JP2008015844A - Mechanical translation device, mechanical translation method, generation rule forming device, generation rule forming method, and programs and recording media therefor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a mechanical translation technique capable of improving translation accuracy. <P>SOLUTION: The mechanical translation device 2 comprises a generation rule searching means 241 for searching a generation rule to a partial hypothesis from a rule table 114 storing generation rules restricted so that the symbol string of a translation destination language of the right side of a generation rule in synchronous context free grammar starts from a terminal symbol; a generation rule with word range generation means 242 for adding a word range showing the range of an input sentence covered by non-terminal symbols of the generation rule to the generation rule; a partial hypothesis score calculation means 243 for forming a translated word and a word range contained in the generation rule with word range as a new partial hypothesis, and successively extending the new partial hypothesis in the arrangement order of non-terminal symbols on the translation destination language side of the generation rule from the beginning to the end of the sentence by applying the generation rule from top to down to calculate partial hypothesis scores; and a hypothesis search means 244 for searching a hypothesis with maximum partial hypothesis score of applicable hypotheses. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a machine translation device, a machine translation method, a production rule creation device, a production rule creation method, a program thereof, and a recording medium.

Conventionally, techniques (statistical machine translation) for realizing machine translation using a statistical model are known (for example, Non-Patent Documents 1 to 4).
Statistical machine translation is formulated as a problem of searching for a target language word string (sentence) that maximizes the translation probability when a source language word string (sentence) is given. Here, when the translation probability is expressed by a logarithmic linear model, statistical machine translation is finally formulated by Equation (1).

Here, an identification code of integers “1” to “J” is assigned to each word position in the word string of the translation source language. That is, the word string in the source language is composed of “J” words. When the translation source language is Japanese, the translation source language word string is composed of “J” independent words, attached words, and punctuation marks (.,). Also, in the word string of the translation destination language, identification codes of integers “1” to “I” are assigned to the respective word positions. That is, the word string in the translation target language is composed of “I” words. M (1 ≦ m ≦ M) represents an integer for identifying a feature function, and M represents the number of feature functions. Each feature function represents a certainty as a translation, a certain one as a translation target language, or the like. The word string set E represents all word string sets that can be generated by any combination of words in the translation target language. Each feature weight λ _m is set such that the translation accuracy value in the feature weight learning parallel translation corpus is maximized by using an error minimization learning method or the like (see, for example, Non-Patent Document 1).

For translation from a source language sentence to a destination language sentence, a weighted synchronous context-free grammar is used to determine whether the source language sentence and the destination language sentence are A method for modeling the association is known (for example, see Non-Patent Document 2).
The weighted synchronization context free grammar is composed of a set of weighted generation rules shown in Expression (2).

  Here, X on the left side of the arrow indicates a non-terminal symbol. Γ on the right side of the arrow is a symbol string composed of terminal symbols or non-terminal symbols and corresponds to the source language. Α is a symbol string composed of a terminal symbol or a non-terminal symbol, and corresponds to the language to be translated. “˜” represents a one-to-one correspondence between a non-terminal symbol included in the symbol string γ and a non-terminal symbol included in the symbol string α. Here, it is assumed that the number of non-terminal symbols included in the symbol string γ is the same as the number of non-terminal symbols included in the symbol string α.

Table 1 shows a specific example of the generation rule shown in Expression (2). Here, X _(k) is a non-terminal symbol, and k (k = 1, 2,...) Indicates the correspondence between the arrangement of the non-terminal symbol in the source language and the arrangement of the non-terminal symbol in the target language. ing.

  As a special form of the method described in Non-Patent Document 2, a method using only a generation rule of the form shown in Expression (3) is also known (see Non-Patent Document 3). The method described in Non-Patent Document 3 performs statistical machine translation in phrase units, and is a translation method based on phrase pairs. In this method, when generating a synchronous context free grammar from a set of extracted phrase pairs, only a generation rule of the form shown in Expression (3) is used. When performing a solution search depending on the form of this generation rule, an efficient beam search technique can be used (see Non-Patent Document 4). In the method described in Non-Patent Document 4, in the solution search, the translated word range is expressed by a bit string in which the position of the translated word is marked and held in the memory.

  In the conventional modeling using the weighted synchronous context free grammar, the derivation D of the weighted synchronous context free grammar is used to convert the word string of the source language and the word string of the target language into f (D) and e, respectively. (D) is described. Here, for example, if the generation rule covering the “j” -th word from the “i” -th word in the word string f (D) of the translation source language is r, the derivation D is a triple <r, i, j>. It is represented by a set of

  In the modeling of Non-Patent Document 2, Formula (4) obtained by modifying Formula (1) formulated from statistical machine translation to a derivation base is used. In this formulation, when a translation source language word string is given, a derivation D ^ that maximizes the linear sum of the product of the feature function and the feature weight is obtained, and the corresponding e (D ^ ) Is the translation result. Here, the symbol “＾ (hat)” indicates a symbol added on the letter “D”, and the symbol “＾ (hat)” is used in the same meaning hereinafter.

Various variations can be considered as to what is used for the value h _m (D) of each feature function shown in Expression (4). For example, the natural logarithm log _e (hereinafter referred to as ln) of the following six function values may be used (for example, see Non-Patent Document 3). The values of these six functions are the translation probabilities P _{e | f} (D) and P _{f | e} (D) shown in Equation (5), and the lexical weight Lex _{e | f} (D), Lex _{f | e} (D), n-gram language model probability P _LM (e (D)), and phrase penalty exp (length (e (D))). Here, length (·) indicates a function that returns the number of words.

The translation probabilities P _{e | f} (D), P _{f | e} (D) and the lexical weights Lex _{e | f} (D), Lex _{f | e} (D) are values for evaluating the likelihood of translation. It is also called a translation model. Specifically, for example, the translation probability P _{e | f} (D) is obtained by using the probability P (α | γ) for each production rule r included in the derivation D as the score for each production rule, as shown in Expression (5). , The scores for each production rule are integrated for all production rules r included in derivation D.

For example, in the method described in Non-Patent Document 2, a solution search in translation is performed according to the following procedure.
First, in bottom-up syntax analysis based on the CKY (Cocke-Kasami-Younger) method, the generation rules on the source language side of the synchronous context free grammar are applied to the source language word strings, and the source language syntax analysis tree Get. Then, an optimum derivation D ^ of the synchronous context free grammar corresponding to the parse tree of the source language is obtained based on the above-described equation (4), and the word string of the translation destination language is obtained based on the optimum derivation D ^. Is generated. However, since there are a large number of solution candidates (hereinafter referred to as hypotheses) in the statistical machine translation solution search, it is virtually impossible to perform a full search to find the true optimal solution from the viewpoint of computational complexity. It has become. Therefore, conventionally, a sub-optimal solution is obtained by performing processing while executing predetermined pruning on the derivation D of the subtree of the synchronous context free grammar partially configured in the solution search process.
Franz Josef Och. Minimum error rate training in statistical machine translation.In Proc.of ACL 2003, p. 160-167, Sapporo, Japan, July 2003 David Chiang. A hierarchical phrase-based model for statistical machine translation. In Proc. Of ACL 2005, p. 263-270, Ann Arbor, Michigan, June 2005 Philipp Koehn, Franz Josef Och, and Daniel Marcu. Statistical phrase-based translation.In Proc. Of NAACL 2003, p. 48-54, Edmonton, Canada, 2003 Philipp Koehn. Pharaoh: A beam search decoder for phrase-based statistical machine translation models.In Proc. Of the 6th Conference of the Association for Machine Translation in the Americas (AMTA), p. L15-124, September-October 2004

  In the solution search of statistical machine translation, in order to realize pruning with few side effects, it is necessary to accurately estimate the likelihood of the hypothesis of the derived D shown in the above equation (4). At that time, in order to improve the translation accuracy, it is important to use the n-gram language model of the translation destination language as one of the feature functions. Therefore, in order to effectively use the score of the n-gram language model as the score of the partial hypothesis that is the basis of the hypothesis, it is desired that the score is generated sequentially from the beginning of the translated language to the end of the sentence. However, in the conventional method, it is not guaranteed that the partial hypotheses are sequentially generated from the beginning of the translated language to the end of the sentence.

