documents, or based on a user's designation. After performing classification, machine learning is performed using multiple sets of important words extracted from invention documents and their corresponding translated words as training data for each group. A learning model for phrase translation is generated, and the corresponding translated words in the partial translation line of the claim translation information and the specification translation information generated based on the one invention document acquired by the document acquisition unit are It may be generated as a result of machine translation using a learning model for important word/phrase translation of a group to which the invention document belongs. This group classification is based on the technical field to which the invention related to the invention belongs, since it can be assumed that the correlation between important word groups extracted from each of multiple invention documents to which the invention belongs to the same technical field is higher than a certain level. It can be a classification based on the user's designation, and it can be a classification in which IPC sections in the case of IPC classification, or sections and classes that are different are classified into different groups, based on the user's designation. Based on the classification, the department to which the inventor of an invention document belongs is divided into different groups to distinguish it from other departments. Although this is possible, it is assumed that the system will be operated in such a way that useful group classifications are made for inventors, patent attorneys, or translators.

As a result, it may be possible to machine translate each important term in an invention document through arithmetic processing using a learning model for important term translation that is suitable for the group to which the invention document belongs. Further, by referring to the partial translation lines of the claim translation information and the specification translation information, the translator may be able to efficiently or appropriately translate important words and phrases of the invention document.

Further, for example, the invention document analysis system may analyze a plurality of invention documents based on a correlation between a group of important words extracted from each of the plurality of invention documents by the important word extraction section, or based on a specification by a user. For each group, a learning model for translation, which is a learning model for machine translation, is generated by machine learning using multiple sets of invention documents and corresponding translation result documents as training data. The complete translation line in the claim translation information and the specification translation information generated based on the one invention document acquired by the document acquisition unit is obtained by using the translation learning model of the group to which the one invention document belongs. It may be generated as a result of machine translation.

As a result, it may be possible to machine translate each important phrase in an invention document through arithmetic processing using a translation learning model that is suitable for the group to which the invention document belongs. Further, by referring to complete translation lines of claim translation information and specification translation information, a translator may be able to efficiently or appropriately translate an invention document.

Further, for example, the group classification of invention documents is performed based on the correlation between important word groups extracted from each of the plurality of invention documents by the important word extraction unit, and The correlation between important word groups is determined based on the degree of importance of each important word in the important word group extracted from each of the plurality of invention documents, which is determined based on the attribute information of the important word. The attribute information of the important words and phrases calculated and extracted from the invention document after performing corresponding weighting is information that identifies whether or not the important words are used in the claims of the invention document, and the information that identifies whether or not the important words are used in the claims of the invention document, Information identifying whether or not a code corresponding to the important word/phrase exists in the document, information identifying whether the key word/phrase is used as a subject in the invention document, and information identifying whether the key word/phrase is used as a subject in the invention document. It is also possible to include information for identifying whether or not the technical term has a predetermined ending character string unique to the technical term.

As a result, the degree of importance of each important phrase is reflected in the correlation between the important phrase groups, so that group classification can become more appropriate for generating an important phrase translation learning model or a translation learning model. By referring to the results of the machine translation of the invention document by the invention document analysis system using the learning model for important word translation or the learning model for translation generated as a result of the group classification, the translator can improve the invention document. It may be possible to translate efficiently.

Furthermore, for example, the degree of importance determined based on the attribute information for each important word/phrase extracted from the invention document is based on the structural inclusion relationship between each element represented by each important word/phrase in the invention document. Therefore, the importance of the important word related to the containing element is greater than the importance of each important word related to all included elements included in the element expressed by the important word related to the containing element. It may be used to calculate the correlation between important word groups after being updated to a value that is equal to or higher than the highest importance value.

As a result, among the important words and phrases related to the included element group between elements that have a structural inclusion relationship, the importance of the important word whose importance has been determined to be the highest based on the attribute information is finally is propagated to the importance of the important words related to the included element, so the result of group classification based on the correlation between the above important word groups, which is calculated based on the compositional inclusion relationship related to the important words, is , it may be more suitable for generating a learning model for translation of important words or a learning model for translation.

Further, for example, the display control information related to an important phrase in the display document related to the patent application related information generated for the invention document acquired by the document acquisition unit may change the display element of the important phrase to the display A process of displaying the attribute information for the important word/phrase according to the operation of the element, and displaying the important word/phrase from the claim information, the specification information based on the invention document, or an invention document different from the invention document. An important word inclusion description enumeration process that enumerates and displays lines containing important words, and a search site on the network that displays web pages that contain the important words, definitions of the important words, synonyms of the important words, or synonyms of the important words. It functions as a GUI that executes the process of searching for related images or the process of translating the relevant important words on a translation site on the network. The lines containing the above are listed and displayed from the claim information, the specification information, or the other invention document, and the important words and phrases in each of the listed and displayed lines are highlighted, and the displayed lines are , display control information for causing the line to function as a GUI for displaying links to multiple lines including the line in the claim information, the specification information, or the other invention document, in response to an operation on the line. It may also be an additional process.

As a result, translators who have viewed patent application-related information related to a display document displayed on a browser, etc., can perform operations on important words in the displayed patent application-related information to obtain attribute information of the important words and phrases. It may be possible to obtain information such as search results (for example, web pages containing important words, definitions of important words, synonyms, related images), translation results of important words, etc., so you can efficiently analyze important words in invention documents. Or it may be possible to understand properly. Therefore, the translator may be able to efficiently or appropriately determine translation terms for important terms.

In addition, the translator or the like can accurately identify the important word/phrase because the important word/phrase is highlighted in the line enumerated by the important word/phrase inclusion description enumeration process by including the manipulated important word/phrase. In addition, by further operating on one of the lines enumerated by the important term inclusion description enumeration process by including the operated important term, the translator etc. Multiple lines in claim information, specification information, or another invention document that precede and follow one line can be easily confirmed.

