JP2011070237A - Machine translation system, machine translation method, and machine translation program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To use the information of a dictionary on another machine translation system at translation processing; and to reduce a data transfer amount between systems. <P>SOLUTION: A first translation system has: a first dictionary storage means storing a dictionary used for translation, a first translating means for translating an inputted sentence by using the dictionary stored in the first dictionary storage means, and a dictionary entry extracting means for extracting a dictionary entry associated with the dictionary entry used for the translation processing by the first translating means from the dictionary stored in the first dictionary storage means by using the dictionary stored in the first dictionary storage means, and transmitting the extracted dictionary entry to the second translation system via a communication line. The second translation system has: a second dictionary storage means for storing a dictionary used for translation; and a dictionary registering means for receiving the dictionary entry transmitted from the first translation system via the communication line and registering the received dictionary entry in the dictionary stored in the second dictionary storage means. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a machine translation system, a machine translation method, and a machine translation program for translating text and voice from a first language to a second language.

  As a system that supports communication between different languages, there is a machine translation system that translates text and speech. These systems receive a natural language input from a user in the form of text or speech, convert between languages by referring to a dictionary, and output a translation result in the form of text or speech.

  In the machine translation system as described above, a user dictionary is often used in addition to a system dictionary storing commonly used words. In the user dictionary, words not included in the system dictionary, such as words specific to the user or field, are registered and used for translation. For example, if the user's name is not included in the system dictionary, this user name is registered in the user dictionary so that his / her name is correctly translated. For example, in a Japanese-English translation system, even if the Japanese name “Nakano” is not included in the system dictionary, the user dictionary changes from “Nakano” to “Nakano” from Japanese to English. A dictionary entry for translation is registered in advance. By doing so, the Japanese sentence “My name is Nakano” can be translated into “My name is Nakano” and appropriate English sentences.

  Also, if you register in advance the dictionary entry for translation from English to Japanese, “Nakano” → “Nakano” in the reverse direction, you can use this terminal to translate “Nice to meet you, Nakano” “san” → “I'm glad to see you, Mr. Nakano” can be executed correctly.

  However, when there are a plurality of machine translation systems, the contents registered in the dictionary in one machine translation system are generally not shared with other machine translation systems. Therefore, even if a dictionary entry necessary for translation of a word is registered in a user dictionary of a certain translation system, a translation result using the dictionary entry cannot be obtained by another translation system.

  For example, it is assumed that a Japanese user has a terminal A equipped with a Japanese-English machine translation system, and an American user has a terminal B equipped with an English-Japanese machine translation system. Then, assume a situation in which text is exchanged with the other party by translating Japanese → English and English → Japanese on their own terminals.

  In the above situation, a Japanese user maintains a user dictionary and registers his name. Specifically, in the user dictionary of terminal A, translation dictionary entries of “Nakano” → “Nakano” (Japanese → English) and “Nakano” → “Nakano” (English → Japanese) are registered. And In the terminal A, when a Japanese user inputs “My name is Nakano”, this dictionary entry is used and translated as “My name is Nakano”. However, since the translation system of terminal A and the translation system of terminal B use different dictionaries, even if an American user inputs “Nice to meet you, Nakano san” to terminal B, The dictionary does not have the required dictionary entry, so “Nakano” cannot be correctly translated as “Nakano”.

  As a related technology for sharing a dictionary between terminals, for example, Patent Literature 1 is configured to freely access and use a user dictionary in another device using a communication function between devices. An automatic interpretation (speech translation) device is disclosed. Patent Document 1 has a function of transmitting a user dictionary of a terminal to another terminal, and performs interpretation processing using both a user dictionary unique to the terminal and a user dictionary transmitted from another terminal. An automatic interpretation device is disclosed. In the automatic interpretation apparatus described in Patent Document 1, when translation is performed in a certain apparatus, processing can be performed with reference to the dictionary contents of another apparatus, so a necessary dictionary entry is registered in any apparatus. If so, the translation processing can be executed by referring to this in other apparatuses.

JP 2001-92822 A

  However, the related technique described in Patent Document 1 has a problem that the amount of data transferred between devices is large, the communication cost is high, and the response is adversely affected. In the related technique described in Patent Document 1, when searching a dictionary, a dictionary of another device is accessed. However, generally, a dictionary file is accessed many times during a dictionary search, and a part of the dictionary file is transferred via a communication line. For this reason, a large amount of data transfer occurs between the apparatuses via the communication line, and there is a problem that the communication cost increases and the response deteriorates.

