JP5148583B2 - Machine translation apparatus, method and program - Google Patents

Description

The present invention relates to a machine translation apparatus , method, and program for translating an original sentence of a first language to be translated into a translated sentence of a second language using an example of translation consisting of a parallel translation of a first language sentence and a second language sentence.

  There is a machine translation device having a function of translating an original sentence of a first language to be translated into a sentence of a second language by a computer. In such a machine translation apparatus, a plurality of translation examples consisting of parallel translations of a first language sentence and a second language sentence are stored in advance in a translation example database, and a translation example similar to the input translation target sentence is searched. Then, it is presented to the user together with the original text to be translated, and the translation is supported. The translation source text often does not exactly match the translation target text, so there are limited cases where the translation text of the translation example can be directly applied as a translation corresponding to the translation target text without editing. The translation of the original text to be translated is obtained after performing such editing on the translation example translation (see, for example, Patent Document 1 and Patent Document 2).

  In translation using such a translation example, it is time consuming because the user needs to determine the editing location and select a translation. Therefore, a method for automatically performing the editing has been proposed. In this case, differences (hereinafter referred to as differences) between the translation target original text and the searched translation example original text are determined in terms of phrases, the translation example translation phrases corresponding to the differences in the translation example original text are determined, and the corresponding phrases are translated. A synthetic sentence is created by replacing the difference words in the target original sentence, and a translation sentence is created by replacing the words in the translation target original sentence included in the synthetic sentence with the translation corresponding to the word (for example, see Patent Document 3).

JP 2003-330924 A JP 2005-339087 A JP 2006-11842 A

However, in Patent Document 3, the difference between the original text to be translated and the original text for translation is determined in units of words, the word / phrase of the example translation for translation corresponding to the difference in the original text for translation is determined, and the original text to be translated is determined. A compound sentence that has been replaced with a difference word in the middle sentence is created, and a translation sentence in which the word or phrase in the original sentence included in the compound sentence is replaced with a translation corresponding to the word or phrase, the difference word or phrase is context information or Translation according to the field information of the document may not be possible. That is, since a translation corresponding to a difference word in the original text to be translated is replaced, a default translation is always used among the translation candidates registered in the dictionary, which may result in an inappropriate translation. Table 1 shows an example of an inappropriate translation.

  In Table 1, the difference between the translation source text and the translation example text is “strain” and “antigen”, and the word in the translation example translation corresponding to “antigen” in the translation example text is “antigen” Is shown. In this case, the compound sentence in which the “antigen” in the translation example translation is replaced with “strain” in the translation target sentence becomes “this strain was used for the prevention of tuberculosis”. Then, referring to the translation candidate of “strain”, the “translation” which is a predetermined translation is selected from the translation candidates, and the automatically generated translation is generated. Thus, the automatically generated translation is "This load was used for TB vaccination."

  However, the desired translation is “This strain was used for the TB injection” and cannot be translated. This is the result of replacing with the default translation without considering the context of the original text to be translated and the field of the document.

The purpose of the present invention is to create contextual or document field information for a difference word or phrase in the translation target original text when creating a translation in which the word or phrase in the translation example translation is replaced with a difference translation in the translation original text. It is an object to provide a machine translation apparatus , method, and program capable of generating a translation of an appropriate word / phrase according to the above.

The machine translation apparatus according to the present invention includes a translation unit that performs morphological analysis and syntax analysis on an input source text to be translated using vocabulary and rules defined in a machine translation dictionary , and translates without using a translation example . A translation example search means for searching a translation example that is the same as or similar to the translation target text from the translation example database using the source text as a search key, and a difference portion between the source text of the searched translation example and the source text to be translated a differential associating means for associating the words in the phrase and the translated original text in the translation of the translation example necessary edits for the translation example in the translation of the phrase in the translated source text associated for the phrase, a translation obtained by using information other than the language in the translated original text by translating the translated original text using the translation means Translated words corresponding to words in said translated original text associated with said difference correlating means of its translation sentence, a difference phrase translation acquisition means for acquiring as a translation corresponding to the context in the translated original text, the difference obtain Bei and translation substitution means a phrase obtained by translation acquisition unit translation by replacing the phrase in translation of the editing needs translation example to complete the translation.

