JPH0682380B2 - Machine translation device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION <Industrial field of application> The present invention relates to a machine translation device capable of producing a correct translation result while a machine and a human interact with each other.

<Prior Art> Conventionally, there is the following as an interactive machine translation device. This machine translation device is, for example, "They consider hi
After the translation source text "ma good teacher" is entered from the keyboard and the translation process is performed, when you move the cursor to the "good" portion of the displayed translation source text, as shown in the upper part of Fig. 8, Both "good" and its translation "good" are highlighted. Furthermore, in this state "goo
When the word translation display key is operated to display the word translation of "d", a window of the translated word group is displayed based on the dictionary as shown in the lower part of FIG.

<Problems to be Solved by the Invention> However, since the above-described conventional machine translation device displays a word for which a translated word is to be displayed and only the translated word when displaying a word translation, However, there is a problem that it is difficult to find out what preposition word (preposition or adverbial adverb) is associated with the above word to form an idiom, and the syntactic analysis result cannot be grasped well.

Therefore, an object of the present invention is, when performing word translation display, when it is explained that a word to be displayed forms an idiom with a preposition word, the preposition word forming an idiom with the above word is described above. Another object of the present invention is to provide a machine translation device capable of correctly and promptly grasping a syntactic analysis result by displaying it together with a translation of an idiom.

<Means for Solving Problems> A morpheme analysis unit that obtains information such as a translated word by dividing an input translation original sentence into morphemes using a dictionary, a syntax analysis unit that analyzes the syntax of the morpheme sequence, and an acquisition unit. In the machine translation device having a conversion unit for converting the structure of the obtained syntax into the structure of the syntax of the target original language, and a translation sentence generation unit for generating a translation sentence according to the structure of the syntax of the target original language, in the dictionary, words and , A translated word of the word, a preposition word that co-occurs with the word to form an idiom, a grammatical rule based on a surface case pattern of the word and the preposition word, and an idiom formed by the word and the preposition word Is registered in association with each other, the word idiom translation display instruction for instructing to display the specific word in the translation original sentence or the idiom formed by the specific word in association with the translated word Means, and when there is an instruction from the word idiom translation display instruction means, when the translation of the specific word is selected by the morphological analysis unit based on the dictionary, the specific word and its translation are associated with each other. On the other hand, when the translated word of the idiom formed by the specific word is selected by the morpheme analysis unit, a word idiom translation display control means for displaying the idiom and the translated word in association with each other is provided. It has a feature.

<Operation> When the translation original sentence is input, the morpheme analysis unit divides it into morphemes using a dictionary to obtain information such as translated words, and the syntax of the morpheme sequence is analyzed by the syntax analysis unit. The structure is converted into the structure of the syntax of the target original language by the conversion unit, and the translation sentence is generated and output by the translation sentence generation unit according to the structure of the syntax of the target original language.

When the translated text obtained by the translation processing operation is displayed together with the original translation, it is possible to display the specific word in the original translation or the idiom formed by the specific word in association with the translated word. When instructed by the display instructing means, when the translation of the specific word is selected by the morpheme analysis unit, when the translation of the idiom formed by the specific word is selected based on the grammatical rule of the dictionary, The word idiom translation display control means displays the idiom and its translation in association with each other.

On the other hand, when the translated word of the specific word is selected by the morphological analysis unit, the specific word and the translated word are displayed in association with each other by the word idiom translation display control means.

Therefore, when performing the word idiom display, if the specified specific word is analyzed to form an idiom with the preposition word, the preposition word can also be displayed.

<Example> Hereinafter, the present invention will be described in detail with reference to illustrated examples.

FIG. 1 is a block diagram of an embodiment of the machine translation device of the present invention. 1 is the main CPU (central processing unit), 2
Is a main memory, 3 is a CRT (cathode ray tube), 4 is a keyboard, 5 is a translation module, and 6 is a memory that stores a dictionary for translation, grammar rules, tree structure conversion rules, and the like.

When the source language is input, the translation module 5
It translates it and outputs the target language.
That is, the source language input from the keyboard 4 is sent to the translation module 5 under the control of the main CPU 1. The translation module 5 uses the dictionary, grammar rules, tree structure conversion rules, etc. stored in the memory 6 to translate the input source language into a target language as described in detail later. The result is once stored in the main memory 2 and displayed on the CRT 3.

The specific configuration of the translation module 5 of this embodiment is shown in FIG.

The translation processing operation will be described below with reference to FIGS. 1 and 2.

First, the read original sentence is used by the dictionary lookup morpheme analysis unit 11 to use the dictionary of the memory 6 for each morpheme sequence (word sequence).
The grammatical information such as the part of speech and the translated word for each word are obtained, and the tense, the personal name, the number, etc. are analyzed. Then, each obtained information is stored in the dictionary lookup buffer. The syntactic analysis unit 12 determines a structural analysis tree indicating a dependency relation between each word according to the dictionary of the memory 6 and the grammar rules.

