JP4350566B2 - Machine translation system - Google Patents

Description

The present invention relates to a machine translation system you translate natural language into another natural language, in particular, relates to a suitable machine translation system to contribute to understanding of the improvement of translation.

  As business has become more internationalized and borderless, the network society has also developed, and digital information exchange has become very common among ordinary users, such as the Internet and e-mail. Accompanying the widespread use of such network environments, information is frequently acquired through the WWW (world wide web) on the Internet.

  Most of the information available on the Internet is in English, and if English is not a native language and is poor, users have typically tried to do machine translation to understand the meaning of the original text. In general, the quality of machine translation that translates from foreign languages to Japanese is at a reasonable level.

  Recently, in addition to this, the necessity of transmitting information from Japan has been screamed, and this requires more linguistic knowledge of the target language than required by the passive ability of reading the target language. The demand for machine translation systems that translate native languages into foreign languages is increasing, but the quality of the translations is not yet commensurate. One reason for this is that Japanese-English translation is taken as an example. From the viewpoint of naturalness, the quality of the translated word in the commercial Japanese-English dictionary used as a reference was improved when developing the bilingual dictionary used in machine translation. May not be appropriate. Therefore, there is the present situation that machine translation is also dragging it.

  Furthermore, unless the original text that is intended for information transmission is written with the intention of translation, it is intended only for native speakers of that language. Therefore, in the case of human translation, annotations are made as necessary. However, as the translation method of translation software is basically a verbatim translation, in principle, only the translations of the words that existed in the original text are displayed. It is not usually taken into account. Even if the parsing is successful as a translation, the purpose of information transmission cannot be achieved unless the reader can understand the intention.

  As an example, some bilingual dictionaries only list transliteration as a translation, but this is a word in the first language, broadly unless you have knowledge of the local culture where the language is used as your native language, The extension of the word is unknown to the reader. Explanations of such words are provided in the monolingual dictionary of the first language, but the reader must examine the meaning of each word that cannot be understood among the words that appear in the translation. Don't be. This itself takes effort, but reading the meaning also takes a burden. However, the contents of the monolingual dictionary cannot be read by a native speaker of the second language who does not know the first language unless there is some support.

  There is an attempt to use a monolingual dictionary of the first language for machine translation. For example, in Japanese Patent Laid-Open No. 2002-334078 (Patent Document 1), the meanings described in the monolingual dictionary of the first language are added to the bilingual dictionary (between the first language and the second language). It is described that different bilingual dictionaries (between the first language and the third language) are constructed by combining with bilingual dictionaries (between the second language and the third language).

  The annotation creation function is described in, for example, Japanese Patent Laid-Open No. 9-44486 (Patent Document 2). This uses the contents of the dictionary as annotation information as it is, and is not related to translation. In addition, annotations are created only for the words specified by the user.

To summarize the current situation, it is not realized that the information of the monolingual dictionary is fully utilized for translation, and the contents in the monolingual dictionary are good for the user (the native speaker of the second language). It is not considered to present information in an easy-to-understand form such as In this way, at present, it is not possible to sufficiently meet the needs of machine translation, which can save labor for the examination as much as possible and present necessary and sufficient information to the reader of the target language in an easy-to-understand manner.
JP 2002-334078 A JP 9-44486 A

As described above, in conventional machine translation, the reader does not use the meaning of the monolingual dictionary of the first language and does not display the information in an easy-to-understand manner. You will feel constraints. The present invention has been made in view of such circumstances, there in the machine translation system you translate natural language into another natural language, Ru can contribute to the understanding of the improvement of the translation machine translation system The purpose is to provide.

Machine translation system according the present invention, with reference to the translation dictionary storage means the stored translation dictionary, and means for translating a sentence in the first language to be translated sentences in the second language, before Symbol second language When a relative pronoun or subordinate clause is used in the translated word that appears in the translated sentence, it corresponds to a means for determining that the translated word is an explanatory translation, and to a translated word determined to be an explanatory translation Means for transliterating words in the sentence in the first language based on the reading data of the words in the word / idiom dictionary stored in the translation dictionary storage means, and appearing in the sentence translated into the second language Means for adding a translated word of the transliteration to a sentence translated into the second language for a portion determined to be an explanatory key translation.