Further, for example, the display control information included in the additional information related to the important words extracted from the invention document may be translated into a corresponding translation as a result of the execution of the determination process by executing the script for executing the determination process. It includes a script for executing processing for generating and outputting translation auxiliary information, which is information associating the important word and the corresponding translated word, for each important word for which the word has been determined, and the document acquisition unit When the translated word information included in the additional information related to one important phrase of the acquired invention document is the translated word candidate set, the elements of the translated word candidate set are In the past, for each of one or more invention documents, the display document related to the patent application related information generated based on the analysis result of the invention document by the analysis unit is output to the output unit for display. The information is identified based on the translation auxiliary information generated for each important term in each of the one or more invention documents as a result of being used in the invention document acquired by the document acquisition unit. If the translated word information included in the additional information related to one important word is a corresponding translated word, the corresponding translated word is determined by using the translation auxiliary information for the one important word. It may also be a thing.

As a result, the result of the selection operation by the translator on the GUI element related to the set of translation word candidates of the display document is then used by the invention document analysis system to determine the corresponding translated word corresponding to the important word/phrase via the translation auxiliary information. There is a possibility of improving the validity of identifying a set of translation word candidates for important words or phrases.

Further, for example, if the translated word information included in the additional information related to one important term of the invention document acquired by the document acquiring unit is the translated word candidate set, the translated word candidate set is The elements include important words and phrases in each of the plurality of translation auxiliary information generated as a result of displaying each of the display documents output based on each of a plurality of invention documents in the past by the invention document analysis system. , the results of machine translation using a learning model for important word translation, which is a learning model generated by machine learning, using the corresponding pairs of translated words as training data for machine learning for important word translation. , and if the translated word information included in the additional information related to one important term of the invention document acquired by the document acquisition unit is a corresponding translated word, the corresponding translated word The word may be determined as a result of machine translation of the one important word/phrase using the important word/phrase translation learning model generated based on the translation auxiliary information.

As a result, the learning model for important word translation, which is generated based on the translation auxiliary information generated as a result of the translator's selection operation on the GUI element related to the set of translation word candidates of the display document, is later used in the invention document analysis system. By using the invention document analysis system, the validity of determining corresponding translation words corresponding to important words or phrases or specifying a set of translation word candidates for important words can be improved.

Further, for example, the document acquisition unit acquires a plurality of invention documents, and the output unit displays each of the display documents output for each of the plurality of invention documents acquired by the document acquisition unit. A set of important words and corresponding translated words in each of the plurality of translation auxiliary information generated as a result is used as training data for machine learning for important word phrase translation, and the information is generated by machine learning. It is also possible to include a learning model output unit that outputs a learning model for important word/phrase translation, which is a learning model.

As a result, in a machine translation system for machine translation of invention documents, the learning model for important word translation can be used to perform arithmetic processing using important words as input. Translation can be accomplished.

By referring to the results of this machine translation, the translator may be able to efficiently or appropriately determine the translated words and phrases that are the important words and phrases in the invention document.

Further, for example, in the display document related to the patent application related information generated for the invention document acquired by the document acquisition unit, the display related to the important words extracted from the invention document by the important word extraction unit The control information includes a process of displaying the attribute information of the key word/phrase in response to the operation of the display element of the key word/phrase, and a process of displaying the attribute information of the key word/phrase based on the claim information and the specification information based on the invention document. Enumeration process for listing and displaying lines containing important words, and web pages that include the important words on search sites on the network, definitions of the important words, synonyms of the important words, or relationships of the important words. It functions as a GUI that executes the process of searching for an image or the process of translating the important words on a translation site on the network, and the attribute information of the important words extracted from the invention document is If the important word or phrase is used in a claim of the document, there is information indicating the section number of the claim section in which the important word or phrase is stated, and there is also a statement in which a code is added to the important word or phrase in the invention document. In this case, information indicating the code may also be included.

As a result, a translator, etc. who viewed patent application related information related to a display document displayed on a browser, etc., can perform operations on important words in the displayed patent application related information to determine the importance of the important words as attribute information. Information such as the item number of the claim in which the words are stated, the code corresponding to the important words, search results for the important words (e.g., web pages containing the important words, definitions of the important words, synonyms, related images), translation results of the important words, etc. Therefore, it may be possible to efficiently or appropriately understand important words and phrases in the invention document. Therefore, the translator may be able to efficiently or appropriately determine translation words for important words.

Further, the invention document analysis processing program in the present disclosure causes a device equipped with a microprocessor to analyze an invention document, which is a document that describes claims or specifications for a patent application expressing a technical idea in Japanese. An invention document analysis processing program for performing an invention document analysis process, wherein the invention document analysis process includes a setting information acquisition step of acquiring setting information including condition information indicating conditions for extracting important words; a document acquisition step of acquiring a document; an analysis step of analyzing the invention document acquired by the document acquisition step based on the setting information; and an analysis by the analysis step of the invention document acquired by the document acquisition step. an output step of outputting information generated based on the result, and the analysis step includes an important word extraction step of extracting important words from the invention document acquired in the document acquisition step based on the setting information. , the output step is an invention document analysis processing program that outputs information generated based on each of the plurality of important phrases extracted in the important phrase extraction step from the invention document acquired in the document acquisition step.
By installing this invention document analysis processing program into a device etc. equipped with a microprocessor, the device etc. can realize invention document analysis processing. By referring to the information outputted by the invention document analysis process realized by this invention document analysis process program, translators and the like may be able to facilitate their understanding of the invention document. Facilitating the understanding of invention documents will be useful for translators to appropriately or efficiently translate invention documents.

Furthermore, the display document in the present disclosure is data that is output by the above-described invention document analysis system, may include a script as a computer program, and includes a document written in a markup language. The display document in the present disclosure includes a document written in a markup language, is subject to interpretation by a browser in a device equipped with a microprocessor, and displays the above-mentioned patent application related information generated by the above-mentioned invention document analysis system. , include it as the content that should be displayed.
By having a browser in a device equipped with a microprocessor interpret this display document, patent application related information is displayed. By referring to this patent application related information, translators and the like may be able to facilitate their understanding of the invention document.