  Therefore, an object of the present invention is to provide a machine translation technique that can solve the above-described problems and improve translation accuracy.

  In order to solve the above problem, the machine translation device according to claim 1, wherein the translation destination language on the right side of a generation rule that generates a word string of the translation destination language from a word string of the translation source language using a synchronous context free grammar Using a rule table that stores a plurality of generation rules generated so that the symbol string of a symbol starts with a terminal symbol, the translation target language corresponding to the input that is the translation result of the word string of the input source language is input. A machine translation device that outputs a hypothesis that is a partial hypothesis that is finally generated by sequentially creating a new partial hypothesis from a predetermined partial hypothesis as a word string and extending the predetermined partial hypothesis. A generation rule search means for searching the rule table for a generation rule applicable for extending the predetermined partial hypothesis, and a translation source language side of the searched generation rule A word range indicating a range of the word string of the source language covered by the non-terminal symbol of the searched generation rule based on the number of words and the word position of the words constituting the input word string of the source language. In addition, a generation rule generation unit with a word range for generating each applicable generation rule with a word range, a translated word on the translation target language side included in the generated generation rule with a word range, and the word range As the new partial hypothesis, and the applicable generation rule is applied in a top-down manner, and the non-terminal symbols on the translation target language side are arranged from the beginning to the end of the sentence in the applicable generation rule. Then, the new partial hypothesis is sequentially expanded to obtain a translation model score representing the likelihood of translation and a word string of the translation target language, which are obtained in advance for each of the generation rules. A partial hypothesis score calculating means for calculating a partial hypothesis score indicating an evaluation value of the created partial hypothesis based on a score of a language model representing the certainty of the input, and a word string of the input source language Search for a predetermined partial hypothesis that can be applied and expand the predetermined partial hypothesis to search for a partial hypothesis finally having the maximum partial hypothesis score as the hypothesis. And a hypothesis search means.

  Further, in the machine translation method according to claim 2, the symbol string of the translation destination language on the right side of the generation rule for generating the translation target language word string from the translation source language word string using the synchronous context free grammar is terminated. Using a rule table storing a plurality of generation rules generated so as to start with a symbol, a translation target language word string corresponding to the input, which is a translation result of the input source language word string, is predetermined. This is a machine translation method of a machine translation device that outputs a hypothesis that is a partial hypothesis finally generated by sequentially creating a new partial hypothesis from the partial hypotheses and extending the predetermined partial hypothesis. Then, the machine translation device searches the rule table for a generation rule applicable for extending the predetermined partial hypothesis, respectively, and a search for the searched generation rule. The source language word covered by the non-terminal symbol of the searched generation rule based on the number of words constituting the input source language word string and the word position on the source language side A generation rule generation step with a word range that adds a word range indicating the range of the column and generates an applicable generation rule with a word range, and a translation-side language side included in the generated generation rule with a word range The translated word and the word range are created as the new partial hypothesis, and the applicable generation rule is applied top-down, and the non-terminal symbol on the translation destination language side is applied from the beginning of the sentence in the applicable generation rule The new partial hypotheses are sequentially expanded in the order in which they are arranged at the end of the sentence, and a translation model score representing the certainty of the translation obtained in advance for each of the generation rules, and the previous A partial hypothesis score calculating step for calculating a partial hypothesis score indicating an evaluation value of the created partial hypothesis based on a score of a language model representing the probability as a word string of a translation target language; and the input translation Search for a predetermined partial hypothesis applicable to a word string in the original language, and among the partial hypotheses finally generated by extending the predetermined partial hypothesis, a portion having the maximum partial hypothesis score A hypothesis searching step for searching for a hypothesis as the hypothesis.

  According to the machine translation device according to claim 1 or the machine translation method according to claim 2, since the symbol string of the translation target language on the right side of the generation rule starts with a terminal symbol, the machine translation device That is, a partial hypothesis starting from a translated word can be generated. Therefore, the machine translation device expands the generation rules from the top down so that the partial hypotheses are sequentially expanded with respect to the untranslated portion in the word string of the input source language, and the non-terminal symbol of the destination language side is expanded. By creating partial hypotheses in the order in which they are arranged, hypotheses as candidates for translation target language word strings corresponding to input can be sequentially generated from the beginning of the translation destination language to the end of the sentence. Further, the machine translation device calculates a partial hypothesis score that is an evaluation value of each partial hypothesis based on the translation model and language model for each generation rule, and searches for a partial hypothesis having the maximum partial hypothesis score as a hypothesis. Therefore, the translation accuracy of the word string in the translation destination language can be improved. In addition, since the machine translation apparatus sequentially generates partial hypotheses from the beginning of the translation target language to the end of the sentence, it is easy to read even a relatively long n-gram language model. Therefore, by adopting a relatively long n-gram language model, it becomes possible to improve the translation performance accordingly. Furthermore, since partial hypotheses are generated sequentially from the beginning of the translation target language to the end of the sentence, it is possible to effectively execute pruning by setting a predetermined threshold value and reduce the solution search space. Become.

  A production rule creation device according to claim 3 is a production rule creation device for creating a production rule used in the machine translation device according to claim 1, wherein a word in the translation source language and a word in the translation destination language From the combination stored in the bilingual corpus having a plurality of combinations of the source language word sequence and the translation destination language word sequence having the same meaning, the translation source language and the translation destination language A phrase pair extraction means for extracting a combination of words or phrases having the same meaning as a phrase pair, and a symbol string of the translation target language on the right side in the generation rule of the synchronous context free grammar based on the extracted phrase pair Generation rule creating means for creating a production rule to which a restriction is added that starts with a terminal symbol.

  A production rule creation method according to claim 4 is a production rule creation method of a production rule creation device for creating a production rule used in the machine translation device according to claim 1, wherein the production rule creation device includes: The translation corpus includes a plurality of combinations of a source language word sequence and a target language word sequence having the same meaning based on a word correspondence between a source language word and a target language word. A phrase pair extraction step for extracting a combination of words or phrases having the same meaning in the translation source language and the translation destination language as a phrase pair from a combination; and generation of a synchronous context free grammar based on the extracted phrase pair And a generation rule creating step for creating a generation rule to which a restriction that the symbol string of the translation target language on the right side starts with a terminal symbol is added. The features.

