JP3389313B2 - Machine translation equipment - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a machine translation device,
In particular, it is suitable for translating a sentence including a complex sentence including a plurality of sections. [0002] In recent years, various types of machine translation devices using electronic computers have been developed. Normally, when performing translation using such a machine translation device, post-editing work is necessary to correct errors in the translated sentence output by the machine translation device or to prepare a more natural and easily understandable sentence. Especially when there are many long sentences or complicated sentences, it takes time and effort to collate with the original sentence, and after that, the editing work often becomes a bottleneck in increasing the efficiency of translation using a machine translation device. From the standpoint of supporting post-editing work of such a translated sentence, many proposals have been made to enhance the understandability of the translated sentence. In particular, JP
In the "machine translation apparatus" described in Japanese Patent Application Laid-Open No. -107769, there is proposed an apparatus that generates an easily understandable translated sentence by eliminating ambiguous modification relations regarding the translation order of a continuous modifier clause, a consecutive modifier clause, and a parallel structure. ing. [0003] However, in the above-mentioned publication, the order reversal of the "section" unit, which causes a great reduction in the ease of understanding and the ease of editing depending on the translation, is described. Not mentioned. [0004] For example, in English-Japanese translation, a subordinate clause is often translated first and a main clause is translated later, as in the following translated example 1. [0005] (1) English: I talked with my wife when my brother called me. He said he was ready. Translation 1 (does not preserve the order of the clauses) When my brother called me, I was my wife I said. He said he was ready. Translation 2 (preserving the order of the clauses) I spoke with my wife. Then my brother called me.
He said he was ready. (2) English: The service of the restaurant was poor because it
That helped him to get the job as a waiter. Translation 1 (Don't preserve the order of the clauses) The cafeteria service was bad due to lack of staff. That helped him get a job as a waiter. Translation 2 (preserving the order of the sections) The cafeteria service was poor. Because there was a shortage of manpower. That helped him get a job as a waiter. (3) English: The judge said that this case was difficultand tha
Judge said that this task was difficult and that he would reconsider his decision. Translation 2 (preserving the order of the verses) The judge said. When this case is difficult. And he will reconsider the decision. In some cases, the translated sentence 2 in which the order of the clauses is preserved is preferable as the output of the machine translation. The reason is that the closer to the order of the original sentences, the easier it is for the user (translator) to understand the correspondence at the time of post-editing. Further, when two pieces of information are included in one sentence and the second piece of information is referred to in the next sentence, it is easier for the user (translator) to translate the original sentence in the order of the original sentence. ,
Similarly, because the reader is easy to understand, it is close to the translation that the user (translator) should complete. Therefore, the amount of correction is also reduced. As a completed translation, it is said that preserving the word order of the original sentence as much as possible is a good translation that makes the flow of the story natural. This is an item common to all translations, not just English-Japanese translations. [0010] However, example above publication is fixed, understandability or editing ease of translation, taking into account the circumstance that varies depending on the preferences and skill of the request level and user postediting I can't say that. Therefore, an object of the machine translation apparatus of the present invention is as follows.
Has developed a machine translation device that can adjust the order of "section" units.
To provide. At that time, save the order of “sections”.
Can be adjusted by user control
And if you want to save the order further,
Various conditions can be controlled by the user . [0012] Means for Solving the Problems The present invention has been made to solve the above problems, the machine translation apparatus of the present invention, the target word
An original sentence input unit for inputting an original sentence of a word, and the original sentence input unit
An original text storage unit for storing the original text to be translated input from
A dictionary section containing knowledge information to be used for translation processing;
Translation of the original text to be translated input using the contents of the book
A translation processing unit for performing processing, and a translation processing unit in the translation processing unit
A translation storage unit for storing translations obtained as a result of
A display unit for displaying the original and translations to be translated, and the display unit
Edits the original or translated text displayed in
And an editing control unit for performing
Word order change that generates a translated sentence that preserves the order of the clauses in the source sentence to be translated
And the word order change unit to reflect the user's specification.
A user instruction input unit for
When translating the original sentence, it was input from the original sentence input unit.
The original sentence to be translated is stored in the original sentence storage unit, and the original sentence input is performed.
The original text to be translated input from is stored in the original text storage unit.
And the input translation is input by the user instruction input unit.
Noun clause,
Whether to preserve the order of each clause, including adjective clauses and adverbial clauses
If you set the clause order to be required to save,
Specify the conditions for what kind of clause order to preserve,
The word order changing unit stores the data in the original text storage unit.