  According to the present invention, when creating a translation in which the word / phrase in the translation example translation is replaced with the translation of the difference in the translation example original, context information and document field information are obtained for the difference phrase in the translation target original. It is possible to provide a machine translation device that can generate a translation of an appropriate word or phrase according to the situation.

The functional block diagram of the machine translation apparatus concerning the 1st Embodiment of this invention. The block block diagram which shows the hardware constitutions of the machine translation apparatus which concerns on embodiment of this invention. The flowchart which shows the processing content of the machine translation apparatus concerning the 1st Embodiment of this invention. The flowchart which shows the content of the difference matching process in the difference matching means in the 1st Embodiment of this invention. Explanatory drawing of the syntax tree of the example sentence for translation and the example translation sentence for translation in which the morphological analysis and syntax analysis in the 1st Embodiment of this invention were carried out. The flowchart which shows the content of the difference word phrase translation acquisition process in the difference word phrase translation acquisition part in the 1st Embodiment of this invention. The functional block diagram of the machine translation apparatus concerning the 2nd Embodiment of this invention. The flowchart which shows the content of the difference word phrase translation acquisition process in the difference word phrase translation acquisition part in the 2nd Embodiment of this invention. The flowchart which shows the content of the difference word phrase translation acquisition process of the difference word phrase translation acquisition part in the 3rd Embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a functional block diagram of the machine translation apparatus 11 according to the first embodiment of the present invention, and FIG. 2 is a block configuration diagram showing a hardware configuration of the machine translation apparatus according to the embodiment of the present invention.

  In FIG. 2, the machine translation device 11 is realized by installing a software program such as a machine translation program in a general computer and executing the software program in the processor 13 of the arithmetic control device 12.

  The arithmetic control device 12 performs various arithmetic operations related to machine translation. The arithmetic control device 12 has a processor 13 and a memory 14. A memory translation program 15 is stored in the memory 14, and the processor 13 performs processing. When executed, the work area 16 is used. Calculation results and the like of the calculation control device 12 are displayed and output on the display device 18 that is the output device 17 and also output to the communication network via the communication control device 19.

  The input device 20 is used to input information to the arithmetic control device 12, and includes, for example, a mouse 21, a keyboard 22, a disk drive 23, and a communication control device 19. For example, the mouse 21 and the keyboard 22 are connected via the display device 18. Then, various commands are input to the arithmetic and control unit 12, and the keyboard 22, the disk drive 23, and the communication control unit 19 input a document to be translated.

  That is, the disk drive 23 inputs / outputs a file of a document to be translated to / from a storage medium, and the communication control device 19 connects the machine translation device 11 to a communication network such as the Internet or a LAN. The communication control device 19 is a device such as a LAN card or a modem, and data transmitted / received to / from the communication network via the communication control device 19 is transmitted / received to / from the arithmetic control device 12 as an input signal or an output signal. Further, a hard disk drive (HDD) 24 is provided for storing a calculation dictionary of the calculation control device 12 and a translation dictionary storing knowledge and rules necessary for translation.

  FIG. 1 is a functional block diagram of a machine translation apparatus 11 according to the first embodiment of the present invention. Each functional block in the arithmetic and control unit 12 shown in FIG. 1 corresponds to each program constituting the machine translation program 15 described above. That is, when the processor 13 executes each program constituting the machine translation program 15, the arithmetic control device 12 functions as each functional block. Each block of the storage device 25 corresponds to a storage area of the memory 14 and the hard disk drive 24 in the arithmetic control device 12.

The input processing unit 26 inputs the original text data of the first language to be translated input from the input device 20 and outputs it to the translation unit 27, and inputs information necessary for the operation from the input device 20 to the translation unit 27. Gives various commands.

  The display processing unit 28 includes the original text of the first language to be translated input from the input device 20 via the input processing unit 26, the translated text of the second language translated by the translation unit 27, and the translation stored in the storage device 25. An example of translation in the example database 29 is displayed on the display device 18. Information necessary for an operation input from the input device 20 via the input processing unit 26 is also displayed and output to the display device 18 via the display processing unit 28.