  According to the production rule creation device according to claim 3 or the production rule creation method according to claim 4, the production rule creation device generates a symbol string of the translation destination language on the right side with respect to the production rule of the synchronization context free grammar. Create a production rule with the constraint that starts with a terminal symbol. In this way, by constraining the destination language side in the generation rule, when searching for hypotheses using the generated generation rule, it is possible to construct a syntax tree that prioritizes the destination language over the source language. It becomes easy to do.

  The machine translation program according to claim 5 causes a computer to execute the machine translation method according to claim 2. By being configured in this way, a computer in which this program is installed can realize each function based on this program.

  According to a sixth aspect of the present invention, there is provided a production rule creation program that causes a computer to execute the production rule creation method according to the fourth aspect. By being configured in this way, a computer in which this program is installed can realize each function based on this program.

  A computer-readable recording medium according to claim 7 is recorded with the machine translation program according to claim 5 or the generation rule creation program according to claim 6. By being configured in this way, a computer equipped with this recording medium can realize each function based on a program recorded on this recording medium.

  According to the present invention, translation accuracy can be improved.

  The best mode for carrying out a machine translation device and a machine translation method, and a production rule creation device and a production rule creation method according to the present invention will be described in detail below with reference to the drawings. . Hereinafter, the production rule creation device and the production rule creation method, and the machine translation device and the machine translation method will be described in order.

[Configuration of generation rule creation device]
FIG. 1 is a functional block diagram showing a configuration of a production rule creation device according to an embodiment of the present invention.
The generation rule creation device 1 creates a generation rule used in a machine translation device that mechanically translates a word string in a translation source language into a word string in a translation destination language. In the following description, it is assumed that the source language is Japanese and the target language is English.
The generation rule creation device 1 is composed of, for example, a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), an HDD (Hard Disk Drive), an input / output interface, and the like. As shown in FIG. 1, input / output means 10, storage means 11, word correspondence creation module 12, and control means 13 are provided.

The input / output means 10 inputs the bilingual corpus 150 to the word correspondence creation module 12 and outputs the result of the arithmetic processing, data stored in the storage means 11, and the like to the output device D. The input / output unit 10 inputs a predetermined command (such as a mode selection command or an operation command) from the input device K to the control unit 13. In the present embodiment, the mode selection command includes a command for selecting the “word correspondence creation” mode and a command for selecting the “rule table creation” mode.
The bilingual corpus 150 includes a plurality of data of combinations of translation source language word strings and translation destination language word strings having the same meaning.

  The storage unit 11 includes, for example, a RAM used for arithmetic processing by the CPU, and a ROM and HDD for storing predetermined programs, various databases, processing results, and the like. For example, the storage unit 11 stores a word correspondence 111, a phrase pair 112, a rule 113, and a rule table 114 as processing results.

  The word correspondence creation module 12 uses a statistic about word co-occurrence obtained from the bilingual corpus 150 when a command for selecting the “word correspondence creation” mode is input, and provides many-to-many (translation source language or translation destination). The word correspondence 111 (including that the language word does not correspond anywhere) is automatically obtained. In order to obtain the many-to-many word correspondence 111, the word correspondence creation module 12 uses, for example, a word-by-word translation model to obtain the optimum one-to-many correspondence and many-to-one correspondence for the entire sentence. Combine both. As an example of the combination method, there is a method of using an intersection of a one-to-many correspondence and a many-to-one correspondence, and adding a neighboring element of a one-to-many correspondence and a many-to-one correspondence (non-patent document). 3). For the translation model for each word, see “Peter F. Brown, Stephen A. Della Pietra, Vincent J. Della Pietra, and Robert L. Mercer. The mathematics of statistical machine translation: Parameter estimation. Computatinal Linguistics, 19 ): 263-311, 1993 ". In addition, when the bilingual corpus 150 itself holds information about the word correspondence 111, the word correspondence creation module 12 may be omitted.

FIG. 2 shows an example of word correspondence of Japanese-English parallel translation. Eight words and black circles indicating punctuation points (periods) correspond to words.
“The Federal Constitutional Court decides on the issue of unconstitution.”
"The Federal Constitutional Court decides on the question of unconstitutionality."

  The control unit 13 includes a mode determination unit 131, a phrase pair extraction unit 132, a generation rule creation unit 133, and a translation score calculation unit 134.

  The mode determination unit 131 determines the mode indicated by the mode selection command input from the input device K via the input / output unit 10. The mode determination unit 131 instructs the word correspondence creation module 12 to input the bilingual corpus 150 when the mode is the “word correspondence creation” mode. Further, when the mode is the “rule table creation” mode, the mode determination unit 131 instructs the phrase pair extraction unit 132 to input the bilingual corpus 150.

  The phrase pair extraction unit 132 obtains a combination of words or phrases having the same meaning in the source language and the target language from the parallel corpus 150 based on the word correspondence 111 between the source language word and the target language word. The phrase pair 112 is extracted. The extracted phrase pair 112 is stored in the storage unit 11. When a word correspondence a is calculated with respect to the parallel translation of the word string in the translation source language and the word string in the translation destination language shown in Expression (1), the phrase pair extraction unit 132 calculates Expression (7). Extract the indicated phrase pair. Here, the word correspondence a is a set of a set of a word position in the translation destination language and a word position in the translation source language. “I, m, j, n” in equation (7) represents an integer, and satisfies the relationship of equation (8) with the word correspondence a.

  For example, the following phrase pairs are extracted from the word correspondence of the parallel translation shown in FIG.

The generation rule creation means 133 is based on the list of phrase pairs stored in the phrase pair 112 extracted from the parallel translation sentence pair for each parallel translation sentence pair (combination of parallel translation sentences) in the parallel translation corpus 150. A rule for generating a synchronous context free grammar is created and stored in the rule 113. Here, in the generation rule to be created, there is a restriction that the translation destination language side must start with a terminal symbol. In addition, the generation rules stored in the rule 113 allow duplication. In the generation rule created in this way, the following glue rule used in Non-Patent Document 2 is not used. Note that the glue rule refers to the following two rules when S is a start symbol and X is a non-terminal symbol.
S → <S ₍₁₎ X ₍₁₎ , S ₍₁₎ X ₍₁₎ >
S → <X ₍₁₎ , X ₍₁₎ >

に対応して、式（９）〜式（９ｄ）の生成規則を生成する。このうち、式（９ａ）〜式（９ｄ）の生成規則は、式（９）の生成規則から自動的に生成することができる。また、式（９ａ）〜式（９ｄ）の生成規則に付与されるスコアについても、式（９）の生成規則と同一値を用いることができる。このような理由から、実装上は、式（９ａ）〜式（９ｄ）の生成規則は明示的にストレージに格納する必要がない。式（９ａ）〜式（９ｄ）の生成規則は、非特許文献２で用いられるグルー規則の非終端記号Ｘを、Ｘを左辺とする個々の規則で１回書き換えたものに対応している。 The generation rule creation means 133 is a phrase pair

Corresponding to the above, generation rules of Expression (9) to Expression (9d) are generated. Among these, the production | generation rule of Formula (9a)-Formula (9d) can be automatically produced | generated from the production | generation rule of Formula (9). Moreover, the same value as the production | generation rule of Formula (9) can be used also about the score provided to the production | generation rule of Formula (9a)-Formula (9d). For this reason, it is not necessary to explicitly store the generation rules of the expressions (9a) to (9d) in the storage for implementation. The generation rules of the equations (9a) to (9d) correspond to the non-terminal symbol X of the glue rule used in Non-Patent Document 2 rewritten once with each rule having X as the left side.