User's finger using the knowledge information of the dictionary unit.
Section based on the storage conditions specified from the
Divide position, and place the original sentence division symbol at the detected position.
The inserted original sentence is inserted by the translation processing unit.
Detects segmentation symbols and translates detected source sentence segmentation symbols
And convert between the source sentence division symbols into one conversion generation unit.
Using the conversion rules of the dictionary unit in the conversion generation process
Converting, further generating a translation using the generation rule, and
Is written in the translation storage unit, and the
The original text to be translated and the translated text stored in the original text storage unit
It is characterized by displaying the translated sentence written in the storage unit . [0013] [0014] [0015] [0016] [0017] In the machine translation apparatus according to the present onset bright [action], enter the original text of the language by the original sentence input section and stored by the original storage unit. Then, the translation processing is executed by the translation processing unit using the contents of the translation dictionary containing the knowledge information used for the translation processing. At the time of this translation process, the word order changing unit divides the data into clause units,
Translate within that range. The result is stored in the translation storage unit and displayed on the display unit. Further, at the time of the translation process, the user designates whether or not to preserve the order of the sections of the original sentence using the user instruction input unit. At that time, do not specify the storage conditions
Two cases, case and designated case, can be selected. Conditions for preserving the order of the above clauses are as follows:
Whether or not the user-specified to save the translation order of the noun clause, is performed using the user instruction input unit. Furthermore, conditions for preserving the order of the above sections
As, or user-specified whether to save the translation order form Yoshibushi is performed using the user instruction input section. Furthermore, conditions for preserving the order of the above sections
As, whether the user-specified save the translation order of the sub-Clause is performed using the user instruction input section. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a machine translation apparatus according to the present invention applied to an English-Japanese translation machine will be described below.
However, the machine translation apparatus of the present invention is applicable to any language pair. Before describing the embodiments, an outline of the machine translation processing unit will be briefly described. FIG. 2 schematically shows the concept of the English-Japanese machine translation process. The contents of each process are as follows. Although a typical example of machine translation is referred to here as a transfer system, the application of the present invention is not limited to the transfer machine translation. (1) In the morphological analysis process, the input sentence is searched in order from the word at the beginning of the sentence while referring to the word dictionary in which the morpheme is stored, and for example, the longest matching method is used. First, processing for dividing into morpheme strings is performed. Then, a process of obtaining syntactic information such as a part of speech and a conversion information such as a translated word for each of the obtained words and further analyzing a tense, a personal name, a number, and the like is performed. (2) In the syntax analysis process, a process of assembling and determining a sentence structure (structure analysis tree), such as the inter-word intercept, is performed based on the part of speech and the variation. At that time, a syntax rule selected from a syntax rule group in the dictionary section is applied, and syntax information obtained from the word dictionary is used. (3) In the conversion process, a process of converting a parse tree of a source language (English) into a parse tree of a target language (Japanese) is performed. At that time, a conversion rule selected from a conversion rule group in the dictionary unit is applied, and conversion information obtained from a word dictionary is used. (4) In the generation processing, the parse tree of the target language is converted into a Japanese character string, and the appropriate particles,
Generate a Japanese sentence by adding an auxiliary verb. At that time, a generation rule selected from a generation rule group in the dictionary unit is applied, and generation information such as a tense obtained by morphological analysis is used. [0028] Figure 1 is a functional block diagram of the English-Japanese machine translation apparatus according to an embodiment of the present invention. Reference numeral 1 denotes an original text input unit for inputting the original text of the target language, and 2 denotes the input unit 1
3 is a dictionary unit that stores knowledge information used for translation processing, and 4 is the original sentence input using the contents of the translation dictionary unit 3. Is a translation processing unit that executes the translation processing of. Reference numeral 5 denotes a translation storage unit for storing translations obtained as a result of the processing by the translation processing unit 4, reference numeral 6 denotes a display unit for displaying the original text and translation, and reference numeral 7 denotes the display unit 6. And 8 is a word order changing unit for generating a translated sentence in which the order of the sections of the original sentence is preserved in order to facilitate understanding and editing, and 9 is a word order changing unit. This is a user instruction input unit for reflecting the user's designation in the change unit 8. FIG. 3 shows a flow of the English-Japanese machine translation process of the present apparatus. Here, the clauses are classified as “noun clauses, adjective clauses, adverbial clauses, and other clauses”. [0031] First, in step S _1, but a text file name to be processed was omitted Ru query the user (here, there is a method of inputting a sentence directly from the keyboard. Also, designated collectively multiple text files also how to there.). Next, in step S _2, inquires the word order conversion specified by the user. In step S _3, performs the translation process. [0032] In step S _4, inquires to the user performing the word order changed again. Returns to step S ₂ if "Y es", the flow proceeds to step S ₅ If "No". In the above-described step S _5, query the user whether to translate other text files. Returns to step S ₁ if "Y es", the process ends if "No". FIG. 4 shows a flow for performing the word order change designation (step S ₂ ) shown in FIG. 3 of the present apparatus. Step S ₆