According to which this, the concept of translation description tone because it is possible to indicate a translation (transliteration) that a cohesive, can contribute to comprehension improvement easier to see that the translation for the reader. Each means can be configured by hardware such as a microprocessor, a memory, and a hard disk, and basic software such as an operating system operating on these hardware, application programs, and the like.

According to the present invention, it can contribute to the machine translation system you translate natural language into another natural language, comprehension improvement of translation.

Based on the above, embodiments of the present invention will be described below with reference to the drawings. First, FIG. 1 is a block diagram showing a configuration of a premise comprising machine translation system implementation form (i.e. see Example) of the present invention. This machine translation system includes a machine translation program. As shown in the figure, an input unit 11, a system control unit 12, an editing unit 13, a translation unit 14, a translation dictionary unit 15, a monolingual dictionary unit 16, and a display are displayed. Each unit of the unit 17 and the document storage unit 18 functions. The relationship between the machine translation program and machine translation system also applies to the following implementation forms. In the following, it will be described a first language as an example Japanese, the second language as English.

  The input unit 11 inputs sentences and sentences in the first language to be translated through communication means such as the Internet, a keyboard, and the like. Also, a command for instructing an operation as a machine translation system is input. The input sentence or command is sent to the system control unit 12.

  The system control unit 12 controls the entire machine translation system in an integrated manner. For example, a sentence or sentence obtained from the input unit 11 is sent to the display unit 17 for display on the display unit 17, sent to the editing unit 13 for editing, sent to the translation unit 14 for translation, or a document storage unit for storage. The operation such as sending to 18 is performed. Further, the operations of the editing unit 13, the translation unit 14, the translation dictionary unit 15, the document storage unit 18, and the display unit 17 are controlled in accordance with commands from the input unit 11. For example, data in the translation dictionary unit 15 is taken out and sent to each unit, a sentence to be translated is selected from a sentence input according to a command, or a dictionary to be worked out is selected from the dictionary of the translation dictionary unit 15.

  In addition to editing the sentence in the first language selected by the system control unit 12, the editing unit 13 displays the target editing displayed on the display unit 17, that is, in response to an instruction such as a command input from the input unit 11. It controls the movement of the cursor above, the setting of the use of the monolingual dictionary unit 16, the display of the translated text obtained by the translation process in the translation unit 14 on the display unit 17 via the system control unit 12, and the like. .

  The translation unit 14 performs translation processing on the sentence in the first language selected by the system control unit 12 while referring to the dictionary of the translation dictionary unit 15 selected by the system control unit 12. The translation dictionary unit 15 stores and holds various dictionaries used when the translation unit 14 performs translation processing. The contents of various dictionaries included in the translation dictionary unit 15 will be described later. The monolingual dictionary unit 16 is composed of a first language entry word and its meaning, and is referred to when the translation unit 14 performs translation.

  The display unit 17 displays, under the control of the system control unit 12, a first language sentence (original sentence) input from the input unit 11, a second language sentence translated by the translation unit 14 (translation sentence), and the like. Is. The document storage unit 18 stores and holds the text in the first language and the text in the second language independently under the control of the system control unit 12, and the second language as the translation result of the first language. Data is stored and retained in association with each other.

  As shown in the figure, the translation dictionary unit 15 includes a utilization change dictionary 15a, an analysis grammar dictionary 15b, a word / idiom dictionary 15c, a conversion grammar dictionary 15d, a generation grammar dictionary 15e, and a morpheme generation grammar dictionary 15f. The utilization change dictionary 15a is a dictionary for converting words and idioms whose endings and the like are changed into original forms of the first language. The analysis grammar dictionary 15b is a dictionary in which a grammar for analyzing the first language is stored. The word / idiom dictionary 15c is a dictionary in which translations of the second language corresponding to the words / idioms of the first language are stored together with their part of speech information and translation type information (for example, classification such as transliteration, explanation translation, one-to-one translation). It is. The conversion grammar dictionary 15d is a dictionary in which structure conversion rules between the first language and the second language are stored. The generation grammar dictionary 15e is a dictionary for determining the structure of the second language sentence. The morpheme generation grammar dictionary 15f is a dictionary for completing a translated sentence by changing the word form such as the end of the second language.