The invention document analysis system and invention document analysis processing program according to the present disclosure can be used to analyze invention documents, and the display document according to the present disclosure can be used to promote understanding of invention documents.

10, 20 Invention document analysis system 10a, 10c, 10d, 10e, 10g, 10h Terminal device 10b, 10f, 10i Server device 101, 201 Setting information acquisition unit 102, 202 Document acquisition unit 103, 203 Important word setting dictionary acquisition unit 104 , 204 Invention document standard provision acquisition unit for translation 105, 205 Analysis unit 106, 2071 Document evaluation unit 107, 207 Machine learning unit 108, 208 Output unit 1051, 2051 Important phrase extraction unit 10511 First extraction unit 10512 Second extraction unit 10513 Third extraction unit 10514 Fourth extraction unit 1052, 20541 Translation word candidate set identification unit 1053, 2053 Classification unit 10531, 20531 Technical field identification unit 1054, 2054 Translation unit 1081, 2081 Learning model output unit 2021 Invention document acquisition unit 2022 Drawing document Acquisition unit 2023 Translation result document acquisition unit 2055 Importance determination unit 2056 Attribute identification unit 2057 Claim citation relationship detection unit 2058 Inter-element relationship determination unit 2059 Warning information generation unit 205a Claim information generation unit 205b Specification information generation unit 205c Claim tree generation unit 2072 Teacher data storage unit 2073 Learning model management unit 2082 Patent application related information output unit 2083 Display document generation unit 2084 Display unit 20841 Operation reception unit

Claims (35)