  The translation example database 29 is a database in which a plurality of translation examples composed of parallel translations of sentences in the first language and second language registered in advance are stored. The machine translation dictionary 30 includes the first language in the second language. The vocabulary and rules necessary for translation into are stored.

  The translation unit 27 translates the first language into the second language using the vocabulary and rules of the machine translation dictionary 30, and includes a translation unit 31, a translation example search unit 32, a difference association unit 33, and a difference phrase. A translation acquisition unit 34 and a translation replacement unit 35 are provided.

  The translation means 31 performs a syntax analysis or a morphological analysis on the sentence or phrase in the first language and translates it into the second language. The translation example search means 32 searches the translation example database 29 for a translation example similar to the translation target original text using the input translation target original text of the first language as a search key. In addition to associating a difference between the translation example original searched for by the search means 32 and the translation target original, the phrase in the translation example translation that needs to be edited corresponding to the difference is associated with the phrase in the translation target original. is there.

  Then, the difference word phrase translated word acquisition unit 34 applies the phrase in the translation target text to the phrase in the translation target text that is associated with the phrase in the translation text of the translation example, that is, the phrase in the translation example translation text that needs to be edited. Uses information other than the word / phrase to obtain an appropriate translation according to the context or field in the original text to be translated, and the translation replacement unit 35 needs to edit the translation acquired by the differential word / phrase translation acquisition unit 34 The translation is completed by replacing words in the translation example.

FIG. 3 is a flowchart showing the processing contents of the machine translation apparatus according to the first embodiment of the present invention. In the following description, a combination of the translation target original text and the translation example (translation example original text and translation example translation text) shown in Table 2 will be described as an example.

  When the translation target original text in the first language is input from the input device 20 to the translation section 27 via the input processing section 26 and a translation request is issued, the translation means 31 of the translation section 27 responds to the translation target text. Dictionary lookup processing and morphological analysis are performed using the machine translation dictionary 30 (S1), and various information such as parts of speech, types of utilization, various attributes, and translations of each morpheme are obtained.

  Subsequently, the translation example search means 32 performs a translation example search on the translation example database 29 (S2). That is, it is searched whether or not the translation example (translation example original sentence and translation example translation) that is the same as or similar to the input sentence is stored in the translation example database 29.

  At this stage, since the morphological analysis result of the translation target text and various information about each word in the translation target text are acquired by the processing of step S1, the katakana and the kana notation, kanji notation and hiragana notation Information on the shaking and the presence / absence of a sending pseudonym is also acquired. With these pieces of information, even if the translation example does not completely match the input sentence with one character, it is possible to perform a vibration absorption process that does not recognize the difference regarding the fluctuation of these notations as a difference.

  On the other hand, a configuration in which the search is successful only when there is an exact original translation example that does not absorb the fluctuation of the notation is also possible. Moreover, even if there is a difference between characters and words, a configuration may be adopted in which the number of different characters regarded as the same value or the lower limit value of the ratio of the number of different words is provided to control the possibility of search. As a result, it is possible to extract a translation example original that is similar in terms and phrases. When such a search method is used, it is not necessary to set the timing of the example search after the dictionary lookup process and the morphological analysis, and the translation example search step may be performed at the beginning of the process of FIG.

  In the translation example search in step S2, there may be cases where a plurality of translation examples can be searched, but the translation example with the highest priority to be finally used is the processing target. In order to determine an example with the highest priority from a plurality of translation examples, it is common to use a criterion such as the registration date of the example when the degree of similarity is the same. The similarity is determined, for example, by determining that the similarity is higher as the number of matching words is larger.

  Now, as an example for translation, the original text for translation is “The antigen was used to vaccinate people against tuberculosis.”, And the translated text for translation is “Prevention of tuberculosis”. Suppose this antigen was used for injection. "

  If such a translation example is searched (S3), the difference association means 33 performs a difference association process on the translation example used for translation and the translation target original (S4). FIG. 4 is a flowchart showing the contents of the difference association process (step S4) in the difference association means 33.