In, the flag is initialized. Specifically, all
_Flag , nn_flag , aj_flag , av_
Substitute 0 for flag . In step S _7, query the user whether or not to preserve the order of the sections. "Ye
s "If the process goes to step S _8, the end of this process, if" No "(the step S _2). [0034] In step S _8, query the user whether or not to specify the conditions of section storage. The process proceeds to step S ₁₀ if "Yes", the process proceeds to step S ₉ if "No". In step S _9, the All_flag 1
And the process (step S ₂ ) ends. [0035] In the step S _10, query the user whether or not to order preservation of the noun clause. "Yes"
Nara proceeds to the step S _12, the process proceeds to step S ₁₁ if the "No". In step S _11, the Nn_flag 1
By substituting, the process proceeds to step S _12. [0036] In step S _12, query the user whether to order storage adjective clause. "Ye
s "If the process goes to step S _14, if" No "step S
Proceed to ₁₃ . In step S _13, aj_flag
Substituting 1, the process proceeds to step S _14. [0037] In step S _14, query the user whether to order storage adverb clause. The process proceeds to step S ₁₅ if "No", and terminates the process if "Yes" (the step S _2). In step S _15, av_
Substitute 1 for flag , and end the process (step S ₂ ). [0038] FIG. 5 shows a flow for performing translation processing shown in FIG 3 of the apparatus (Step S _3). The original sentence input to the English-Japanese machine translation apparatus of the present embodiment and the translated sentence output from this apparatus are as shown in the following translated example. Hara: I talked with my wife when my brother call
ed me. When my brother called me, I spoke with my wife. Translation 2: I talked to my wife / then my brother called me. The translation 1 is an example of a normal translation result, and the translation 2 is an example of a translation result in which the order of clauses is stored. [0039] In step S _16, performs a morphological analysis process. The dictionary is searched in order from the first word of the sentence, and information such as the part of speech for each obtained word is added. The details of the morphological analysis processing are not essential to the present invention and are therefore omitted, but the resulting data is, for example, as shown in FIG. [0040] In step S _17, performs the parsing process. In step S _18, performs conversion-generating process. FIGS. 7 and 8 show the flow of the syntax analysis processing (step S ₁₇ ) shown in FIG. 5 of the present apparatus. FIG. 9 is an example of a syntax rule used in the above syntax analysis processing. For example, “verb phrase → verb, noun phrase” indicates that “a verb phrase is composed of a verb and a noun phrase”. [0042] In step S _19, to check whether there are applicable rules to parsing rules group. Proceeds to step S ₂₀ is applied if, terminates the absence if parsing process (step S _17). In step S _20, the process proceeds to step S ₂₁ if there is a noun clause on the right side of the applied the above rules, the process proceeds to step S ₂₆ otherwise. [0043] In step S _21, noun clause it is checked whether or not are surrounded by original division symbol. If surrounded proceeds to step S _22, the process proceeds to step S ₂₆ if it is not enclosed. In step S _22, noun clause to check whether there is on the right side of the right-hand side of the rule. The process proceeds to step S ₂₃ if "Y es", the flow proceeds to step S ₂₄ if "No". In step S _23, noun clause to check whether there is on the left side of the right-hand side of the rule. The process proceeds to step S ₂₆ if "Y es", the flow proceeds to step S ₂₅ if "No". [0044] In step S _24, to insert the "it" after the original split symbol after the noun clause. In step S _25, to insert the "it" in front of the front of the original split symbol of the noun clause. In step S _26, the left side of the applied rule proceeds to step S ₂₇ if noun clause, the process proceeds to step S ₂₉ otherwise. [0045] In step S _27, all_flag
Or nn_flag proceeds to S ₂₈ if it is 1, returns to the S ₁₉ otherwise. In step S _28, to insert the original split symbols before and after the noun clause which is configured by applying the rules. If an original sentence division symbol is added before or after a clause on the right side of the rule, the position is moved to before or after the noun clause. In step S _29, the left side of the applied rule proceeds to step S ₃₀ if the adjective clause, the process proceeds to step S ₃₂ otherwise. [0046] In step S _30, all_flag
Or aj_flag proceeds to S ₃₁ if it is 1, returns to the S ₁₉ otherwise. In step S _31, to insert the original split symbols before and after the section adjectives that are configured by applying the rules. If an original sentence segmentation symbol is given before or after the clause on the right side of the rule, move that position to before or after the adjective clause. In step S _32, the left side of the applied rule proceeds to step S ₃₃ if adverb clause, the process proceeds to step S ₃₅ otherwise. [0047] In step S _33, all_flag
Or av_flag proceeds to step S ₃₄ if 1, the flow returns to step S ₁₉ otherwise. Step S ₃₄
In, an original sentence division symbol is inserted before and after a clause constructed by applying the rule. If original split symbols are given to the front and rear sections on the right-hand side of the rule, moving the position before and after the adverb clause. In step S _35, if the left side section of the applied rule proceeds to the step S _36, returns to the step S ₁₉ otherwise. In step _S36 , all_flag
But the process proceeds to step S ₃₇ if it is 1, the flow returns to step S ₁₉ otherwise. In step S _37, to insert the original split symbols before and after the section adverb constructed by applying the rules.