  Next, the operation of the machine translation system shown in FIG. 1 will be described with reference to FIG. FIG. 2 is a flowchart showing an operation flow of the machine translation system shown in FIG. First, an original sentence (first language sentence to be translated) is taken into this machine translation system using the input unit 11 (step 21). The taken-in sentence is displayed on the display unit 17 by the system control unit 12 and stored / held in the sentence storage unit 18. Here, as a specific example, it will be described below that the captured sentence is Japanese displayed on the display unit 17 as shown in FIG. FIG. 5 is a diagram showing an example of a window displaying a sentence in the first language (Japanese) to be translated.

  When the sentence in the first language is displayed on the display unit 17, the cursor is positioned on the sentence to be translated among the sentences. The first sentence is usually the first sentence (step 22), but here the translation for the first sentence is over and the cursor is the second sentence: The following steps will be explained.

  Next, the user inputs a command for instructing execution of an event from the input unit 11 (step 23). If the input command is not “end” or “translation”, the processing corresponding to the command is executed (step 29), and the process returns to step 23 where the command can be input again. If the input command is “end” (Y in step 24), the operation is ended. If the input command is “translation” (Y in step 25), the process proceeds to a translation process (step 26).

  The movement of the cursor that designates the translation target is automatically moved to the next sentence when the translation of the sentence is completed in order from the first sentence. In addition, the movement of the cursor can be specified by a command. In this case, the cursor moves by executing the process of step 29. Internally, the editing unit 13 moves the cursor on the display unit 17 via the system control unit 12 according to the command. In this embodiment, a sentence in which the cursor is positioned (for example, the cursor is positioned at the head thereof) is a translation target when a translation command is instructed. This is because the movement of the cursor corresponds to, for example, when it is desired to translate not only the sentence at the current cursor position but also the sentence before and after the original sentence.

  The translation processing (step 26) is performed by the system control unit 12 in an integrated manner. The details of this processing flow will be described from the next paragraph with reference to FIG. When the translation process is completed, the translation process result is displayed on the display unit 17 (step 27). Further, the cursor for designating the translation target is moved to the next sentence (step 28). Then, the process returns to step 23 where the command can be input again. Thereafter, the translation process can proceed in the same flow.

  Step 26 of the translation process will be described with reference to FIG. FIG. 3 is a flowchart showing the details of step 26 shown in the flowchart shown in FIG. 2. Here, the case of translating from Japanese to English will be described. Here, the translation object is the second sentence of the sentence shown in FIG. 5 as an example as described above: “Because the traditional Japanese needlework and doll festival are held at this time”.

  First, as step 26, the system control unit 12 causes the translation unit 14 to execute a translation process for translating the original text in the first language into the translated text in the second language (step 31). For this procedure itself, a known machine translation procedure may be used. For example, the translation unit 14 first uses a utilization change dictionary 15a or an analysis grammar dictionary 15b for a sentence to be translated, and converts a word or phrase that has a change in ending to its original form, Perform grammar analysis. Next, morphological analysis is performed with reference to word information in the word / idiom dictionary 15c. Next, referring to the analysis result using the analysis grammar dictionary 15b, the structure conversion from the first language to the second language is performed using the conversion grammar dictionary 15d. Further, it is a procedure of referring to the word / idiom dictionary 15c, converting each word individually, and generating a translation using the generation grammar dictionary 15e and the morpheme generation grammar dictionary 15f.

  Next, it is assumed that the assumed translated reader has little knowledge of the area where the first language is used (for example, knowledge of the culture), and a process for dealing with it is executed. Specifically, the translation unit 14 first extracts words / phrases whose translation word type information in the word / phrase dictionary 15c is “explanatory” or “transliteration” from phrases in the Japanese sentence (step 32). . The fact that the translation type information is “explanatory” or “transliteration” means that the concept does not match as a simple translation, and the original phrase is different in the region where the second language is used. It is possible.

  The explanatory translation is a translation such as “fish and vegetables boiled in soy with sugar”, and the transliteration is a translation such as “Suizobyo Kenkyu Zaidan”. Examples of other translation types include a one-to-one correspondence translation, which is a translation such as “magnetic resonance imaging”. In this embodiment, it is assumed that the word / idiom dictionary 15c has such translation type information for each translation of the second language.