Then, the conversion unit 13 uses the tree structure conversion rule of the memory 6 to convert the structure of the input parse tree for the translated original sentence into the structure of the parse tree for the translated sentence.
Then, the translated sentence generation unit 14 attaches an appropriate particle or auxiliary verb as a translated sentence in the target language, and the obtained translated sentence is output from the translation module 5, and the main memory 2
It is stored in and displayed on CRT3 along with the translated text.

After that, when the word translation display key that is the word idiom translation display instruction means is operated as necessary, a specific word, or an idiom formed by a word and a preposition word, is displayed in association with the translation word. Idiom translation display processing is performed.

FIG. 3 shows a part of the contents of the dictionary lookup buffer that stores each information obtained as a result of the dictionary lookup by the dictionary lookup morpheme analysis unit 11.

Fig. 4 shows "come" in the dictionary lookup buffer of Fig. 3.
Is a part of the information described in the dictionary of the memory 6. One of this information is the verb "com
The grammatical rule based on the surface case pattern related to "e" indicates the following. That is, in order to output the translated word “come” of the translated word No. 1, the minimum unit verb phrase must be composed of only verbs. Also, translation No. 2
In order to output the translated word "encounter" of, the minimum verb phrase must be composed of a verb, a preposition = on, and a noun. Here, the grammatical rule in this case is that "verb phrase → verb + preposition + noun" applied in translation word No. 2 takes precedence over "verb phrase → verb" applied in translation word No. 1. Shall be.

FIG. 5 shows a flow chart of the above translation processing operation. Hereinafter, the translation processing operation according to this embodiment will be specifically described with reference to FIGS. 3, 4, and 5.

In step S ₁ , the translation source text is input.

Here, it is assumed that (1) He came here. Or (2) He came on a friend there. Is input as the original translation.

In step S ₂ , the dictionary lookup processing is executed by the dictionary lookup morpheme analysis unit 11 to obtain grammatical information such as a part of speech and a translated word, and further analysis of tense, person, number, etc. is performed.

In step S ₃ , the parsing by the parsing unit 12 is performed,
The structure of the parse tree is determined.

In step S ₄ , the conversion unit 13 converts the structure of the source language parse tree to the structure of the target language parse tree.

In step S ₅ , an appropriate translated sentence is generated by the translated sentence generation unit 14.

In step S ₆ , the translation generated in step S ₅ above is
It is displayed on CRT3 along with the translated text.

In such a translation processing operation, when translation is performed according to the information in the dictionary shown in FIG. 4, when the translation original sentence (1) is input, the grammatical rule “verb phrase → verb” is applied, He has come here. "

On the other hand, when the translated original sentence (2) is input, the grammatical rule of "verb → verb + preposition (on) + noun" is applied,
The output sentence "He met a friend there." Is obtained. In this case, as a result of dictionary lookup by the dictionary lookup morpheme analysis unit 11, as shown in FIG. 3, "preposition" information stored in the dictionary lookup buffer is "on" in the translation source sentence (2).
As shown in FIG. 4, among the information described in the verb “come” of the dictionary, the preposition word “on” of the information whose grammatical rule is “verb phrase → verb + preposition + noun” This is because the grammatical rules have been established as a result of matching.

In step S _7, whether a word translation display key by the operator has been operated or not. As a result, if it is operated, the process proceeds to S ₈ to perform the word idiom translation display process, and if not, the process proceeds to step S ₁₁ .

Step S _8, in step S ₉ and step S _10, the designated one of the words in the translation original "come", the preposition words forming the idiom with the "come", "on", idiom "come on" The translated word groups and are displayed in a window as shown in FIG. 6, and the process returns to step S ₇ . At this time, if the one instructed word does not form an idiom, only the instructed word and the translation group of the word are displayed in the window.

Here, the translated word 0 “encountered” in FIG. 6 is “come on”
Is the translated word currently displayed on the CRT3 as a result of being translated by the translation processing operation, and the translated word 1 "encountered" selected in the dictionary lookup processing has its ending changed to match the tense, state, etc. of the original translation. It is a thing.

In step S _11, whether or not the next sentence is determined. As a result, if there is a next sentence, the process returns to step S ₁ , and if not, the translation processing operation ends.

As described above, when the translation source text is translated in this embodiment, when the verb "come" and the preposition "on" are analyzed as having an idiom relationship as in the translation source text (2), the word idiom translation is performed. In the display, the verb "come"
Since the preposition word “on” and the translated word group of the idiom are displayed, it is possible to clearly indicate the relationship between the words forming the idiom and the preposition word. Therefore, the syntactic analysis result can be better understood, and the operator's burden is reduced.

Further, when the idiom is parsed into an incorrect idiom, it is necessary to break the idiom relationship between the word forming the incorrect idiom and the preposition word. In that case, if the preposition word is displayed as described above, the idiom relationship between the word and the preposition can be easily broken using the preediting support function, and the correct translation process can be performed quickly. it can.

In the above example, the verb "co that forms the idiom
The explanation is given by taking the case where "me" and the prepositional word "on" are chained as an example. However, in this invention,
The words forming the idiom and the prepositional words do not necessarily have to be linked.