  As shown in FIG. 4, the difference association unit 33 first performs syntax analysis on the translation target original (S11), and performs morphological analysis and syntax analysis on the translation example original and translation example translation (S12). . As a result, the original text to be translated, the original text for translation, the word string constituting the text of the translated text for translation, the part of speech of each word, the type of use / translation, various information necessary for translation processing, and the syntax structure are obtained. Using these pieces of information, the difference between the translation target original and the translation example original is associated (S13), and which words and phrases in the translation example original and the translation example translation are associated with each other. (S14) Further, using this result, the difference between the translation target original text and the example original text is associated with which word / phrase in the example translated text (S15).

Table 3 shows the analysis results of the original text to be translated. Table 3 also shows the phrase ID of the translation example original corresponding to the phrase of the original text to be translated.

  In the process of associating the difference between the translation target original text and the translation example original text (step S13), the differences between the translation target original text and the translation example original text are correlated by comparing Table 3 and Table 4. As a result, the difference between the translation target original and the translation example original is found to be a difference between “strain” of the translation target original and “antigen” of the translation example original.

  Next, the association process (step S14) of which words in the translation example original sentence and which words in the translation example translation correspond to each other will be described.

Figure 5 is an illustration of syntax tree and the translation example sentence and morphological analysis and parsed translation example original sentence, 5 (a) is the syntax tree diagram of the translation example original sentence, 5 (b) is translated example sentence FIG. In FIG. 5 , “TW” indicates a list of translation candidates that exist as translation candidates for the translation example translations that are the second language. Multiple items are listed separated by commas.

  With reference to FIG. 5, the association process of which words in the translation example original text and the translation example translation text correspond to which words in step S14 will be described.

(1) The following processing is performed for each component of the analysis result of the translation example original.

(1a) For translation headwords (each element after morphological analysis), search for an element including the “TW” attribute in the translation result of translation example translation, and extract a translation corresponding to the headword as a corresponding candidate. If there are a plurality of pairs, a plurality of candidate pairs are extracted.

(1b) It is checked whether or not one of the translation word candidates of the translation example original sentence element matches the component in the analysis result of the translation example translation sentence, and if there is a match, it is extracted as a corresponding candidate. If there are a plurality of pairs, a plurality of candidate pairs are extracted.

(2) As a result of the above (1), those whose correspondences do not overlap and whose correspondence is uniquely determined are determined as corresponding terms. In addition, even when the number of elements does not match, such as when one element in one structure matches two or more continuous elements in the other structure, it is extracted as a corresponding phrase.

(3) As a result of the above (1), when the correspondence is ambiguous, for example, when the same word is used twice, it is directly modified in the analysis result with other corresponding words / phrases. Further, based on the presence / absence of crossing of the modification relationship and the coexistence relationship with other corresponding word / phrase candidates in the partial structure, a more likely correspondence relationship is selected to eliminate the ambiguity.

(4) When the corresponding phrase candidates are extracted by the above (3), structurally continuous among the analysis result structures of the first language sentence and the analysis result structure of the second language sentence of the corresponding phrase candidates. Detecting the part that is.

  The following correspondence is obtained by this association processing. “Antigen”, “use”, “tuberculosis”.

Tables 4 and 5 show lists of internal data after the association. Table 4 shows the analysis result of the translation example original text, and includes individual components of the translation example original text, part of speech information, and an ID number indicating which word in the data structure of the translation example translation text. An element including “−1” indicates that there is no corresponding phrase.

Table 5 shows the analysis result of the translation example translation. Similarly to Table 4, each component of the translation example translation, part-of-speech information, and an ID number indicating which word / phrase of the translation example original data structure corresponds to And an element containing “−1” indicates that there is no corresponding phrase.

  From this correspondence, the difference between the translation target text and the translation example text is the difference between “strain” and “antigen”, and the translation example text “antigen” corresponds from the correspondence between the translation example text and the translation example translation text. It can be seen that the phrase of the translated translation example is “antigen”. From this result, in step S15, it is obtained that the correspondence between the difference between the translation target original and the translation example original and the translation example translation is “strain-antigen”.

  In this way, for the difference phrase “strain-antigen” between the translation target original and the translation example original obtained by the difference association process (step S4) in FIG. As a translation of “Antigen”, an appropriate translation instead of “antigen” is acquired (S5).