  When a word / phrase whose translation type information is “explanatory” or “transliteration” is extracted, the translation unit 14 acquires the meaning of the extracted word / phrase from the monolingual dictionary unit 16. The extracted word and the meaning acquired from the monolingual dictionary unit 16 for the word are displayed on the display unit 17 via the system control unit 12 (step 33). The display result is as shown in FIG. 6, for example. FIG. 6 is a diagram showing an example of a window displaying words and phrases that need the extracted supplementary items and meanings obtained from the monolingual dictionary for the words and phrases. Here, "needle offering" is extracted and its meaning is: "February 8 (December 8 in some regions), needles broken for plain sewing, etc. are stabbed in tofu and konjac or delivered to Awashima Shrine “To give me a rest. That day, I will cease sewing for one day.” Is displayed. Of course, there may be a plurality of pairs of displayed words and their meanings. Note that “needle service” is “memorial service for broken needles” as a translated word in the word / idiom dictionary 15 c and is included in the internal translation by the processing in step 31.

  Next, the user instructs the input unit 11 as to whether or not to use the meaning displayed on the display unit 17 in the translated sentence (step 34). When there is an instruction for importing into the translated sentence, the translation unit 14 translates the acquired meaning using the translation dictionary unit 15 and displays the result on the display unit 17 via the system control unit 12. (Step 35). For example, `` Hold a memorial service on February 8 (it is December 8 depending on a district) by stabbing tofu and konnyaku with the needle which broke used usually in needlework, etc., or dedicating it to Awashima Jinja. '' Display of the correct translation result. This translation result is stored and retained in a pair with the translation “memorial service for broken needles” in the word / idiom dictionary 15c.

  In this display, for example, a window as shown in FIG. 4 is simultaneously displayed on the display unit 17. FIG. 4 is a diagram illustrating an example of a display window in which a user specifies a method for using the meaning of a Japanese dictionary (single language dictionary). By checking a window as shown in FIG. 4, the user can specify how to use the meaning of the Japanese dictionary (step 36). Here, “add as supplement” and “use as translation” can be selected. “Add as supplement” means that the translated meaning is added in brackets after the translated word, for example, and “use as translated word” means the corresponding translated word (word This means that the translated meaning is used so that it is replaced with the translated word in the idiom dictionary 15c.

  If the method of using the meaning translation result is designated by the user, the usage method information is retained (step 37). This information is referred to in step 27 (FIG. 2) for displaying the subsequent translation processing result.

  When the translation process is performed on all the Japanese sentences shown in FIG. 5 according to the operation flow shown in FIGS. 2 and 3, for example, the display unit 17 displays the translation result as shown in FIG. Can get to. In this example, the user selects “add as supplement” for the meaning of “needle feeding” in step 36. Further, when the user does not specify the translation itself in step 34, a display as shown in FIG. FIG. 7 shows an example of a window displaying a sentence translated from Japanese into English according to a normal procedure (FIG. 7A), and a sentence with supplements added to corresponding words that require supplementary items in the translated sentence. It is a figure which shows an example (FIG.7 (b)) of the displayed window.

  In the display in FIG. 7B, the supplementary matter is added in parentheses immediately after the explanatory word translation. However, since the translated word usually consists of a plurality of words, in the case of “add as supplement”, it is difficult to understand which range is the target of supplementation only by looking at the translated sentence. Therefore, in the example of this display, as shown in the figure, for example, an underline is added to the translation part of the explanatory note for the purpose of clarifying the supplementary target range, the font is changed, the size is changed, etc. It can be identified. In this example, the translated word meaning was not used because it was long, but “Use as translation” was selected instead of “Add as supplement” on the designation screen (FIG. 4) for using the monolingual dictionary. Does not use the translated word in the word / idiom dictionary 15c, and displays only the translation part from the meaning of the monolingual dictionary unit 16.

  In the above, the case where the translation word type information in the word / idiom dictionary 15c is a translation having an explanatory key translation has been described. The case where the translation word type information is a transliteration will be supplementarily described below. For example, the translation of the third sentence of the first language sentence shown in FIG. 5 is “The junior high school student of this school went out for inspection of the ceremony of the”, as shown in FIG. “memorial service for broken needles performed in Awashima Jinja in Asakusa on February 8.”, where “Awashima Jinja” has transliteration type information.

  There are several options for such transliteration. The first option is to search the monolingual dictionary unit 16 for “Awashima Shrine”, and when it is listed as a heading in the monolingual dictionary unit 16, the meaning of the word is as described in the explanation of “needle offering” Is to present. Alternatively, if the word / idiom dictionary 15c is listed as a heading but is not a heading word in the monolingual dictionary unit 16, as shown in FIG. It is conceivable that the language dictionary unit 16 is searched, and if it is a headword there, the meaning is translated and used. FIG. 8 is a flowchart showing a modification of step 33 in the flowchart shown in FIG.