  The generation rule creation unit 133 generates the generation rule represented by Expression (10) when the generation rule represented by Expression (10) is generated and the phrase pair on the right side of Expression (9) satisfies the relationship of Expression (11). Is generated.

  Here, α has a restriction that it must start with a terminal symbol. Although not essential to the present embodiment, the generation rule creation unit 133 also employs the following restrictions, for example. First, both γ and α must contain at least one terminal symbol. Second, production rules can have a maximum of two nonterminal symbols. However, nonterminal symbols must not be adjacent in the source language.

The translation score calculation means 134 counts the generation rules stored in the rule 113 with duplication allowed, the translation probabilities P _{e | f} (r) and P _{f | e} (r) corresponding to each generation rule r, and the lexical Weights Lex _{e | f} (r) and Lex _{f | e} (r) are calculated and stored in the rule table 114 in association with each generation rule r. FIG. 3 shows an example of the rule table. Table 3 shows calculation formulas for each translation probability and lexical weight corresponding to the generation rule r. This score calculation is based on Non-Patent Document 2.

[Concrete example]
Specifically, the generation rule creation unit 133 generates a rule (only the right side is shown) as shown in Table 4 from the parallel translation shown in FIG. In the example of Table 4, the translation destination language side on the right side of each production rule always starts with a terminal symbol (word). On the other hand, in the conventional method, the translation destination language side on the right side of the generation rule does not necessarily start with a terminal symbol. For example, as shown in Table 1, it may start with a non-terminal symbol X.

  Note that the mode determination unit 131, the phrase pair extraction unit 132, the generation rule creation unit 133, and the translation score calculation unit 134 have the CPU expand a predetermined program stored in the HDD or the like of the storage unit 11 in the RAM. It is realized by executing.

[Operation of generation rule creation device]
The operation of the production rule creation device shown in FIG. 1 will be described with reference to FIG. 4 (see FIG. 1 as appropriate). FIG. 4 is a flowchart showing the operation of the generation rule creation device shown in FIG.
The production rule creation device 1 determines the mode by the mode determination means 131 (step S1). As a result of the determination, if the mode is the “word correspondence creation” mode, the generation rule creation device 1 inputs the bilingual corpus 150 to the word correspondence creation module 12 via the input / output means 10 (step S2). A word correspondence is created by the correspondence creating module 12 (step S3). The created word correspondence 111 is stored in the storage unit 11.

  On the other hand, if the result of determination in step S1 is that the mode is “rule table creation” mode, the production rule creation device 1 inputs the bilingual corpus 150 to the control means 13 via the input / output means 10 (step S4) The phrase pair extraction unit 132 extracts a phrase pair from the parallel corpus 150 (step S5: phrase pair extraction step). The extracted phrase pair 112 is stored in the storage unit 11.

  Subsequently, the production rule creation device 1 adds a restriction that the symbol string of the translation target language on the right side of the production rule of the synchronous context free grammar starts from the terminal symbol based on the phrase pair 112 by the production rule creation unit 133. A generation rule is created (step S6: generation rule creation step). The created rule 113 is stored in the storage unit 11. And the production | generation rule preparation apparatus 1 matches each translation score calculated from each production | generation rule of the rule 113 with each production | generation rule by the translation score calculation means 134 (step S7). The associated generation rule and translation score are stored in the storage unit 11 as the rule table 114.

  The production rule creation device 1 can also be realized by executing a production rule creation program that causes a general computer to execute each of the steps described above. This program can be distributed via a communication line, or can be written on a recording medium such as a CD-ROM for distribution.

  According to the production rule creation device 1 of the present embodiment, the number of production rules can be greatly reduced as compared with the conventional case. For example, in the case of word correspondence of the parallel translation sentence shown in FIG. 2, there are actually 71 extracted phrase pairs in total. Even when the number of non-terminal symbols of each production rule is limited to “2”, the conventional method creates 533 production rules based on this phrase pair (except for the glue rule). On the other hand, in the present embodiment, since the generation rule is created by imposing restrictions, the number of generation rules can be reduced to 110. As a result, when searching for a translation solution using the generation rule of this embodiment, the solution search space can be significantly reduced as compared with the conventional method.

[Configuration of machine translation device]
FIG. 5 is a functional block diagram showing the configuration of the machine translation apparatus according to the embodiment of the present invention.
The machine translation device 2 uses the rule table created by the generation rule creation device 1 (see FIG. 1) to convert the input source language word string into the translation destination language word string corresponding to the input. It is intended to translate. The machine translation apparatus 2 includes, for example, a CPU, a RAM, a ROM, an HDD, an input / output interface, and the like. As shown in FIG. 5, the input / output means 20, the storage means 21, and a feature weight learning module. 22, a word information extraction module 23, and a control means 24.

The input / output means 20 inputs a translation source language word string from the input device K to the control means 24, and outputs a translation destination language word string as a translation result to the output device D from the control means 24. It is. The input / output unit 20 inputs the feature weight learning parallel corpus 250 to the feature weight learning module 22.
The feature weight learning bilingual corpus 250 is prepared separately from the bilingual corpus 150 used when the generation rule creation device 1 (see FIG. 1) creates a generation rule.

  The storage unit 21 includes, for example, a RAM used for arithmetic processing by the CPU, and a ROM and HDD for storing predetermined programs, various databases, processing results, and the like. For example, the storage means 21 stores a feature weight 211, word information 212, a rule with word range 213, a partial hypothesis 214, and a partial hypothesis score 216 as processing results. The storage unit 11 stores a rule table 114 created by the production rule creation device 1 (see FIG. 1) and a language model 215 created separately in advance. The language model 215 stores an n-gram representing the certainty as a translation destination language. This n-gram is separately learned from an enormous volume of translated corpora.

  The feature weight learning module 22 learns the weight corresponding to the value of each feature function based on the feature weight learning parallel translation corpus 250, the rule table 114, and the language model 215, and stores the learning result as the feature weight 211. The information is stored in the means 21.

  The word information extraction module 23 divides the sentence of the translation source language input from the input device K through the input / output means 20 into words, and obtains information (word information) about the words constituting the sentence of the translation source language. To extract. The word information includes, for example, a word string, a word position, the number of words, and the like. The extracted word information 212 is stored in the storage means 21. When the sentence of the translation source language input from the input device K has already been divided into words, the word information extraction module 23 can be omitted.

  The control unit 24 obtains a hypothesis that covers the entire sentence while expanding the partial hypothesis in the procedure described later, and obtains an optimum (actually suboptimal) one of them. The generation rule search unit 241 and the word range The generation rule generation unit 242 includes a partial hypothesis score calculation unit 243, a partial hypothesis score calculation unit 243, and a hypothesis search unit 244.