  In the related technique described in Patent Document 1, the entire user dictionary file of a certain terminal is transmitted to another terminal and used for translation processing on the destination terminal. In general, the file size of the dictionary for translation is very large, and in this case as well, a large amount of data is transferred between apparatuses via a communication line. Therefore, similarly, there is a problem that the communication cost increases and the response deteriorates.

  Therefore, the present invention has been made to solve the above-described problem, and it is possible to use dictionary information on another machine translation system during translation processing, and between systems. It is an object to provide a machine translation system, a machine translation method, and a machine translation program that require a small amount of data transfer.

  A machine translation system according to the present invention includes a first translation system and a second translation system, and the first translation system includes a first dictionary storage means for storing a dictionary used for translation, and a first dictionary. Using the dictionary stored in the storage means, the first translation means for translating the input sentence and the dictionary stored in the first dictionary storage means, the dictionary entry used by the first translation means for the translation process Dictionary entry extraction means for extracting the associated dictionary entry from the dictionary stored in the first dictionary storage means, and transmitting the extracted dictionary entry to the second translation system via the communication line, the second translation The system receives second dictionary storage means for storing a dictionary used for translation, and dictionary entries transmitted from the first translation system via a communication line, and the second dictionary storage means receives the received dictionary entries. It characterized in that it comprises a dictionary registration means for registering a dictionary 憶.

  In the machine translation method according to the present invention, the first translation system translates the input sentence using the dictionary stored in the first dictionary storage means, and the first translation system includes the first dictionary storage means. A dictionary entry associated with the dictionary entry used in the translation process is extracted from the dictionary stored in the first dictionary storage means, and the first translation system extracts the extracted dictionary entry. Transmitting to the second translation system via the communication line; the second translation system receives the dictionary entry transmitted from the first translation system via the communication line; Registering in a dictionary stored in the dictionary storage means.

  The machine translation program according to the present invention includes a computer having a dictionary storage means for storing a dictionary used for translation, a process for translating an input sentence using the dictionary stored in the dictionary storage means, and a dictionary stored in the dictionary storage means. The process of extracting the dictionary entry associated with the dictionary entry used in the translation process from the dictionary stored in the dictionary storage means, and the process of transmitting the extracted dictionary entry to another computer via the communication line It is for execution.

  According to the present invention, it is possible to use dictionary information on another translation system during translation processing, and to reduce the amount of data transferred between translation systems.

Embodiment 1. FIG.
Hereinafter, a first embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration example of a machine translation system according to the present invention. As shown in FIG. 1, the machine translation system includes a terminal A10 and a terminal B11.

  Terminal A10 includes an input unit 101A, a translation unit 102A, a dictionary 103A, an output unit 104A, and a dictionary entry extraction unit 105A. The terminal A10 is specifically realized by an information processing terminal such as a personal computer that operates according to a program. Further, the terminal A10 may be realized as a data processing terminal by a mobile terminal such as a mobile phone that operates according to a program.

  Specifically, the input unit 101A is realized by a CPU of an information processing terminal that operates according to a program and an input device such as a keyboard and a key operation unit. The input unit 101A has a function of inputting an original text to be translated in accordance with a user operation.

  Specifically, translation unit 102A is realized by a CPU of an information processing terminal that operates according to a program. The translation unit 102A has a function of executing a translation process on the input original text.

  The dictionary 103A is a translation dictionary that the translation unit 102A refers to in the translation process. The dictionary 103A is specifically stored in a storage device such as a hard disk device or a flash ROM provided in the information processing terminal.

  Specifically, the output unit 104A is realized by a CPU of an information processing terminal that operates according to a program, and a display device such as a display device or a liquid crystal display unit. The output unit 104A has a function of outputting the translation result by the translation unit 102A. For example, the output unit 104A displays the translation result on a display device such as a display device or a liquid crystal display unit. Note that the output unit 104A may output the translation result to a printing apparatus such as a printer, for example. The output unit 104A may be provided with a file output function that outputs a file including the translation result, for example.

  Specifically, the dictionary entry extraction unit 105A is realized by a CPU of an information processing terminal that operates according to a program and a network interface unit. The dictionary entry extraction unit 105A has a function of extracting a dictionary entry from the dictionary 103A with reference to information (translated translated sentence) from the translation unit 102A and the translation dictionary 103A. In the present embodiment, the dictionary entry extraction unit 105A extracts, from the dictionary 103A, a dictionary entry for translating the word in the reverse direction for the word included in the translation translated by the translation unit 102A. The dictionary entry extraction unit 105A has a function of transmitting the extracted dictionary entry to the terminal B11 via the communication line L12.