  For example, “Awashima Shrine” can be divided into “Awashima” and “Shrine” by morphological analysis. If these words are listed as headwords using the monolingual dictionary unit 16, the meaning of "Awashima" and the meaning of "Shrine" can be used in the translation. Furthermore, after the morphological analysis, it is also possible to present the translation of the word / idiom dictionary 15c for the divided element without using the monolingual dictionary unit 16. According to this, a translation such as Awashima Shrine can be obtained. The display method at this time can be “Awashima (Awashima) Jinja (Shrine)” for each word break, or “Awashima Jinja (Awashima Shrine)” at once.

  Next, the case where an unknown word is generated in step 31 of FIG. 3 will be supplementarily described with reference to FIG. FIG. 9 is a flowchart showing a modification of the detailed flowchart shown in FIG. In FIG. 9, the same reference numerals are assigned to the processes corresponding to those in FIG. A description of this part is omitted. Here, the unknown word means a character string that cannot be further divided and cannot be subjected to morphological analysis. As a result, there is no translation corresponding to the word / idiom dictionary 15c.

  As an example of a sentence to be translated, mention is made of a sentence in the first language, “He compared micro credits in North Africa and Japanese Yoromo no Makoto”. It is assumed that “Yoromo-no-ko” became an unknown word in step 31 for translating this sentence. In this case, the translation of the conventional machine translation is output with the unknown word part in the first language character type, as in “He compared micro credit in Africa with Yoromo Maiko in Japan.” If there is data, it may be displayed in romaji.)

Therefore, in this example, it is assumed that a sequence of characters determined as an unknown word in the morphological analysis (a sequence of characters for which there is no translation in the word / idiom dictionary 15c) is a unit of a group of words. (Step 32a), and the monolingual dictionary unit 16 is searched by using it as a headword to acquire the meaning (step 33). Thus, the meaning, in this case, for example, “a mutual aid organization for the purpose of monetary interchange. A member pays a certain amount on a certain date and receives a certain amount of money by lottery or bidding. You can get information such as “What you do until you reach all the union members. What happened in the Kamakura period.” Therefore, by executing step 34 and subsequent steps, a translation of the meaning can be generated. Then, the translation result obtained as a result can be displayed by being written together with the character string of “Yori-Magoko” or its Roman alphabet, or by substituting it into that portion.

  When extracting the unknown word, if the first language is an agglutinative word such as Japanese, the range of the unknown word cannot be specified, and a plurality of possibilities may appear. This case will be supplementarily described with reference to FIG. FIG. 10 is a flowchart showing a modification of step 32a in the flowchart shown in FIG.

  In such a case, it is conceivable to display all possibilities and perform the same processing as described above for other possibilities. For example, there is a character string "... ABCDE ...". ABC and DE, 2. AB and CDE, 3. If there are three candidates that A is a part of the previous morpheme and BCDE is a group, all these possibilities are displayed (step 32aa). Here, meaning information may be provided to each candidate. The user selects what is considered to be correct from the candidates (step 32ab). It is also possible that a correct candidate cannot be derived as a machine translation system. In that case, the user may directly specify the correct unknown word range and set the specified character string as the target of monolingual search.

Will now be described with reference to FIGS. 11 to 13 for a machine translation system according to the implementation embodiments of the present invention. Figure 11 is a block diagram showing the configuration of a machine translation system according to the implementation embodiments of the present invention, parts that are the same as those described in FIG. 1 are denoted by the same reference numerals. The description of this part is omitted when there is no function to be added.

  This embodiment does not require the monolingual dictionary unit 16, and the configuration of the translation dictionary unit 15A is slightly different from the configuration of the translation dictionary unit 15 shown in FIG. Unlike the word / idiom dictionary 13c, the word / idiom dictionary 15ca of the translation dictionary unit 15A does not have translation type information, but has first-language reading kana data. In addition, a utilization change dictionary 15g and an analysis grammar dictionary 15h for the second language are newly provided. In this embodiment, the translated word type information is determined and detected from the translated words in the word / idiom dictionary 15ca, and a translated sentence is generated in consideration of the legibility as a sentence translated into the second language.