  The generation rule search means 241 searches the rule table 114 for a generation rule applicable for extending a predetermined partial hypothesis.

  The generation rule generation unit with word range 242 searches the generation rule searched by the generation rule search unit 241 based on the number of words and the word position of the words constituting the input word string of the translation source language. A word range indicating the range of the word string of the translation source language covered by the non-terminal symbol of the generated generation rule is added to generate applicable generation rules with word ranges.

Here, the conversion target part (the range covered by the non-terminal symbol) of the input source language word string is expressed as [l, r] at the left end (left) and right end (right) of the word position. I decided to. In the initial stage, for example, if the input is 11 words, [l, r] = [1,11]. In this case, the word range (untranslated word range) of each non-terminal symbol on the source language side of the generation rule is expressed as [l ₁ , r ₁ ], [l ₂ , r ₂ ],. For example, [l ₁ , r ₁ ] = [1, 2]. If there are two non-terminal symbols on the translation source language side of a certain generation rule, two word ranges are also set. In addition, there are a plurality of possibilities as the value of “l ₁ ” or “r ₁ ” for a non-terminal symbol on the translation source language side of a generation rule.

  The partial hypothesis score calculation means 243 creates the translated word and the word range on the translation target language side included in the generation rule with word range as a new partial hypothesis, and applies the applicable generation rule from the top down. In addition, the new partial hypotheses are expanded in the order in which the non-terminal symbols on the translation target language side are arranged from the beginning of the sentence to the end of the sentence in the applicable production rules, and the certainty of translation obtained in advance for each production rule. The partial hypothesis score indicating the evaluation value of the created partial hypothesis H ′ is calculated based on the score of the translation model to be expressed and the score of the language model indicating the certainty of the translated language as a word string.

  Specifically, the partial hypothesis score calculation means 243 includes a combination of “a word string from the beginning of a sentence in the translation target language” and “a stack that holds a range of untranslated words in the word string in the translation source language”. Based on the partial hypothesis H, a new partial hypothesis H ′ is generated using the applicable generation rule with word range. Creating H ′ based on H is called extension of the partial hypothesis. As the partial hypothesis is expanded, the word string from the beginning of the translated language in the partial hypothesis is added sequentially from the beginning of the sentence to the end of the sentence. Further, the partial hypothesis score calculating means 243 saves memory. Share word strings with other partial hypotheses, and if the number of words in the source language is J, the initial partial hypothesis (initial value of the partial hypothesis) is “empty string” and “[1, J] is the only stack.

Partial hypothesis score calculating means 243, and the number of words that are in the source language portion hypothesis H 'translated and m (0 ≦ m ≦ J) , the priority queue Q _0, Q _1, ..., a Q _J Enter a partial hypothesis H ′. That is, the partial hypothesis score calculation means 243 starts with Q ₀ = {initial partial hypothesis} and expands the partial hypothesis stored in the priority-added queue Q _m in synchronization with the number of translated words in the translation source language. .

When expanding a partial hypothesis, the partial hypothesis score calculation means 243 pops the untranslated word range [l, r] from the top of the stack (from the top of the stack). Generation rules corresponding to input sentences in the source language are managed using the Earley chart structure. By using the chart structure, it is possible to efficiently find a generation rule applicable to the partial hypothesis. For example, when the partial hypothesis score calculation means 243 extracts a partial hypothesis from the generation rule with a word range as shown in the equation (13) generated from the equation (9d), the word range [l ₂ , r ₂ ], [l ₁ , r ₁ ] are pushed onto the stack in order, so that [l ₁ , r ₁ ] is processed first. In this way, the translation target language side is guaranteed to always generate a translation from the beginning of the sentence.

  The partial hypothesis score calculation means 243 adds the difference score to the score S (H) of the original partial hypothesis H, thereby increasing the score S of the expanded partial hypothesis H ′ as shown in the equation (14). (H ′) is calculated.

と表記する。ここで、生成された翻訳先言語の単語列は、「ｉ」番目の単語から「ｉ＋ε」番目の単語で構成されている。 Here, m (1 ≦ m ≦ M) indicates an identification number of the value of the feature function. In the present embodiment, the partial hypothesis score calculation means 243 sets M = 6, the values h _m (H ′) of the six feature functions shown in Table 5, and the feature weights corresponding to the h _m (H ′). λ _m (H ′) is used. Also, when generating a hypothesis, the word string of the target language that was newly generated by expanding the partial hypothesis,

Is written. Here, the generated word string of the translation destination language is composed of the “i + ε” -th word from the “i” -th word.

Here, P _{e | f} (r), P _{f | e} (r), Lex _{e | f} (r), and Lex _{f | e} (r) are generation rules defined in the rule table 114 of the storage unit 21. This is the value of the feature function corresponding to r. The feature weight λ _m is defined in advance in the feature weight 211 stored in the storage unit 21.

That is, in this embodiment, the partial hypothesis score calculation means 243 finally calculates the score S (H ′) of the partial hypothesis H ′ from the equation (14) and Table 5 using the equation (15). The right side of the second equal sign in equation (15) indicates that the difference score is added to the score S (H) of the original partial hypothesis H. Further, the partial hypothesis score calculation means 243 holds only the maximum Y (for example, 1000) partial hypotheses in the priority queues Q ₀ , Q ₁ ,..., Q _J. The partial hypothesis score calculation unit 243, partial hypothesis score S (H ') is the priority queue Q _0, Q _1, ..., if less than the product of a constant with the largest part hypothesis score in Q _J Discard the partial hypothesis H ′. Thereby, the partial hypothesis score calculation means 243 can perform effective pruning on each partial hypothesis stored in the priority-added queues Q ₀ , Q ₁ ,..., Q _J.

The hypothesis search means 244 searches for a predetermined partial hypothesis that can be applied to the input word string of the source language, and expands the predetermined partial hypothesis so that the entire sentence of the source language finally generated Among the partial hypotheses corresponding to (referred to as a hypothesis), a hypothesis having the maximum partial hypothesis score is searched. Specifically, the hypothesis searching means 244 searches the hypothesis H ^ having the maximum hypothesis score from the relationship of the equation (16) among a predetermined number of hypotheses obtained from the entire sentence of the translation source language. Since the hypothesis H shown in the equation (16) is the partial hypothesis H included in the priority queue Q _J , the hypothesis H ^ obtained in the equation (16) maximizes the value of the partial hypothesis score S (H). This can be realized by obtaining a partial hypothesis H (that is, hypothesis H).

  Further, the hypothesis searching means 244 outputs a word string from the beginning of the translation language corresponding to the obtained hypothesis to the end of the sentence as a translation result to the output device D via the input / output means 20.

  It should be noted that the generation rule search unit 241, the word range added generation rule generation unit 242, the partial hypothesis score calculation unit 243, the partial hypothesis score calculation unit 243, and the hypothesis search unit 244 are predetermined ones stored in the HDD of the storage unit 21 by the CPU This program is realized by developing the program in RAM and executing it.