  FIG. 6 is a flowchart showing the contents of the difference word / phrase translation acquisition process (step S5) in the difference word / phrase translation acquisition means 34. First, the original text to be translated is translated (S21). That is, the original text to be translated is translated using the translation means 31 without using a translation example. After that, the translation of the difference word between the translation target original and the translation example original is cut out (S22).

  The translation process in step S21 is a general machine translation process. In other words, the translation means 31 does not use a translation example, and applies the translation rules defined in the machine translation dictionary 30 using information such as clause dependency and co-occurrence relationship obtained by parsing. Get the translation.

  As shown in Table 1, there are a plurality of translated words (load to pathogen) in the phrase “strain” of the translation target text of the difference phrase between the translation target text and the translation example original text, and which translation word is appropriate is determined. As knowledge, the following rules for determining translation are stored in the machine translation dictionary 30.

<Rule> strain + vaccinate → strain = bacteria (Meaning: When “strain” and “vaccinate” co-occur, “strain” is translated as “bacteria”)
According to this rule, the translation of the original text to be translated "The strain was used to vaccinate people against tuberculosis."

Translation: “The fungus was used to vaccinate people with TB.”
Next, in step S22, the translated word “strain” is extracted from the translated sentence obtained in step S21 “The fungus species were used to give people a preventive injection of tuberculosis.” In this translation, the translation information of “strain” is obtained as “bacterial species”, so that “bacterial species” that is the translation of the difference phrase between the translation target original and the translation example original is acquired.

  In this way, when the difference word / phrase translation acquisition (step S5) by the difference word / phrase translation acquisition unit 34 of FIG. 3 acquires the “fungus species” that is the translation of the difference word / phrase between the translation target original and the translation example original, The translated word replacement means 35 performs translated word replacement processing (S6). In this translated word replacement process, this antigen was used for the translation of the translated example translation “tuberculosis” using the “strain-antigen” correspondence obtained in the difference matching process (step S4). The “antigen” in is replaced with the “fungus species” obtained in the differential word phrase translation acquisition (step S5) to complete the translation. As a result, the translation: “This bacterium was used for vaccination for tuberculosis” is obtained, and a translation with an appropriate translation according to the context in the original text to be translated can be obtained.

  FIG. 7 is a functional block diagram of a machine translation apparatus according to the second embodiment of the present invention. This second embodiment is different from the example shown in FIGS. 1 to 6 in that the difference word / phrase translation acquisition unit 34 uses the translation unit 31 to translate the original text to be translated instead of an external translation system 36. Is used to translate the original text to be translated.

  This second embodiment of the present invention will also be described by taking the combination of the translation target original text and the translation example (translation example original text and translation example translation text) shown in Table 2 as an example. In the second embodiment, with respect to the example of Table 2, the processes up to the difference association process (step S4) in FIG. 3 are the same as those in the first embodiment. The difference word / phrase translation acquisition process (step S5) in FIG.

  FIG. 8 is a flowchart showing the contents of the difference word / phrase translation acquisition process in the difference word / phrase translation acquisition unit 34 according to the second embodiment of this invention. First, a translation of the original text to be translated is requested (S41). That is, the difference word / phrase translation acquisition unit 34 requests an external translation system 36 to translate the original text to be translated. Thereafter, a translation of the translation result is received from the external translation system 36 (S42). Then, a differential word phrase translation is acquired from the received translation (S43). In this case, it is not necessary to know how the external translation system 36 performs translation processing to generate a translation.

  In this case, in step S43, in order to obtain the translation of the difference word / phrase from the translation received from the external translation system 36, the same processing as in step S14 in FIG. 4 is performed on the translation target original sentence and the received translation pair. It will be.

  Now, for the example in Table 2, the difference word / phrase translation acquisition means 34 asks the external translation system 36 to translate, and this bacterial species is used for the vaccination of the translation of the external translation system “human tuberculosis”. ”. In order to acquire a translated word from the received translated sentence, the difference word / phrase translated word acquisition unit 34 performs the same processing as step S14 of FIG. 4 in step S43. In other words, the translation target text “The strain was used to vaccinate people against tuberculosis.” And the translation received from an external translation system “This strain was used for vaccination against human tuberculosis.” Thus, the same processing as step S14 in FIG. 4 is performed. By this process, “fungus species” is acquired as a translation of “strain”.