  FIG. 12 is a flowchart showing an operation flow of the machine translation system shown in FIG. In FIG. 11, the same reference numerals are given to the same processes as those in the flowchart shown in FIG. A description of this part is omitted. FIG. 12 is the same as FIG. 2 except that step 26 of translation processing and step 27 of displaying the result of translation processing are replaced with step 26A of translation processing.

  FIG. 13 is a flowchart showing details of step 26A shown in the flowchart shown in FIG. First, as step 26A, the system control unit 12 causes the translation unit 14 to execute a translation process for translating the original text in the first language into the translated text in the second language (step 41). For this procedure itself, a known machine translation procedure may be used. For example, the translation unit 14 first uses a utilization change dictionary 15a or an analysis grammar dictionary 15b for a sentence to be translated, and converts a word or phrase that has a change in ending to its original form, Perform grammar analysis. Next, morphological analysis is performed with reference to the word information in the word / idiom dictionary 15ca. Next, referring to the analysis result using the analysis grammar dictionary 15b, the structure conversion from the first language to the second language is performed using the conversion grammar dictionary 15d. Further, it is a procedure of referring to the word / idiom dictionary 15c, converting each word individually, and generating a translation using the generation grammar dictionary 15e and the morpheme generation grammar dictionary 15f.

  Next, a process for determining and detecting a part that is an explanatory key translation from the corresponding translation from Japanese to English is performed as follows. Here, nouns will be described as examples. As specific examples of nouns, “Awashima Jinja” and “Awashima Jinja” (transliteration), “Taste” and “fish and vegetables boiled in thick and soy with sugar” (explanatory translation), “gender equality society” and “ Assume four translations: “gender-free society in which both men and women can participate fully” (translation), “magnetic resonance imaging” and “magnetic resonance imaging” (one-to-one translation).

  In order to determine and detect a part that is an explanatory translation, first, the reading of a Japanese word is replaced with an alphabet (step 42). Next, the character string after the alphabet substitution is collated with the translated word (step 43), and if it matches, it is determined that it is a transliteration (step 44). In the above example of each noun, the pair of “Awashima Shrine” and “Awashima Jinja” corresponds to this. If the Japanese headword is a katakana and a foreign word, a collation technique between the original word and the katakana character string has already been established and can be used.

  If no match is found (N in step 43), whether or not the word / phrase can be divided using the utilization change dictionary 15a and the analysis grammar dictionary 15b storing the grammar for analyzing the first language is determined. Determine (step 45). If it can be divided, it is divided and the word / idiom dictionary 15ca is searched for each morpheme to obtain a possible translation (step 46).

Next, the acquired translation is analyzed using the target language (English) utilization change dictionary 15g and the analysis grammar dictionary 15h, and content words such as nouns and verbs are extracted. The relationship allowance nouns in its translation, if it meets the criteria, such as that the subordinate clause is being used (step 47) and determines that described regulating them as the type of translation is regarded as a descriptive element (Step 48). In the above example of each noun, the pair of “gender-equal society” and “gender-free society in which both men and women can participate fully” fall into this category.

  Next, for those that were not judged to be explanatory notes, a combination of these content words and possible translations obtained earlier was made, and either of the possible translations of each morpheme of the original word and the given translation were too short or large If there is no match, it is determined that the translated word has a one-to-one correspondence. In the above, “magnetic resonance imaging” and “magnetic resonance imaging” pairs fall into this category. Further, when the words partially match, it is determined as an approximate translation, and when the number of content words of the translated word exceeds the number of morphemes of the original word, it is determined as an explanatory note (steps 52 and 53). Whereas the criterion at step 47 is based on syntactic features, the criterion at step 52 is based on the degree of coincidence of the content words.

  If the original language cannot be further divided in step 45, it is analyzed using the target language (English) utilization change dictionary 15g and the analysis grammar dictionary h, and if the criteria such that the content word is 2 or more is satisfied ( Step 49), it is determined that it is an explanatory tone (Step 50). In the above, the pair of “boiled” and “fish and vegetables boiled in thick and soy with sugar” corresponds to this. Note that if the analysis processing in the above steps is not successful, the type of the translated word cannot be determined (steps 51 and 53).

  When the translated word type information is determined and detected in this way, the information is added to the translated word. Subsequently, the translation unit 14 can perform an appropriate translation process based on this information. For example, in the case of explanatory translation, it cannot be read in the translated state that it was originally a concept, so a transliteration (for example, “umani” for “boiled”) is created and used as the corresponding translation. Display in association with each other (step 54). Thereafter, the transliteration is used as a translation in the same sentence. As a result, the translation can be avoided from becoming unnecessarily complicated, and the translation can be easily viewed.