[Operation of machine translation device]
The operation of the machine translation apparatus shown in FIG. 5 will be described with reference to FIG. 6 (see FIG. 5 as appropriate). FIG. 6 is a flowchart showing the operation of the machine translation apparatus shown in FIG. In advance, the machine translation device 2 learns the weight of the feature function value based on the feature weight learning parallel translation corpus 250, the language model 215, and the rule table 114 by the feature weight learning module 22, and uses the learning result as the learning result. A certain feature weight 211 is stored in the storage means 21.

  Then, the machine translation device 2 inputs the sentence of the translation source language input from the input device K via the input / output means 20 to the word information extraction module 23 (step S11). The machine translation device 2 uses the word information extraction module 23 to divide the input source language sentence (input sentence) into words, and extract the word string, the number of words J, and the word position of each word (step S12). ). The extracted word string, word string word count J, and word position are stored in the storage unit 21 as word information 212.

The machine translation device 2 creates an initial partial hypothesis H ₀ by the hypothesis search means 244, empties _J priority queues Q ₀ ,..., Q _J according to the number of input words J, of which The initial partial hypothesis H ₀ is stored in the priority queue Q ₀ (step S13). Then, the machine translation apparatus 2 initializes the partial hypothesis score S (H ₀ ) for the initial partial hypothesis H ₀ and the identification variable m of the priority queue with the hypothesis search means 244. That is, S (H ₀ ) = 0 and m = 0 are set (step S14). Subsequently, the machine translation device 2 sequentially pops each applicable partial hypothesis H stored therein from the m-th priority queue Q _m by the generation rule search means 241, and each partial hypothesis. Each applicable generation rule r that can expand H is searched from the rule table 114. Then, the number of translated words (number of translated words) V (r) described on the translation source language side of each generation rule r as a search result is acquired, and the identification variable m of the current priority queue. Is added to this value, the value of the number n of translated words for the partial hypothesis H to be processed is updated. That is, n = m + V (r) is set (step S15: generation rule search step).

Then, the machine translation device 2 generates each generation rule r ′ with word range applicable to the searched generation rule by the generation rule generation unit 242 with word range (step S16: generation generation with word range). Step). Then, the machine translation apparatus 2 creates partial hypotheses H ′ by expanding the partial hypotheses H to be processed by the generation rules r ′ with word ranges by the partial hypothesis score calculating means 243, respectively. A partial hypothesis score S (H ′) is calculated for H ′ based on the above equation (15). Then, the created partial hypothesis H ′ is stored in the n-th priority queue Q _n and predetermined pruning is performed (step S17: partial hypothesis score calculation step). Here, the partial hypothesis H ′ that has become unnecessary due to pruning is deleted from the priority queue Q _n . The calculated partial hypothesis score S (H ′) is stored in the partial hypothesis score 216 of the storage means 21, but the partial hypothesis score S (H ′) for the partial hypothesis H ′ deleted by pruning is deleted. The Rukoto.

  Then, the machine translation apparatus 2 determines whether or not all applicable generation rules r ′ with word ranges have been selected by the hypothesis search means 244 (step S18). If applicable r ′ still exists (step S18: No), the machine translation device 2 returns to step S16. On the other hand, if all r ′ are selected (step S18: Yes), the hypothesis searching means 244 determines whether all applicable generation rules r have been selected (step S19). If applicable r still exists (step S19: No), the machine translation device 2 returns to step S15. On the other hand, if all r are selected (step S19: Yes), the hypothesis searching unit 244 determines whether all applicable partial hypotheses H have been selected (step S20). If applicable H still exists (step S20: No), the machine translation device 2 returns to step S15. On the other hand, if all H are selected (step S20: Yes), the hypothesis searching means 244 determines whether or not the value of the identification variable m of the current priority queue is equal to the number of input words J (m = J). (Step S21). If m ≠ J (step S21: No), the machine translation device 2 increments the identification variable m of the priority queue by the hypothesis search means 244. That is, m = m + 1 is set (step S22). Then, it returns to step S15.

On the other hand, when m = J (step S21: Yes), the J-th priority queue Q _J stores a plurality of partial hypotheses H covering the entire sentence of the source language. These partial hypotheses H can be regarded as hypotheses H for sentences in the source language. Similarly, the partial hypothesis score S (H) for the partial hypothesis H stored in the priority queue Q _J is referred to as a hypothesis score S (H) in this sense. In this case, the machine translation device 2 searches the hypothesis search means 244 for a hypothesis H having the maximum hypothesis score S (H) from the J-th priority queue Q _J (step S23: hypothesis search). Step). Then, the machine translation device 2 outputs the searched hypothesis as a sentence in the translation destination language by the hypothesis search means 244 (step S24). As a result, a word string from the beginning of the translation target language to the end of the sentence corresponding to the searched hypothesis is output to the output device D.

  The machine translation apparatus 2 can also be realized by executing a machine translation program that causes a general computer to execute the above steps. This program can be distributed via a communication line, or can be written on a recording medium such as a CD-ROM for distribution.

[Concrete example]
Specific examples will be described with reference to FIG. 7, Table 6, Table 7, and Table 8.
FIG. 7 is a diagram illustrating an extension example from the partial hypothesis illustrated in FIG. 5 to the hypothesis. Table 6 shows a translation source language sentence composed of 11 words, and Table 7 shows a generation rule applicable to the translation source language sentence shown in Table 6. In addition, the “type of generation rule” in Table 7 indicates which of the above formulas (9) to (9d) corresponds. Table 8 shows the order of application of the generation rules shown in Table 7 and the generation rules with word ranges based thereon.

  In the initial state, the partial hypothesis score calculation means 243 empties the stack and, as shown in FIG. 7, the word range [1, 11] covering the entire word string (sentence) of the translation source language in the state “0”. To push. In the state “0”, since the partial hypothesis H ′ (0) is the initial partial hypothesis, it is a set of the word range [1, 11] pushed onto the stack and an empty string. The partial hypothesis score S (H ′) of the partial hypothesis H ′ (0) is “0”.

Next, in the state “1”, the partial hypothesis score calculation unit 243 pops the range [1, 11] from the stack, and the generation rule search unit 241 generates a generation rule applicable to the popped range from Table 7. Select r (3). Since the word position in the input sentence of “ha” corresponding to the translation solution “The” in the generation rule r (3) is “3”, the generation rule generation unit with word range 242 generates the generation rule as shown in Table 8. A generation rule r ′ (3) with a word range such that the word ranges of the non-terminal symbol X ₍₁₎ and the non-terminal symbol X ₍₂ ) in r (3) are [1, 2] and [4, 11], respectively. Generate. In this generation rule r ′ (3) with word range, the non-terminal symbol X ₍₁₎ must be processed before the non-terminal symbol X ₍₂₎ on the translation destination language side. Therefore, as shown in FIG. 7, the partial hypothesis score calculation means 243 pushes the word range [4, 11] corresponding to the nonterminal symbol X ₍₂₎ on the stack in the generation rule r ′ (3) with word range. After that, the word range [1, 2] corresponding to the non-terminal symbol X ₍₁₎ is pushed. The partial hypothesis score calculation means 243 converts the word ranges [1, 2] and [4, 11] pushed onto the stack and the translation solution “The” as a word string from the head of the translation target language into the partial hypothesis H ′ ( 1). The partial hypothesis score calculation means 243 calculates a partial hypothesis score for the partial hypothesis H ′ (1).