  Finally, the translation replacement process is performed by the translation replacement unit 35, and the translation example "translation" is used for the preventive injection of tuberculosis using the "starin-antigen" correspondence obtained by the difference correlation processing unit 33. The “antigen” in “” is replaced with “fungal species” to complete the translation. As a result, the following translation is obtained. As a result, the translation: “This bacterium was used for vaccination for tuberculosis” is obtained, and a translation with an appropriate translation according to the context in the original text to be translated can be obtained.

  In the second embodiment, only the translation is received from the external translation system 36, but in addition to the translation, the correspondence relationship between the translation target original sentence and the phrase of the translation is also acquired from the external translation system 36. Good. In this case, in the “acquire difference word / phrase translation” process in step S43 of the difference word / phrase translation acquisition unit 34, the translation of the difference word / phrase can be easily acquired from the correspondence between the translation target original sentence and the phrase of the translation, FIG. It is not necessary to perform the same process as step S14.

  Next, a third embodiment of the present invention will be described. In the first embodiment and the second embodiment, the translation source 31 is translated using the translation means 31 and the external translation system 36 without using a translation example, and the translation obtained by the translation is The translation of the difference word / phrase is made based on the correspondence between the translation target sentence and the phrase of the translation sentence, but the field of the first language document including the translation target sentence is determined using the translation means 31 or the external translation system 36. It is also possible to acquire the field information and collate it with the field information of the translated word of the word / phrase in the original text to be translated associated by the difference associating means 33 to obtain a translation having the same field information as an appropriate translation. Good.

Table 6 shows an example in which the translation of the difference word / phrase of the original text to be translated is acquired using the field information of the translation original text document.

  In this case, the “base-fleet” correspondence is obtained by the difference association process (step S4) by the difference association unit 34 shown in FIG. If you translate the original text “Formation and equipment of this base were reported” without using the translation examples, you will get a translation of “This basic configuration and equipment has been reported.”

  In the first embodiment and the second embodiment, the “Fleet” in the translation example translation is replaced with the “basic” translation translated without using the translation example. Therefore, in the first embodiment and the second embodiment, a translation of “This foundation organization and equipment has been reported” is obtained. In other words, “base” is translated as “basic” as the default translation.

Here, some of the translated words in the machine translation dictionary have field information. For example, “base” is translated as “base” in the baseball field, “base” in the military field, and “base” in the other fields. In the machine translation dictionary 30, for example, the field information is stored as shown in Table 7.

  Therefore, the differential word phrase translation acquisition unit 34 obtains an appropriate translation of the differential phrase using the field information of the document. FIG. 9 is a flowchart showing the contents of the difference word / phrase translation acquisition process of the difference word / phrase translation acquisition means 34.

  First, the difference word / phrase translation acquisition unit 34 performs field determination on the original document using the translation unit 31 or the external translation system 36 (S31). The translation word candidates of the difference words between the translation target original text and the translation example original text are collated in order (S32). It is determined whether the field information is attached to the candidate translation (S33). If the field information is attached, it is determined whether the field information of the original document obtained in S31 matches the field information of the translation (S34). . If they match, the translation candidate is regarded as a translation and the process is terminated (S35). On the other hand, if field information is not attached to the candidate translation in the determination in step S33, and it does not match in the determination in step S34, it is determined whether there is a next translation (S36), and if there is a next translation, step Return to S32. When there is no next translation, the specified translation of the translation candidate is used as a translation (S37).

  To explain using the example in Table 6, first, in step S31, a field determination is performed on the “translation original document” in Table 6. As a result, “field = military” is obtained. From step S32, individual translations of “base” are collated in order, and the translation “base” having the same “field = military” is obtained.

  Then, by the translation replacement process in step S6 of FIG. 3, the “base-fleet” correspondence is used to replace “fleet” in the translation example translation with “base” to complete the translation. As a result, the translation “The formation and equipment of this base has been reported.” Is obtained.