Here, it is assumed that the original sentence division symbol is //. The resulting data is, for example, from an example in which the original sentence division symbol shown in FIG. 10 is not inserted to an example in which the original sentence division symbol shown in FIG. 11 is inserted. Originally, the original sentence division symbol is desirably a symbol that is not used in the original sentence, but it is represented by // for the sake of simplicity. FIG. 12 shows the flow of the conversion / generation processing (step S ₁₈ ) shown in FIG. 5 of the present apparatus. In step S _38, 1 is substituted into a variable n.
In step S _39, and ends larger than the length of the variable n is statements present conversion-generating process (Step S _19), the process proceeds to step S ₄₀ otherwise. [0049] In step S _40, n, n + 1-th character is check whether it is the original division sign, "Y e
If "s", the process proceeds to step _S41 , and if "No", the process proceeds to step S41.
Proceed to ₄₃ . In step S _41, and writes the translation division symbol translation buffer. In step S _42, adding the number of characters of the original split symbol n. If the original sentence division symbols are consecutive, add the total number of characters. Then,
Back to flop S _39. [0050] In step S _43, the n-th character to the next original division sign (if there is no split symbol to the end of the sentence) and the conversion generating units. In step S _44, the conversion-generating unit checks whether the adjective clause,
The process proceeds to step S ₄₅ if "Y es", the flow proceeds to step S ₄₆ if "No". In step S _45, performs replacement as shown in FIG. 13 for the appropriate relative pronouns, relative adverbs, so Kudaseru translated in conjunction manner. [0051] In step S _46, performs conversion generation process. Although the conversion generation processing itself uses a conventional technique, even an incomplete sentence must be able to obtain a result without interrupting the processing. In step _S47 , the conversion generation result is written into the translation buffer. In step S _48, adding the length of the original text that is processed n. The contents of the translated sentence buffer are, for example, the translated sentence 1 and the translated sentence 2 described above.
Originally, the translation segmentation symbol is desirably a symbol that is not used in the translation, for the sake of distinction. As another embodiment, (1) In the above embodiment, steps S ₂₀ to S ₂₀ are applied to a noun clause.
S _25, were subjected to special processing of steps S ₄₄ to S ₄₅ relative adjective clause, if whether the special processing user-selectable, can provide a more user-friendly system. (2) The present invention is particularly long sentence, because it is effective for complex sentences, applied to only the long sentence, sentence number of nodes is large
If it is possible to specify the user so that it can be performed, the system can be applied as a user-friendly system. (3) A user-friendly system can be provided if the type of a delimiter (such as “,” or “;”) that separates translation units can be specified by the user in advance. (4) If the user can pre-register words and collocations that separate translation units, a system such as user friendliness can be provided. [0054] (5) 1 3 at the indicated such relative pronoun in adjective clause, a replacement for the relative adverbs, can provide a more user-friendly system if the user so as to be registered and corrected. (6) A more efficient system can be provided if the user can specify which sentence to apply by inserting a preedit symbol in advance. (7) If the user can specify which part is a node by inserting a pre-edit symbol in advance, a more efficient system can be provided. (8) For Japanese , a reading point or a connecting particle can be used as a delimiter of a translation unit. The machine translation apparatus according to the present invention has the above-described configuration. (1) In this machine translation apparatus capable of adjusting the order of “sections”, a translated sentence close to the order of the original sentence Is output, the understanding is facilitated, and the amount of post-correction can be reduced. (2) In this machine translation apparatus in which the order of “sections” can be controlled by the user, the order of the sections can be customized for each user, for each text, and for each type of section.