An invention document analysis system that analyzes an invention document, which is a document that describes technical ideas in Japanese and describes claims and specifications for patent applications,
a settings information acquisition unit that acquires settings information including condition information indicating conditions for extracting important words;
a document acquisition unit that acquires invention documents;
an analysis unit that analyzes the invention document acquired by the document acquisition unit based on the setting information;
an output unit that outputs information generated based on an analysis result by the analysis unit of the invention document acquired by the document acquisition unit,
The analysis unit includes an important word extraction unit that extracts important words from the invention document acquired by the document acquisition unit based on the setting information,
The output unit outputs information generated based on each of the plurality of important phrases extracted by the important phrase extraction unit from the invention document acquired by the document acquisition unit ,
The output unit outputs patent application related information generated based on the analysis result by the analysis unit of the invention document acquired by the document acquisition unit,
The patent application related information includes, as contents to be displayed, each of a plurality of important words extracted by the important word extraction unit from the invention document acquired by the document acquisition unit,
The patent application related information includes claim information representing the content of the claim of the invention document acquired by the document acquisition unit to be displayed,
The claim information representing the content of the claim of the invention document acquired by the document acquisition unit includes a display of the important phrase for each of a plurality of important phrases in the claim extracted from the invention document by the important phrase extraction unit. including display control information for controlling aspects;
The patent application related information includes specification information representing the content of the specification of the invention document acquired by the document acquisition unit, which is to be displayed;
The specification information representing the content of the specification of the invention document acquired by the document acquisition unit includes information on each of the plurality of important terms and phrases in the specification extracted from the invention document by the important term extraction unit. Contains display control information for controlling the display mode of words and phrases,
The display control information for controlling the display mode of each of the plurality of important words extracted by the important word extraction unit in the claim information and the specification information corresponds to the display control information. including highlighting information for differentiating the display mode of the important words and phrases from the display mode of words other than the plurality of important words and phrases in the claim information and the specification information,
Among the important words and phrases expressed in the claim information, each of the display control information for controlling the display mode corresponds to a plurality of important words and phrases in which the character strings constituting the important words are different from each other. including mutually different identification information for controlling each word to be displayed in a mutually different display manner;
Among the important words and phrases represented in the specification information, the display mode of an important word that is the same as any one of the plurality of important words and phrases in which the character strings forming the important words in the claim information are different from each other. The display control information for controlling the display element for the important word/phrase is controlled to be displayed in a display manner that is at least partially similar to the display manner of the display element for the one important word/phrase in the claim information. Contains information for
Invention document analysis system. An invention document analysis system that analyzes an invention document, which is a document that describes technical ideas in Japanese and describes claims and specifications for patent applications,
a settings information acquisition unit that acquires settings information including condition information indicating conditions for extracting important words;
a document acquisition unit that acquires invention documents;
an analysis unit that analyzes the invention document acquired by the document acquisition unit based on the setting information;
an output unit that outputs information generated based on an analysis result by the analysis unit of the invention document acquired by the document acquisition unit,
The analysis unit includes an important word extraction unit that extracts important words from the invention document acquired by the document acquisition unit based on the setting information,
The output unit outputs information generated based on each of the plurality of important phrases extracted by the important phrase extraction unit from the invention document acquired by the document acquisition unit,
The output unit outputs patent application related information generated based on the analysis result by the analysis unit of the invention document acquired by the document acquisition unit,
The patent application related information is output as a display document in which each of the plurality of important terms extracted by the important term extraction unit from the invention document acquired by the document acquisition unit is included as content to be displayed. ,
The patent application related information includes claim information representing the content of the claim of the invention document acquired by the document acquisition unit, which should be displayed in the display document,
The claim information representing the content of the claim of the invention document acquired by the document acquisition unit includes a display of the important phrase for each of a plurality of important phrases in the claim extracted from the invention document by the important phrase extraction unit. including display control information for controlling aspects;
The patent application related information includes specification information representing the content of the specification of the invention document acquired by the document acquisition unit, which should be displayed in the display document,
The specification information representing the content of the specification of the invention document acquired by the document acquisition unit includes information on each of the plurality of important terms and phrases in the specification extracted from the invention document by the important term extraction unit. Contains display control information for controlling the display mode of words and phrases,
The display control information for controlling the display mode of each of the plurality of important words extracted by the important word extraction unit in the claim information and the specification information corresponds to the display control information. including highlighting information for differentiating the display mode of the important words and phrases from the display mode of words other than the plurality of important words and phrases in the claim information and the specification information,
Among the important words and phrases expressed in the claim information, each of the display control information for controlling the display mode corresponds to a plurality of important words and phrases in which the character strings constituting the important words are different from each other. including mutually different identification information for controlling each word to be displayed in a mutually different display manner;
Among the important words and phrases represented in the specification information, the display mode of an important word that is the same as any one of the plurality of important words and phrases in which the character strings forming the important words in the claim information are different from each other. The display control information for controlling the display element for the important word/phrase is controlled to be displayed in a display manner that is at least partially similar to the display manner of the display element for the one important word/phrase in the claim information. Contains information for
Invention document analysis system. The display related to each of the plurality of important words and phrases in the claim extracted by the important word and phrase extraction unit, which is included in the claim information representing the content of the claim of the invention document acquired by the document acquisition unit. The control information includes information indicating a display color to be added to the important word/phrase,
For each of a plurality of important terms and phrases in the specification extracted by the important term extraction unit, the specification information representing the contents of the specification of the invention document acquired by the document acquisition unit includes: The display control information includes information indicating a display color to be added to the important word/phrase,
Among the important words and phrases expressed in the claim information, the display control information relating to each of the plurality of important words and phrases whose character strings are different from each other controls the display elements of each of the plurality of important words and phrases for each important word. including information for controlling display in a display mode with a different display color added in response to the identification information of
The display for controlling the display mode of an important phrase that is different from any of the plurality of important phrases in which the character strings constituting the important phrase in the claim information are different from each other among the important phrases represented in the specification information. The control information includes information for controlling the display element of the important word/phrase to be displayed in a display mode different from any of the display elements of the plurality of important words/phrases in the claim information,
Among the important words and phrases expressed in the specification information, the character strings constituting the important words in the claim information are the same as any one of the plurality of important words and phrases that are different from each other. The display control information controls the display element of the important term to be displayed in a display mode with the same display color as the display color added as the display mode of the display element of the one important term in the complaint information. 3. The invention document analysis system according to claim 2, wherein the invention document analysis system includes information for. The display document is an HTML document that includes the patent application related information as content to be displayed and includes an HTML file written in HTML (Hypertext Markup Language),