  According to the embodiment of the present invention, in the translation using the translation example, the difference between the translation target original and the translation example original is determined in units of words, and the translation example translation corresponding to the difference in the translation example original is determined. Judgment and use of context information and document field information for the difference words in the translation source text when creating a translation that replaces the words in the translation example translation with the translation differences in the translation example source text Thus, a translation is determined and a translation is generated by replacing the translation with the translation, so that an appropriate translation can be obtained.

DESCRIPTION OF SYMBOLS 11 ... Machine translation apparatus, 12 ... Operation control apparatus, 13 ... Processor, 14 ... Memory, 15 ... Machine translation program, 16 ... Work area, 17 ... Output device, 18 ... Display apparatus, 19 ... Communication control apparatus, 20 ... Input Device 21 ... Mouse 22 ... Keyboard 23 ... Disk Drive 24 ... Hard Disk Drive 25 ... Storage Device 26 ... Input Processing Unit 27 ... Translation Unit 28 ... Display Processing Unit 29 ... Translation Example Database 30 ... Machine translation dictionary, 31 ... translation means, 32 ... translation example search means, 33 ... difference association means, 34 ... difference word phrase translation acquisition means, 35 ... translation replacement means, 36 ... external translation system

Claims (3)

A translation means for performing morphological analysis and syntax analysis on the input original text to be translated using the vocabulary and rules defined in the machine translation dictionary , and translating without using a translation example ;
A translation example search means for searching a translation example that is the same as or similar to the translation target text from a translation example database using the translation target text as a search key;
A differential associating means for associating the words in the phrase and the translated original text in the translation of the translation examples require edits to difference portions of the original and the translated original text of the matching translation example,
For the word in the translated original text associated with the word in the translated sentence of the translation example, the information other than the words of the translated in the original sentence by translating the translated original text using the translation means A translation obtained by using the translation corresponding to the phrase in the translation target original text correlated by the difference correlation means in the translated text is acquired as a translation corresponding to the context in the translation target text Differential word phrase translation acquisition means to
The difference phrase the translation obtained by the translation acquisition unit replaces the word in the translation of the editing is required translations example obtain Preparations and translation substitution means to complete the translation machine translation apparatus. A translation step for translating the input original text to be translated without using a translation example by performing morphological analysis and syntax analysis using the vocabulary and rules defined in the machine translation dictionary;
A translation example search step for searching a translation example identical or similar to the translation target text from a translation example database using the translation target original as a search key;
A differential mapping step for associating the words in the phrase and the translated original text in the translation of the translation examples require edits to difference portions of the original and the translated original text of the matching translation example,
For the word in the translated original text associated with the word in the translated sentence of the translation example, the information other than the words of the translated in the original sentence by translating the translated original text using the translation step A translation obtained by using the translation corresponding to the phrase in the source text to be translated, which is correlated in the difference matching step in the translation, is acquired as a translation according to the context in the source text to be translated A differential word / phrase translation acquisition step,
Machine translation method to obtain Preparations and translation substitution step to complete the replaced phrase translation in the translation of the difference phrase translation Translation examples require the editing the acquired translation by the obtaining step. To a computer used as a machine translation device that inputs a source language of a first language to be translated and outputs a translation of a second language,
A translation function that performs morphological analysis and syntactic analysis on the input source text to be translated using vocabulary and rules defined in the machine translation dictionary , and translates without using a translation example ;
A translation example search function for searching the translation example database for the same or similar translation example from the translation example database using the translation source text as a search key;
A differential correlation function for associating the words in the phrase and the translated original text in the translation of the translation examples require edits to difference portions of the original and the translated original text of the matching translation example,
For the word in the translated original text associated with the word in the translated sentence of the translation example, the information other than the words of the translated in the original sentence by translating the translated original text using the translation function A translated word obtained by using the translation corresponding to the phrase in the source text to be translated, which is correlated with the difference matching function in the translated text, is acquired as a translated word according to the context in the source text to be translated A differential word translation acquisition function to
A program for realizing a translation replacement function that completes a translation by replacing the translation acquired by the differential phrase translation acquisition function with a phrase in the translation of the translation example that needs to be edited.

Images