Next, in the state “2”, the partial hypothesis score calculation unit 243 pops the word range [1, 2] from the stack, and the generation rule search unit 241 generates a generation rule applicable to this word range from Table 7. Select r (1). Since the word position in the input sentence “international” corresponding to the translation solution “international” in the generation rule r (1) is “1”, the generation rule generation unit with word range 242 generates the generation rule as shown in Table 8. word range of r (1) non-terminal symbol X in ₍₁₎ to produce a conditioned word range generation rule r '(1) such that [2,2]. As shown in FIG. 7, the partial hypothesis score calculation means 243 pushes the word range [2, 2] onto the stack in the generation rule with word range r ′ (1). The partial hypothesis score calculation means 243 translates the word range [2, 2] pushed onto the stack, the previously pushed word range [4, 11], and the translation solution “The international” as a word string from the beginning of the target language. Is a partial hypothesis H ′ (2). The partial hypothesis score calculation means 243 calculates a partial hypothesis score for the partial hypothesis H ′ (2).

  Next, in the state “3”, the partial hypothesis score calculating unit 243 pops the word range [2, 2] from the stack, and the generation rule searching unit 241 uses the generation rule applicable to this word range from Table 7. Select r (2). In the generation rule r (2), the translation solution “terrorism” corresponding to “terrorism” is described, but since there is no non-terminal symbol, the generation rule generation unit with word range 242 uses the generation rule r (2) as it is. The generation rule r ′ (2) with word range is used. The partial hypothesis score calculation means 243 does not operate the stack because the word range is not specified in the generation rule r ′ (2) with word range. As shown in FIG. 7, the partial hypothesis score calculation means 243 partially generates the word range [4, 11] that was previously pushed onto the stack and the translation solution “The international terrorism” as a word string from the head of the target language. Let it be hypothesis H ′ (3). The partial hypothesis score calculation means 243 calculates a partial hypothesis score for the partial hypothesis H ′ (3).

In the same manner, when an operation corresponding to the state “4” to the state “9” is performed in the application order described in Table 8, the stack becomes empty, and the partial hypothesis score calculation unit 243 develops the partial hypothesis. To generate a hypothesis. At this time, the translation solution as a word string from the sentence head of the translation target language is a word string of 10 words as follows.
"The international terrorism also is a possible threat in Japan"

  Here, since the generation rule r (9) has two terminal symbols (words) “is a”, the partial hypothesis score calculation means 243 translates 10 words in nine state transitions. To do. FIG. 7 shows that the translation destination language side was generated from the beginning of the sentence to the end of the sentence in the process of partial hypothesis development.

  FIG. 8 is a diagram showing a tree structure of the synchronous context free grammar corresponding to the extension example to the hypothesis shown in FIG. This is a sequence of generation rules used for partial hypothesis expansion expressed as a tree. In FIG. 8, (1) to (9) represent the expansion order of the generation rules of the synchronous context free grammar. The generation rules start with a terminal symbol on the destination language side on the right side, and the tree is expanded from top to bottom in the order from the beginning of the target language to the end of the target language. Generated automatically. In contrast, the source language side is not always analyzed from the beginning to the end of the sentence, and this point is greatly different from the conventional method.

Table 9 shows an example different from the generation rule with word range shown in Table 8. Table 9 is the same as Table 8 except that the contents of the application orders “1” and “2” are different. In this case, the generation rule r (1) is selected in the state “1”, and the generation rule r (3) is selected in the state “2”. At this time, since the translation solution as a word string from the head of the target language is as follows, the hypothesis score of the hypothesis in the case of Table 8 is as follows. Smaller than.
"International The terrorism also is a possible threat in Japan"

  According to the machine translation apparatus 2 of the present embodiment, the generation rules are expanded top-down so that partial hypotheses are sequentially expanded with respect to untranslated parts in the input word string of the source language, and the corresponding parts By creating a hypothesis, it is possible to sequentially generate a hypothesis of the translation target language corresponding to the input from the beginning of the sentence to the end of the sentence. Further, the machine translation device 2 calculates a partial hypothesis score based on the translation model and language model for each generation rule, and determines the hypothesis having the maximum partial hypothesis score (in this case, the hypothesis score) as a word in the translation destination language. Search as a column. Therefore, the translation accuracy of the word string in the translation destination language can be improved.

  As mentioned above, although embodiment of this invention was described, this invention is not limited to this, It can implement in the range which does not change the meaning. For example, in the present embodiment, Japanese-English translation has been described as an example, but the present invention is not limited to this. If a production rule is created by the production rule creation device, this translation device can be used between any multilinguals.

  Translation experiments were conducted to confirm the effects of the present invention. Specifically, from a Japanese-English bilingual corpus of approximately 180,000 newspaper articles, a feature-weight learning bilingual corpus (development set) and an evaluation set (test set) for evaluating translation results are each 1500 sentences. Sampling. The Japanese-English bilingual corpus of newspaper articles is described in “Masao Utiyama and Hitoshi Isahara. Reliable measures for aligning Japanese-English news articles and sentences. In Pro. Of ACL 2003, pages 72-79, 2003”. In addition, a bilingual corpus (training set) for learning of the “generation rule creation device” is obtained by adding a Japanese-English dictionary (about 1.3 million entries) to the remaining parallel translations. Table 10 shows the size of each training set, development set, and test set.

  In trainingset, we learned a word-based HMM (Hidden Markov Model) translation model and used it to find a many-to-many word correspondence. The HMM translation model is described in “Franz Josef Och and Hermann Ney. A systematic comparison of various statistical alignment models. Computational Linguistics, 29 (1): 19-51, March 2003”.

  Furthermore, the phrase pair was extracted using it, and the production | generation rule was produced | generated (refer Table 11). Phrase and Hierarchal rules / phrases in Table 11 do not contain the number of glue rules. In addition, the number of rules of Normalized-2 and Normalized-3 counts only Expression (9) among Expressions (9) to (9d) described above.

  Phrase indicates a translation method (Comparative Example 1) using a phrase pair described in Non-Patent Document 3. On the other hand, Normalized-2 and Normalized-3 represent translation methods according to this embodiment. Here, Normalized-2 corresponds to a case where a generation rule is obtained under the restriction that the number of non-terminal symbols is two (Example 1). Further, Normalized-3 corresponds to the one for which the generation rule is obtained under the restriction that the number of non-terminal symbols is three (Example 2). Moreover, Hierarchical represents the translation method by a nonpatent literature 2 (comparative example 2). As is clear from Table 11, in this embodiment (Normalized-2, Normalized-3), although the number of generation rules is larger than that of Phrase, the number of generation rules can be significantly suppressed as compared with Hierarchical. Furthermore, Table 12 shows the translation accuracy.

  In an experiment for comparing translation accuracy, statistical machine translation based on phrases (Phrase is used as a translation technique) and statistical machine translation according to the present embodiment (Nomalized-2 is used as a translation technique) were compared. For each, 3-gram and 5-gram were used as language models. As the evaluation scale, BLEU and NIST, which are indices based on the accuracy rate of n-gram, were used.