Display control related to important phrases in the claim information, in which the character strings constituting the important phrases in the claim information are composed of the same character string as any one important phrase among the plurality of mutually different important phrases. The information is for controlling the important word/phrase to be displayed in a display mode with the same display color as the display color added as the display mode of the one important word/phrase,
The display control information relating to each of a plurality of important words and phrases in which the character strings forming the important words are different from each other in the claim information has a predetermined width in which each of the plurality of important words is displayed in a different color for each important word. This is to control display in a display mode with a left border added,
For each of the plurality of important words in the specification information, if the important word is the same as any one of the plurality of important words in the claim information, then the important word in the description information The display control information relates to the important word/phrase, and a left border line having the predetermined width is added to the important word/phrase in the same display color as the left border line according to the display mode of the one important word/phrase in the claim information. If the important word/phrase in the description information is different from any of the plurality of important words/phrases in the claim information, the important word/phrase in the description information is 4. The invention document analysis system according to claim 3, wherein the display control information is for controlling the important word/phrase to be displayed in a display mode in which the left border line having the predetermined width is not added. The predetermined width is the width of one character,
In the claim information, for each of the plurality of important phrases in which the character strings constituting the important phrases are different from each other, the display control information related to the important phrase may This is to control the display in a display mode without lines.
The display color that differs for each important word is a display color that has a different hue component for each important word, or a display color that has different components of RGB values representing a display color for each important word.
The invention document analysis system according to claim 4 . The patent application related information includes, for each of the plurality of important words extracted from the invention document by the important word extraction section, at least a part of the attribute information of the important word or phrase, or at least one of the attribute information of the important word or phrase. Contains information based on
The invention document analysis system according to any one of claims 3 to 5. The analysis unit determines the relationship between elements expressed by each important phrase extracted by the important phrase extraction unit from the claims described in the invention document acquired by the document acquisition unit, and determines the relationship between elements expressed by each important phrase extracted by the important phrase extraction unit, and including an inter-element relationship determination unit that generates word relationship information,
The patent application related information includes the important word/phrase related information,
The inter-element relationship determining unit includes, for each of the plurality of important terms extracted by the important term extracting unit from the claims described in the invention document , part of the attribute information of the important term in the important term relationship information. 7. The invention document analysis system according to claim 6 , wherein inclusion relationship information expressing a structural inclusion relationship between the element represented by the important word and the element represented by other important words is included. The setting information acquired by the setting information acquisition unit includes comment specifying information that is information specifying a comment line or comment,
The analysis unit includes a claim tree generation unit that generates a claim tree expressing in a tree structure a citation relationship of each term of the claim expressed in the claim information in the patent application related information,
The patent application related information includes the claim tree,
The claim tree generation unit includes:
For each section of the claim in the claim information, a new line is written in descending order of the section number, and information indicating the section number of the section that the section cites is written using a symbol for expressing the citation relationship and the section number. generate a tree,
For each item related to each line of the claim tree,
If the line immediately preceding the relevant section in the claim information is a comment line in a predetermined format specified based on the comment identification information, the contents of the comment line are converted to the item number of the relevant section in the claim tree. Add it to the line you want to write,
In the case where the line immediately preceding the relevant section in the claim information is not the comment line, and a comment is written in the same line as the heading of the relevant section in a predetermined format specified based on the comment specifying information, Add the content of the comment to the line where the item number of the section in the claim tree is written,
In the case where the line immediately preceding the relevant section in the claim information is not the comment line, and the comment is not written on the same line as the heading of the relevant section, the important information stated in the relevant section of the claim of the invention document When there are one or more important words extracted by the word extraction unit, at least one of the one or more important words is added to the line in which the item number of the item in the claim tree is written. ,
The claim tree generation unit generates a constituent claim tree that expresses the citation relationship of each term of the claim expressed in the claim information in the patent application related information in a tree structure format,
The patent application related information includes the constituent claim tree,
The claim tree generation unit rearranges the lines of each claim in the claim information in descending order of item number, and adds information indicating the item number of the item cited by the item to a symbol for expressing a citation relationship. A constituent claim tree is generated, which is expressed by the section number and item number, and for each term related to each row of the constituent claim tree, one or more important terms written in the relevant term in the claim information and extracted by the important term extraction unit are generated. If it exists, each important term is added to the line in which the item number of the relevant section is written, and if there are multiple important words to be added to the relevant section and the elements represented by each of the multiple important words are structurally If there is an inclusion relationship, add a statement expressing the relationship between important words to the line.
The invention document analysis system according to claim 7 . The patent application related information includes claim-based specification information representing the content of the specification of the invention document acquired by the document acquisition unit, which should be displayed in the display document,
The claim-based specification information includes each of a plurality of important words extracted by the important word extraction unit from the claims described in the invention document as part of the description following the heading of the description of the invention. As part of the attribute information of the important term, inclusion is permitted based on inclusion relationship information that expresses the structural inclusion relationship between the element expressed by the important term and the element expressed by other important terms. Contains descriptions of character strings with unfinished messages to support descriptions, which represent but do not necessarily indicate inclusion relationships, and can be modified and used as appropriate.
The invention document analysis system according to any one of claims 6 to 8. The invention document analysis system is configured to analyze important words and phrases extracted by the important words and phrases extracting unit from the claims written in the invention document acquired by the document acquiring unit, which are important words whose codes do not correspond in the invention document. For words and phrases, the key word and phrase are determined based on inclusion relationship information that expresses the structural inclusion relationship between the element expressed by the key term and the element expressed by other key terms, as part of the attribute information of the key term. determine the sign as part of the attribute information of
The patent application related information includes, for each of a plurality of important terms extracted by the important term extracting unit from the claims described in the invention document, a code as part of the attribute information of the important term. containing a code list represented in association,
The patent application related information includes claim-based drawing information,
The claim-based drawing information is a configuration diagram based on a code as part of the attribute information of each of the plurality of important words extracted by the important word extraction unit from the claims described in the invention document. including
The invention document analysis system according to any one of claims 6 to 9. The patent application related information includes an important word list in which a new line is listed for each important word extracted from the invention document by the important word extraction unit , and at least a part of the attribute information of the important word ;
Each line of the important word list is arranged in a predetermined order, and the order is determined based on at least part of the attribute information for the important words in each line, and the character code for the important words in each line. Claim 6: The order is according to the sorting result in the following order, the order of the number of uses of important words and phrases in each line in the invention document, or the order of the number of uses of important words and phrases in each line in the claims of the invention document.Claim 6 10. The invention document analysis system according to any one of items 10 to 10 . The display control information relating to each of a plurality of important words and phrases in which the character strings forming the important words are different from each other in the claim information has a predetermined width in which each of the plurality of important words is displayed in a different color for each important word. This is to control display in a display mode with a left border added,