For example, if the prepositional word is an adverbial noun,
The word order is "verb + adverbial noun + noun phrase" or "verb + noun phrase + adverbial noun". In the case of the latter word order, the word (verb) forming the idiom and the preposition word (adverbial adverb) are located at positions apart from each other.

Even when the preposition word is a preposition, an idiom is formed with a word located at a distant place as in the following example sentence.

This is the same type as AAA. “AAA”: word Further, in the case of an original sentence having a relative clause, etc., the word order forming the idiom and the preposition word are interchanged as in the following example sentence.

AAA on which BBB depends. "AAA", "BBB": word In this way, even if the words that make up the idiom to be displayed and the preposition words are located at distant positions, or even if they are positioned interchangeably. As shown in FIG. 4, in the dictionary of the memory 6, the words forming the idiom and the preposition words, the grammatical rules based on the surface pattern of the words and the preposition words, and the translation words of the idiom are associated with each other. By registering the idiom and its translation, the idiom and its translation can be easily associated and displayed.

In the above embodiment, as shown in FIG. 6, the designated word, the preposition word, and the translated word group of the idiom are displayed, but as shown in FIG. 7, the word in the designated translation original sentence is displayed. The "come", the preposition word "on" forming the idiom together with the "come", and the translated word "encountered" corresponding to the idiom "come on" may be highlighted.

<Advantages of the Invention> As is apparent from the above, the present invention provides a machine translation device having a dictionary, a morpheme analysis unit, a syntax analysis unit, a conversion unit, and a translation sentence generation unit, and a word idiom translation display instruction means and a word idiom translation. When the display control means is provided and the word idiom translation display instruction means instructs to display the specific word or the idiom formed by the specific word in association with the translated word, the specific word is formed by the morphological analysis unit. When the translated word of the idiom to be selected is selected, the word idiom translation display control means displays the idiom and the translated word in association with each other, so that the specific word for which the translated word is to be displayed forms the preposition word and the idiom. The above preposition word can also be displayed if it is analyzed. Therefore, the syntactic analysis result can be correctly and promptly grasped.

At that time, in the above dictionary, the word, the translation of the word,
Preposition words that co-occur with the above words to form idioms,
Grammar rules based on the surface case pattern relating to the word and the preposition word, and the translation of the idiom formed by the word and the preposition word are registered in association with each other, so the word forming the idiom and the preposition The idiom and its translated word can be easily and correctly associated and displayed even if the word is located at a place distant from the word or if the word is located in an alternate position.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of a machine translation apparatus of the present invention, FIG. 2 is an operation configuration diagram of a translation module in the above embodiment, FIG. 3 is a diagram showing an example of contents of a dictionary lookup buffer, and FIG. FIG. 7 is a diagram showing an example of information for “come” written in the dictionary, FIG. 5 is a flow chart of the translation process according to the above embodiment, FIG. 6 is a diagram showing a display example by the word idiom translation display process, and FIG. Is a diagram showing a display example different from FIG. 6, and FIG. 8 is a diagram showing a display example by the conventional word translation display processing. 1 ... Main CPU, 2 ... Main memory, 3 ... CRT, 4 ... Keyboard, 5 ... Translation module, 6 ... Memory, 11 ... Dictionary morphological analysis unit, 12 ... Syntax analysis unit, 13 ...... Conversion unit, 14 …… Translation sentence generation unit,

 ─────────────────────────────────────────────────── ─── Continuation of front page (72) Inventor Shuzo Kugimiya 22-22 Nagaike-cho, Abeno-ku, Osaka-shi, Osaka Within Sharp Co., Ltd. (72) Noriyuki Hirai 22-22 Nagaike-cho, Abeno-ku, Osaka-shi, Osaka Incorporated company (56) References JP 61-54565 (JP, A) JP 61-150068 (JP, A)

Claims (1)

[Claims] 1. A morpheme analysis unit that obtains information such as a translated word by dividing an input translation original sentence into morphemes using a dictionary, a syntax analysis unit that analyzes the syntax of the morpheme string, and a structure of the obtained syntax. In a machine translation device that includes a conversion unit that converts a to a target language syntax structure and a translation sentence generation unit that generates a translation sentence according to the target language syntax structure, the dictionary includes a word and a translated word of the word. And a preposition word that co-occurs with the word to form an idiom, and a grammatical rule based on a surface case pattern relating to the word and the preposition word,
A translation of an idiom formed by the word and a preposition word is registered in association with each other, and a specific word in the translation original text or an idiom formed by the specific word is displayed in association with the translation word. If there is an instruction from the word idiom translation display instruction means to instruct, and the word idiom translation display instruction means, if the translation of the specific word is selected by the morphological analysis unit based on the dictionary, the specific word And the corresponding word are displayed in association with each other, and when the translation word of the idiom formed by the specific word is selected by the morphological analysis unit, the word idiom translation display in which the idiom and the translation word are displayed in association with each other A machine translation device comprising control means.