  BLEU is described in “Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. BLEU: a method for automatic evaluation of machine translation. In Proc. Of ACL 2002, p. 311-318, 2002”. Has been. For NIST, see “National Institute of Standards and Technology. Automatic evaluation of machine translation quality using n-gram co-occurrence statistics. Http://www.nist.gov/speech/tests/mt/doc/ngram-study .pdf, 2002 ".

  The experimental results in Table 12 indicate that Normalized-2 (Example 1) can obtain translation with higher accuracy than Phrase (Comparative Example 1). In addition, when changing from 3-gram to 5-gram, the accuracy improvement of Normalized-2 is significant compared to Phrase, suggesting that Example 1 can effectively use a long n-gram. Although the embodiment of the present invention has a strong restriction on the generation rule, it can be said that the restriction provided on the generation rule is valid as a translation model from the experimental results comparing the translation accuracy of Phrase and Normalized-2.

It is a functional block diagram which shows the structure of the production | generation rule production apparatus which concerns on embodiment of this invention. It is a figure which shows the example of word correspondence of a Japanese-English parallel translation. It is a figure which shows the example of the rule table shown in FIG. It is a flowchart which shows operation | movement of the production | generation rule preparation apparatus shown in FIG. It is a functional block diagram which shows the structure of the machine translation apparatus which concerns on embodiment of this invention. It is a flowchart which shows operation | movement of the machine translation apparatus shown in FIG. It is a figure which shows the example of an extension from the partial hypothesis shown in FIG. 5 to a hypothesis. It is a figure which shows the tree structure of the synchronous context free grammar corresponding to the example extended to the hypothesis shown in FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Generation rule creation apparatus 2 Machine translation apparatus 10 Input / output means 11 Storage means 111 Word correspondence 112 Phrase pair 113 Rule 114 Rule table 12 Word correspondence creation module 13 Control means 131 Mode determination means 132 Phrase pair extraction means 133 Generation rule creation means 134 Translation score calculation means 150 Bilingual corpus 20 Input / output means 21 Storage means 211 Feature weight 212 Word information 213 Rule with word range 214 Partial hypothesis 215 Language model 216 Partial hypothesis score 22 Feature weight learning module 23 Word information extraction module 24 Control means 241 Generation Rule search means 242 Generation rule generation means with word range 243 Partial hypothesis score calculation means 244 Hypothesis search means 250 Bilingual corpus for feature weight learning K input device D output device

Claims (7)

Stores multiple generation rules generated so that the destination language symbol string on the right side of the generation rule that generates the target language word string from the source language word string using the synchronous context free grammar starts from the terminal symbol A new partial hypothesis that is longer than a predetermined partial hypothesis is generated as a word sequence in the translation target language corresponding to the input, which is the translation result of the input word sequence in the translation source language. A machine translation device that outputs a hypothesis that is a partial hypothesis finally generated by sequentially creating and extending the predetermined partial hypothesis,
Generation rule search means for searching each rule table for a generation rule applicable to extend the predetermined partial hypothesis;
The non-terminal symbol of the searched generation rule is determined based on the number of words and the word position of the word constituting the input word string of the input source language with respect to the source language side of the searched generation rule. A word range generation rule generating means for adding a word range indicating a range of a word string of the translation source language to be generated and generating each applicable generation rule with a word range;
A translated word on the translation target language side included in the generated generation rule with a word range and the word range are created as the new partial hypothesis, and the applicable generation rule is applied top-down, and In the applicable production rule, the new partial hypothesis is expanded in the order in which the non-terminal symbols on the translation target language side are arranged from the head of the sentence to the end of the sentence. A partial hypothesis for calculating a partial hypothesis score indicating an evaluation value of the created partial hypothesis based on a score of the translation model representing the likelihood and a score of the language model representing the probability as the word string of the translation target language A score calculation means;
Among the partial hypotheses finally generated by searching for a predetermined partial hypothesis applicable to the input word string of the source language and extending the predetermined partial hypothesis, the partial hypothesis score A hypothesis search means for searching for a partial hypothesis with the maximum as the hypothesis,
A machine translation device comprising: Stores multiple generation rules generated so that the destination language symbol string on the right side of the generation rule that generates the target language word string from the source language word string using the synchronous context free grammar starts from the terminal symbol A new partial hypothesis that is longer than a predetermined partial hypothesis is generated as a word sequence in the translation target language corresponding to the input, which is the translation result of the input word sequence in the translation source language. A machine translation method of a machine translation device that outputs a hypothesis that is a partial hypothesis finally generated by sequentially creating and expanding the predetermined partial hypothesis,
The machine translation device includes:
A production rule search step of searching the rule table for a production rule applicable to extend the predetermined partial hypothesis;
The non-terminal symbol of the searched generation rule is determined based on the number of words and the word position of the word constituting the input word string of the input source language with respect to the source language side of the searched generation rule. Adding a word range indicating a range of a word string of the translation source language to be covered, and generating a generation rule with a word range, each generating a generation rule with a word range,
A translated word on the translation target language side included in the generated generation rule with a word range and the word range are created as the new partial hypothesis, and the applicable generation rule is applied top-down, and In the applicable production rule, the new partial hypothesis is expanded in the order in which the non-terminal symbols on the translation target language side are arranged from the head of the sentence to the end of the sentence. A partial hypothesis for calculating a partial hypothesis score indicating an evaluation value of the created partial hypothesis based on a score of the translation model representing the likelihood and a score of the language model representing the probability as the word string of the translation target language A score calculating step;
Among the partial hypotheses finally generated by searching for a predetermined partial hypothesis applicable to the input word string of the source language and extending the predetermined partial hypothesis, the partial hypothesis score A hypothesis search step for searching for a partial hypothesis with the maximum as the hypothesis,
A machine translation method comprising: A production rule creation device for creating a production rule used in the machine translation device according to claim 1,
The combination stored in the bilingual corpus having a plurality of combinations of a source language word string and a target language word string having the same meaning based on the word correspondence between the source language word and the target language word A phrase pair extraction means for extracting a combination of words or phrases having the same meaning in the translation source language and the translation destination language as a phrase pair;
Based on the extracted phrase pair, in the generation rule of the synchronization context free grammar, a generation rule creating means for creating a generation rule to which a restriction that the symbol string of the translation target language on the right side starts with a terminal symbol is added;
A production rule creation device comprising: A production rule creation method of a production rule creation device for creating a production rule used in the machine translation device according to claim 1,
The generation rule creating device is
The combination stored in the bilingual corpus having a plurality of combinations of a source language word string and a target language word string having the same meaning based on the word correspondence between the source language word and the target language word A phrase pair extraction step of extracting a combination of words or phrases having the same meaning in the translation source language and the translation destination language as a phrase pair;
Based on the extracted phrase pair, in the generation rule of the synchronous context free grammar, a generation rule creation step of creating a generation rule with a restriction that the symbol string of the translation target language on the right side starts with a terminal symbol;
A production rule creation method characterized by comprising:   A machine translation program for causing a computer to execute the machine translation method according to claim 2.   A generation rule creation program for causing a computer to execute the production rule creation method according to claim 4.   A computer-readable recording medium in which the machine translation program according to claim 5 or the production rule creation program according to claim 6 is recorded.

Images