For each important word in the important word list, if the important word is the same as any one of the plurality of important words in the claim information, the important word in the important word list includes: The important word/phrase is displayed in a display mode in which a left border line having the predetermined width is added in the same display color as the left border line according to the display mode of the one important word/phrase in the complaint information. The invention document analysis system according to claim 11 , wherein display control information for controlling is added. The analysis unit generates translation word candidates that are a set of corresponding translation word candidates for the one or more important words extracted by the important word extraction unit from the invention document acquired by the document acquisition unit. including a translation word candidate set identification unit for identifying the set;
The invention document analysis system according to claim 11 or 12, wherein each line of the important word list includes one important word and a corresponding translation word or a set of translation word candidates corresponding to the important word. In the display document related to the patent application related information generated for the invention document acquired by the document acquisition unit, the display control information related to important words extracted from the invention document by the important phrase extraction unit, The display element of the important word/phrase functions as a graphical user interface for displaying at least part of the attribute information of the important word/phrase according to the operation.
The invention document analysis system according to any one of claims 6 to 13 . In the display document related to the patent application related information generated for the invention document acquired by the document acquisition unit, the display control information related to important words extracted from the invention document by the important phrase extraction unit, To display the item number of the claim section that defines the important term and the item number of the claim section that refers to the important term, in response to a mouse-over operation of the display element of the important term. It functions as a graphical user interface for
The invention document analysis system according to claim 14. In the display document related to the patent application related information generated for the invention document acquired by the document acquisition unit, the display control information related to important words extracted from the invention document by the important phrase extraction unit, Depending on the operation of the display element of the key term, search for web pages containing the key term, definition of the key term, synonyms of the key term, or related images of the key term on a search site on the network. The invention document analysis system according to any one of claims 6 to 15, wherein the invention document analysis system functions as a graphical user interface that executes processing to translate the important words or phrases at a translation site on a network. In the display document related to the patent application related information generated for the invention document acquired by the document acquisition unit, the display control information related to important words extracted from the invention document by the important phrase extraction unit, Functions as a graphical user interface for executing an important word inclusion description enumeration process that enumerates and displays lines including the important word from the claim information and the specification information according to the operation of the display element of the important word. It is something that makes
The invention document analysis system according to any one of claims 6 to 16 . The invention document analysis system includes, for important words and phrases whose codes correspond to the important words extracted from the invention document by the important word extraction unit, the code and the important words as part of the attribute information of the important words . Generate a code list that is a list representing words and phrases in relation to each other and include it in the patent application related information,
The invention document analysis system includes:
Listing the codes written subsequent to the important words extracted by the important words extracting unit from the invention document, and representing the codes and the corresponding important words in association with each other;
If the description in which a code is added to the key word extracted by the key word extraction unit does not exist in the invention document, the element represented by the key word is constitutive of the element represented by the other key word or phrase. or when it is structurally included in the element represented by another important word, the code is determined based on the structural inclusion relationship between the important word and the other important word, and the determined code and the relevant By expressing it in association with important words,
The invention document analysis system according to any one of claims 6 to 17, wherein the code list is generated. The patent application related information includes a specification heading list in which lines including headings representing paragraph numbers in the specification of the invention document acquired by the document acquisition unit are listed in the order of description in the specification,
In the specification heading list, for a line that includes a heading representing a paragraph number in the specification of the invention document, the paragraph range corresponding to the paragraph number in the specification includes the paragraph from the claims of the invention document. If there is an important phrase that is the same as the important phrase extracted by the important phrase extraction unit, add the important phrase to the relevant line of the specification heading list.
The invention document analysis system according to any one of claims 3 to 18 . The analysis unit generates warning information based on the invention document acquired by the document acquisition unit, including one or more indication information representing indications regarding inappropriate descriptions included in the invention document. including the
The patent application related information includes the warning information,
The warning information generation unit includes:
If it is detected that the invention document acquired by the document acquisition unit includes a description that does not satisfy the condition that the same element name is written immediately before the same code, the same including in the warning information indication information indicating a code and element names of a plurality of different expressions to which the code is added ;
If a word or phrase is written following a reference character string in any section of a claim of an invention document acquired by the document acquisition unit, the referenced phrase that is the same as the word or phrase is the reference character in that section. Contains descriptions of reference character strings that do not meet the conditions that they must be stated in the part that precedes the column, or that the relevant paragraph must be stated in the cited section based on the citation chain. If it is detected that the preceding referenced word/phrase does not exist, the item number of the section that includes the description related to the detected reference character string, and the description related to the reference character string. The warning information includes indication information indicating the contents of the warning information.
The invention document analysis system according to any one of claims 3 to 19 . The warning information generation unit displays the claim information or the specification corresponding to the indication of the indication information in at least one display element of the indication information in the warning information included in the patent application related information as the display document. 21. The invention document analysis system according to claim 20 , further comprising: adding display control information for causing the display element to function as a graphical user interface for displaying a link to a description location in the document information. The display control information relating to each of a plurality of important words and phrases in which the character strings forming the important words are different from each other in the claim information has a predetermined width in which each of the plurality of important words is displayed in a different color for each important word. This is to control display in a display mode with a left border added,
The patent application related information includes each line of the claim information, and for each line or item of the included claim information, immediately after the content of the line or item, Contains claim translation information in which a translation line containing the result of replacing the included important words with corresponding translated words corresponding to the important words is inserted,
The patent application related information includes each line of the specification information, and for each line of the included specification information, immediately after the content of the one line, the important words included in the one line correspond to the important words. Contains the specification translation information in which the translation line containing the result of replacing with the corresponding translated word is inserted,
The claim translation information and the specification translation information are configured to provide information on corresponding translated words corresponding to the same important words as each of the plurality of important words and phrases in which the character strings constituting the important words in the claim information are different from each other. Contains display control information for controlling the display mode of the corresponding translated word,
The display control information related to the corresponding translated word corresponding to the important word in the claim translation information and the specification translation information displays the corresponding translated word in the same color as the display color added to the important word and phrase in the predetermined width. The invention document analysis system according to any one of claims 3 to 21, wherein the invention document analysis system is for controlling display in a display mode in which a left border line having a left boundary line is added. An invention document analysis system that analyzes an invention document, which is a document that describes technical ideas in Japanese and describes claims and specifications for patent applications,
a settings information acquisition unit that acquires settings information including condition information indicating conditions for extracting important words;
a document acquisition unit that acquires invention documents;
an analysis unit that analyzes the invention document acquired by the document acquisition unit based on the setting information;
an output unit that outputs information generated based on an analysis result by the analysis unit of the invention document acquired by the document acquisition unit,