It is possible to improve the editing efficiency after users familiar with Ru Do <br/>. (3) In this machine translator that can adjust the translation order of noun clauses ,
Regarding the noun clause, by outputting a translated sentence close to the order of the original sentence, it is easy to understand, and the amount of post-correction can be reduced. [0058] (4) in an adjustable book machine translation apparatus translation order of adverbial clause regard adverb clause, to facilitate understanding by outputting translated text closer to the order of the original, yet, is possible to reduce the post-correction amount It becomes possible.

It is a functional Bed <br/> lock view of the English-Japanese machine translation apparatus according to an embodiment of the BRIEF DESCRIPTION OF THE DRAWINGS [Figure 1] present invention. FIG. 2 is a diagram schematically illustrating the concept of English-Japanese machine translation processing. FIG. 3 is a diagram showing a flow of English-Japanese machine translation processing of the present apparatus. FIG. 4 is a diagram showing a flow of a word order change processing designation (S ₂ ) in the present apparatus. FIG. 5 is a diagram showing a flow of a translation process (S ₃ ) in the present apparatus. FIG. 6 is a diagram showing an example of data as a result of a morphological analysis process (S ₁₆ ) in the present apparatus. FIG. 7 is a diagram showing a flow of a syntax analysis process (S ₁₇ ) in the present apparatus. FIG. 8 is a diagram showing a flow of a syntax analysis process (S ₁₇ ) in the present apparatus. Is a diagram illustrating an example of the syntax rules used in [9] syntax analysis processing in this apparatus (S _17). FIG. 10 is a diagram illustrating an example of data in which a division symbol is not inserted as a result of the syntax analysis processing (S ₁₇ ) in the present device. FIG. 11 is a diagram showing an example of data into which a division symbol resulting from the syntax analysis processing (S ₁₇ ) in the present apparatus is inserted. FIG. 12 is a diagram showing a flow of a conversion / generation process (S ₁₈ ) in the present apparatus. FIG. 13 is a diagram showing an example of replacement of a relative pronoun and a relative adverb in a morphological clause by the present device. [Description of Signs] 1 Original sentence input unit 2 Original sentence storage unit 3 Dictionary unit 4 Translation processing unit 5 Translated sentence storage unit 6 Display unit 7 Edit control unit 8 Word order change unit 9 User instruction input unit

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G06F 17/27-17/28

Claims (1)

(57) [Claims] [Claim 1] An original sentence input section for inputting an original sentence of a target language, an original sentence storage section for storing a translation target original sentence inputted from the original sentence input section, A dictionary unit that contains knowledge information to be used, a translation processing unit that executes a translation process of the original text to be translated input using the contents of the dictionary unit, and a translation processing result obtained by the translation processing unit. A machine including: a translation storage unit that stores a translation; a display unit that displays the translation target original and the translation; and an edit control unit that performs editing processing of the translation target original or the translation displayed on the display unit.
In the translation apparatus, word order to produce a translation that preserves the order of the sections to be translated original text
And a user unit for reflecting the user's designation in the word order changing unit.
An input unit for translating the original sentence to be translated, wherein the original sentence to be translated inputted from the original sentence input unit is translated into the original sentence.
Stored in a storage unit, and the input translation pair is stored by the user instruction input unit.
Noun clauses and adjectives for elephant sentences according to user instructions
Specify whether to preserve the order of each clause, including lyric clauses and adverbial clauses.
If the clause order is set to require storage,
Specify the condition of preserving the order of the clauses of the class, and the word order change unit stores the condition in the original text storage unit.
User's finger using the knowledge information of the dictionary unit.
Section based on the storage conditions specified from the
Divide position, and place the original sentence division symbol at the detected position.
Inserted by the translation processing unit, the inserted original sentence division symbol
Is detected, and the detected original sentence segmentation symbol is replaced with the translated sentence segmentation symbol.
That is, the space between the original text division symbols is converted as one conversion generation unit.
In the conversion generation process, conversion is performed using the conversion rule of the dictionary unit,
Further, a translation is generated using the generation rule, and the translation is
Written in a sentence storage unit and stored in the original sentence storage unit by the display unit
The translation source text and the translation written in the translation storage unit are displayed.
Machine translation apparatus, characterized in that it shows.