  As a supplement to making it easier to see, for example, the translation of “The government conducted a public opinion survey on gender-free society in which both men and A translation such as “women can participate fully and human rights” is obtained, but in this translation, the translations corresponding to “society” and “human rights” are separated. From the translation, it is difficult to read that they are in parallel, which hinders understanding of the meaning. In this regard, the parallel relationship can be easily understood by adding, for example, transliteration of “gender equality society” to the corresponding translation word. In addition, the corresponding translated words of the subsequent “gender equality society” are shown only by transliteration, making it easier to see the translation.

  In the translation process shown in FIG. 13, the determination is performed step by step using a plurality of scales in order to determine and detect the translated word type information, but it may be simplified to use only one scale. Is possible. That is, it is conceivable that the degree of coincidence between the translation candidate of the original word content word and the content word in the translated word is calculated, and explanation translation, approximate translation, etc. are determined based on the degree of coincidence.

As described above, in the embodiment with reference to FIG. 11 to FIG. 13, the form in which the translated sentence is generated in consideration of the legibility as the sentence translated into the second language has been described. The method of determining / detecting translated word type information described therein can of course be used in the previous reference example described with reference to FIGS. According to this, the word / idiom dictionary 15c in FIG. 1 does not need to have translation type information in advance.

As described above, the following effects can be obtained according to the implementation mode of the. That is , it is possible to avoid that the explanatory translation is directly embedded in the translation of the original sentence, so that the translation is easy to understand and view. In particular, when the frequency of appearance of the word in the document is high, it is possible to avoid repeating a long explanatory note many times.

Block diagram showing the configuration of a premise comprising machine translation system implementation embodiment of the present invention (i.e., see Example). The flowchart which shows the operation | movement flow of the machine translation system shown in FIG. 3 is a flowchart showing details of step 26 shown in the flowchart shown in FIG. 2. The figure which shows an example of the display window which a user designates the usage method of the meaning of a Japanese dictionary (single language dictionary). The figure which shows an example of the window which displayed the text of the 1st language (Japanese) which should be translated. The figure which shows an example of the window which displayed the phrase which needs the supplementary matter extracted, and the meaning acquired from the monolingual dictionary about the phrase. The figure which shows an example of the window which displayed the text which added the supplementary word to the corresponding word which requires the supplementary matter in a translation sentence, and the example which displayed the sentence translated from Japanese into English according to a normal procedure. The flowchart which shows the modification in step 33 in the flowchart shown in FIG. The flowchart which shows the modification of the detailed flowchart shown in FIG. The flowchart which shows the modification in step 32a in the flowchart shown in FIG. Block diagram showing the configuration of a machine translation system according to the implementation embodiments of the present invention. 12 is a flowchart showing an operation flow of the machine translation system shown in FIG. 11. The flowchart which shows the detail of step 26A shown in the flowchart shown in FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 ... Input part, 12 ... System control part, 13 ... Editing part , 14 ... Translation part, 15, 15A ... Translation dictionary part, 15a ... Utilization change dictionary (1st language), 15b ... Analysis grammar dictionary (1st language) 15c, 15ca ... word / idiom dictionary, 15d ... conversion grammar dictionary, 15e ... generation grammar dictionary (second language), 15f ... morpheme generation grammar dictionary (second language), 15g ... utilization change dictionary (second language), 15h: Analytical grammar dictionary (second language), 16: Monolingual dictionary unit, 17 ... Display unit, 18 ... Document storage unit.

Claims (1)

Means for translating a sentence in the first language to be translated into a sentence in the second language with reference to the translation dictionary stored in the translation dictionary storage means;
Means for determining with the translation is always described tone when any one of the previous SL subordinate clauses or relative pronoun in translation appearing in statements translated into the second language is used,
Means for transliterating words and phrases in the sentence of the first language corresponding to the translated words determined to be explanatory translations based on the reading kana data of the words in the word / idiom dictionary stored in the translation dictionary storage means When,
For a translation that appears in a sentence translated into the second language and is determined to be an explanatory translation, the translation by the transliteration is added to the text of the sentence translated into the second language. And a machine translation system.

Images