The analysis unit includes an important word extraction unit that extracts important words from the invention document acquired by the document acquisition unit based on the setting information,
The output unit outputs information generated based on each of the plurality of important phrases extracted by the important phrase extraction unit from the invention document acquired by the document acquisition unit,
The condition information is
Important term candidate character string pattern information indicating an important term candidate character string pattern representing a pattern of character strings that are candidate important terms;
Reference character string pattern information indicating a reference character string pattern representing a reference character string pattern used for referencing the previously described character string;
Code string pattern information indicating a code string pattern representing a pattern of a code string appended to a string representing an element name;
and specific particle string pattern information indicating a specific particle string pattern representing a pattern of strings that are predetermined particles;
The important word/phrase candidate character string pattern information is determined such that at least a predetermined character string whose first character is a kanji and a predetermined character string whose first character is a katakana correspond to the important word/phrase candidate character string pattern. has been
The reference character string pattern information is defined such that at least the character string "said" corresponds to the reference character string pattern,
The code string pattern information is determined such that at least a predetermined character string whose first character is a number corresponds to the code string pattern,
The specific particle string pattern information includes at least one of the characters "wo", "ni", "ga", "to", "kara", "he", and "wa" in the specific particle string. It is determined that it corresponds to the pattern,
The important word extraction unit is
If the invention document acquired by the document acquisition unit includes a second character string that corresponds to the important word candidate character string pattern following the first character string that corresponds to the reference character string pattern, a first extraction unit that extracts a character string that is the same as the two character strings as an important phrase;
In the claim of the invention document acquired by the document acquisition unit, when there is a period character following the third character string that corresponds to the important word candidate character string pattern, a character string that is the same as the third character string; a second extraction unit that extracts as important words;
In the invention document acquired by the document acquisition unit, when there is a fifth character string corresponding to the code character string pattern following the fourth character string corresponding to the important word candidate character string pattern, the fifth character string corresponding to the code character string pattern exists. a third extraction unit that extracts a character string that is the same as the four character strings as an important phrase;
In the invention document acquired by the document acquisition unit, when there is a seventh character string that corresponds to the specific particle character string pattern following the sixth character string that corresponds to the important word/phrase candidate character string pattern, a fourth extraction unit that extracts the same character string as the sixth character string as an important phrase;
In the extraction of important words, when a plurality of character strings correspond to the important word/phrase candidate character string pattern, one of the plurality of character strings is selected from other character strings. is included in the description of the invention document, the remaining character strings excluding the one character string among the plurality of character strings are selected as the selection target, and important words are selected and extracted.
Invention document analysis system. An invention document analysis system that analyzes an invention document, which is a document that describes technical ideas in Japanese and describes claims and specifications for patent applications,
a settings information acquisition unit that acquires settings information including condition information indicating conditions for extracting important words;
a document acquisition unit that acquires invention documents;
an analysis unit that analyzes the invention document acquired by the document acquisition unit based on the setting information;
an output unit that outputs information generated based on an analysis result by the analysis unit of the invention document acquired by the document acquisition unit,
The analysis unit includes an important word extraction unit that extracts important words from the invention document acquired by the document acquisition unit based on the setting information,
The output unit outputs information generated based on each of the plurality of important phrases extracted by the important phrase extraction unit from the invention document acquired by the document acquisition unit,
The document acquisition unit acquires a plurality of invention documents and a plurality of translation result documents that are translation results of each of the plurality of invention documents,
The invention document analysis system includes:
an acquisition unit for obtaining standard provisions for invention documents for translation, which obtains standard provisions for invention documents for translation that define a plurality of conditions that an invention document to be translated should satisfy;
Regarding each of the plurality of invention documents acquired by the document acquisition unit, as to whether the invention document satisfies at least some of the conditions defined in the standard provisions for invention documents for translation. a document evaluation unit that makes a determination and selects an invention document from among the plurality of invention documents that satisfies the condition for which the determination was made to a degree that exceeds a predetermined degree;
A machine that generates a learning model for translation, which is a learning model for machine translation, by machine learning using a set of an invention document selected by the document evaluation unit and a translation result document corresponding to the invention document as training data. including the learning department,
The output unit includes a learning model output unit that outputs the translation learning model generated by the machine learning unit.
Invention document analysis system. A device equipped with a microprocessor is caused to perform an invention document analysis process that analyzes an invention document, which is a document containing claims and specifications for a patent application expressing a technical idea in Japanese, and generates a display document. An invention document analysis method, comprising:
The invention document analysis process includes:
a document obtaining step of obtaining an invention document;
an analysis step of analyzing the invention document acquired in the document acquisition step by extracting important words based on predetermined conditions for extracting important words;
an output step of generating and outputting the display document for displaying the patent application related information generated based on the analysis result of the analysis step for the invention document acquired by the document acquisition step;
The display document includes a document written in a markup language, is subject to interpretation by a browser in a device equipped with a microprocessor, and is the content of a claim in an invention document for a patent application that expresses the technical idea in Japanese. Patent application related information including claim information representing the content of the invention document and specification information representing the content of the specification in the invention document is included as content to be displayed,
The claim information representing the content of the claim includes the display mode of the display element of each important word extracted based on the predetermined conditions for extraction of important words among the words of the claim. Contains display control information for controlling the
Among the important words and phrases expressed in the claim information, each of the display control information relating to each of a plurality of important words and phrases in which the character strings constituting the important words and phrases are different from each other is configured to control the display elements of each of the plurality of important words and phrases to be different from each other. Contains information for controlling display in a display mode with added display colors,
Among the important words and phrases expressed in the claim information, the important words and phrases are composed of the same character string as any one of the plurality of important words and phrases whose character strings are different from each other. The display control information is for controlling the display element of the important term to be displayed in a display mode with the same display color added as the display mode of the display element of the one important term. Contains information;
The specification information representing the contents of the specification includes, for each important term extracted based on the conditions for extracting important terms among the terms in the specification, the display mode of the display element of the important term. Contains display control information for controlling the
Among the important words and phrases expressed in the specification information, the character strings constituting the important words in the claim information are the same as any one of the plurality of important words and phrases that are different from each other. The display control information controls the display element of the important term to be displayed in a display mode with the same display color as the display color added as the display mode of the display element of the one important term in the complaint information. Contains information for
Invention document analysis method. The display document is an HTML document including an HTML file in which the patent application related information is written in HTML (Hypertext Markup Language) as content to be displayed;
The display control information relating to each of the plurality of important words and phrases in which the character strings forming the important words are different from each other in the claim information is such that the display element of each of the plurality of important words and phrases is displayed in a different display color for each important word. This is for controlling the display in a display mode in which a left border line having a predetermined width is added.
For each important phrase in the specification information, if the important phrase is the same as any one of the plurality of important phrases in the claim information, the above-mentioned important phrase related to the important phrase in the specification information The display control information includes a left border line having the predetermined width and a display color that is the same as the display color of the left border line according to the display mode of the display element of the one important word/phrase in the claim information. If the key word/phrase in the specification information is different from any of the plurality of key words/phrases in the claim information, the key word/phrase in the description information is The invention document analysis method according to claim 25 , wherein the display control information is for controlling the display element of the important word /phrase in a display mode in which the left border line having the predetermined width is not added. . The predetermined width is the width of one character,
The patent application related information includes an important word list in which a plurality of important words in the claim information each having different character strings constituting the important words are listed on a new line for each important word,
For each important word in the important word list, display control information for controlling the display mode of the display element of the important word is added,
The display control information regarding the important phrase in the important phrase list is such that the important phrase is one of the important phrases of the plurality of important phrases in the claim information in which the character strings forming the important phrases are different from each other. If it is the same as the word/phrase, the display element of the important word/phrase is set to the left side having the predetermined width in the same display color as the left border line according to the display mode of the display element of the one important word/phrase in the claim information. 27. The invention document analysis method according to claim 26 , wherein the invention document analysis method is for controlling display in a display mode with border lines added. Each line of the important word phrase list includes one important word and a corresponding translation word corresponding to the important word or phrase, or a translation word candidate set that is a set of candidates for corresponding translation words for the important word phrase,
The display control information related to the important words in the important words list is:
Controlling the display elements of the set of translation word candidates for the important word/phrase so as to function as a graphical user interface that allows any element of the set of translation word candidates to be selected in response to an operation. and
An element of the translation word candidate set that is selected in response to an operation on a display element of the translation word candidate set that is functioning as the graphical user interface is used as a corresponding translated word corresponding to the important word. 28. The invention document analysis method according to claim 27 , further comprising a script for executing a process of determining the corresponding translation word, and replacing and displaying the content of the display element of the set of translation word candidates with the determined corresponding translation word. The display control information related to the important term in the patent application related information changes the display element of the important term to a web page containing the important term, the definition of the important term, on a search site on the network, depending on the operation. , functions as a graphical user interface that executes processing to search for synonyms of the important word or phrase, or images related to the important word or phrase, or processing to have the important word or phrase translated by a translation site on the network. 29. The invention document analysis method according to any one of 25 to 28 . The display control information related to the important term in the patent application related information changes the display element of the important term to display at least part of the attribute information of the important term or the attribute of the important term, depending on the operation. It functions as a graphical user interface for displaying the importance of the important word/phrase determined based on the information,
The attribute information of the important term is
If there is a claim section that defines the important term in the claim represented by the claim information, information indicating the section number of the section;
If there is a claim section that refers to the important word/phrase in the claim represented by the claim information, information indicating the section number of the section;
The invention document analysis method according to any one of claims 25 to 29, further comprising information indicating a code when there is a description in which a code is added to the important word in the specification represented by the specification information. . The display control information related to the important word/phrase in the patent application related information changes the display element of the important word/phrase to an important word/phrase that lists lines including the important word/phrase from the claim information and the specification information according to the operation. It functions as a graphical user interface for executing inclusion description enumeration processing,
The above-mentioned important word inclusion description enumeration process enumerates and displays the lines containing the important words to be operated from the above-mentioned claim information and the above-mentioned specification information, and highlights the important words in each of the enumerated and displayed lines. , and the displayed line has a function for making the line function as a graphical user interface for displaying links to multiple lines including the line in the claim information or statement information, depending on the operation on the line. The invention document analysis method according to any one of claims 25 to 30, which is a process of adding display control information. The patent application related information includes claim translation information representing the result of machine translation for the claim represented by the claim information, and specification translation information representing the result of machine translation for the specification represented by the specification information. including
In the claim translation information, a line or paragraph representing the result of machine translation of a line or paragraph of the claim information is a complete translation of the entire row or paragraph of the claim information in the target language. line, or a partial translated line that is the result of replacing an important word or phrase included in the line or section of the claim information with a corresponding translated word corresponding to the important word or phrase, or a set of translation word candidates for the important word or phrase. can be,
In the specification translation information, a line representing the result of machine translation for one line of the specification information is a complete translation line expressing the entire line of the specification information in the target language, or A partial translation line that is the result of replacing an important term included in a line of specification information with a corresponding translation word corresponding to the important term or a set of translation word candidates for the important term;
The claim translation information and the specification translation information may be used to translate important words and phrases that are the same as any one of the plurality of important words and phrases in which the character strings forming the important words in the claim information are different from each other. Contains display control information for controlling the display mode of a display element as a result of machine translation that targets words for machine translation, and the display control information related to the display element as a result of machine translation includes the display mode of the display element. for controlling the display mode of the display element so that the same display color as that added to the display element of the important word in the claim information and the specification information is added to at least a part of the display element; The invention document analysis method according to any one of claims 25 to 31 . The claim translation information includes each line of the claim information, and for each line or item of the included claim information, immediately after the content of the line or item, a machine translation of the content of the line or item is provided. The complete translation line or the partial translation line representing the result of machine translation corresponding to the target is inserted,
The specification translation information includes each line of the specification information, and for each line of the included specification information, immediately after the line, the above-mentioned statement indicating the result of machine translation corresponding to the machine translation target that is the one line is provided. The complete translation line or the partial translation line is inserted,
The claim translation information and the specification translation information include the display mode of display elements of at least a portion of the machine translation target in the claim translation information and the specification translation information, and the result of machine translation corresponding to the machine translation target. In order to control the display mode of at least a portion of display elements to be different in at least one of text background color, text color, text decoration, font size, font style, and font weight, each display The invention document analysis method according to claim 32 , wherein display control information is added to the element. The patent application related information includes warning information,
The warning information is
Inspecting whether the claims or the specification of the invention document satisfy a plurality of conditions stipulated in the predefined standard provisions for invention documents for translation, and pointing out if the conditions are not satisfied. information and
predetermined based on the results of an examination as to whether the claim or the specification of the invention document satisfies at least some of the conditions set forth in the standard provisions for invention documents for translation. invention document evaluation information indicating an evaluation value of the conformity of the invention document to the invention document standard provisions for translation, which is calculated by the calculation;
provisional access information for accessing the invention document standard provisions for translation;
The invention document analysis method according to any one of claims 25 to 33, further comprising version information indicating a version of the invention document standard provisions for translation. A device equipped with a microprocessor is made to perform an invention document analysis process that analyzes an invention document, which is a document containing claims and specifications for a patent application expressing a technical idea in Japanese , and generates a display document. An invention document analysis processing program for realizing the invention document analysis method according to any one of claims 25 to 34 .

Images