  The terminal B11 includes an input unit 101B, a translation unit 102B, a dictionary 103B, an output unit 104B, and a dictionary entry registration unit 106B. The terminal B11 is specifically realized by an information processing terminal such as a personal computer that operates according to a program. The terminal B11 may be realized as a data processing terminal by a mobile terminal such as a mobile phone that operates according to a program.

  Specifically, the input unit 101B is realized by a CPU of an information processing terminal that operates according to a program and an input device such as a keyboard and a key operation unit. The input unit 101B has a function of inputting an original text to be translated in accordance with a user operation.

  Specifically, translation unit 102B is realized by a CPU of an information processing terminal that operates according to a program. The translation unit 102B has a function of executing a translation process on the input original text.

  The dictionary 103B is a translation dictionary that the translation unit 102B refers to in translation processing. The dictionary 103B is specifically stored in a storage device such as a hard disk device or a flash ROM provided in the information processing terminal.

  Specifically, the output unit 104B is realized by a CPU of an information processing terminal that operates according to a program, and a display device such as a display device or a liquid crystal display unit. The output unit 104B has a function of outputting a translation result (translation) by the translation unit 102B. For example, the output unit 104B displays the translation result on a display device such as a display device or a liquid crystal display unit. Note that the output unit 104B may output the translation result to a printing apparatus such as a printer, for example. The output unit 104B may be provided with a file output function for outputting a file including the translation result, for example.

  Specifically, the dictionary entry registration unit 106B is realized by a CPU of an information processing terminal that operates according to a program. The dictionary entry registration unit 106B has a function of receiving (receiving) a dictionary entry from the terminal A10 via the communication line L12. The dictionary entry registration unit 106B has a function of additionally registering the received dictionary entry in the dictionary 103B.

  As shown in FIG. 1, in the present embodiment, the terminal A10 and the terminal B11 are connected via a communication line L12 such as the Internet or a wireless communication network (for example, a mobile phone network).

  In the present embodiment, the storage device (not shown) of the terminal A10 stores various programs for translating text and speech. For example, the storage device of the terminal A10 uses the dictionary stored in the dictionary storage unit in the computer to translate the input sentence, and uses the dictionary stored in the dictionary storage unit to store the dictionary entry used in the translation process. A machine translation program for executing a process of extracting the associated dictionary entry from the dictionary stored in the dictionary storage means and a process of transmitting the extracted dictionary entry to another computer via a communication line is stored. .

  In the present embodiment, the storage device (not shown) of the terminal B11 stores various programs for translating text and speech. For example, the storage device of the terminal B11 has a process of registering a dictionary entry transmitted from another computer via a communication line to a computer in a dictionary stored in the dictionary storage unit, and a dictionary stored in the dictionary stored in the dictionary storage unit. After the entry is registered, a machine translation program for executing a process for translating an input sentence using a dictionary entry included in the dictionary stored in the dictionary storage means is stored.

  Next, the operation will be described. FIG. 2 is a flowchart showing an example of processing for extracting and registering dictionary entries in the present embodiment. In addition, the process shown below is performed, for example, when communication by a chat or interactive form is performed between the terminal A10 and the terminal B11. In this case, first, the input unit 101A of the first terminal A10 inputs the original text to be translated in accordance with the user's operation (step SA101). Next, the translation unit 102A of the first terminal A10 translates the input original sentence with reference to the dictionary 103A (step SA102). Then, the output unit 104A of the first terminal A10 outputs the translation result by the translation unit 102A (step SA103).

  Next, the terminal A10 activates the dictionary entry extraction unit 105A. The dictionary entry extraction unit 105A extracts a dictionary entry from the dictionary 103A with reference to the translation result (translation, etc.) translated by the translation unit 102 (step SA104). Then, the dictionary entry extraction unit 105A transmits the extracted dictionary entry to the second terminal B11 via the communication line L12 (step SA105).

  The dictionary entry registration unit 106B of the second terminal B11 receives (receives) a dictionary entry from the terminal A10 via the communication line L12. Then, the dictionary entry registration unit 106B additionally registers the received dictionary entry in the translation dictionary 103B (step SA106).

  Next, the flow of operation in which the second terminal B11 performs translation processing using the dictionary entry registered in the processing of FIG. 2 will be described. FIG. 3 is a flowchart illustrating an example of translation processing in the second terminal B11. It is assumed that a dictionary entry previously transmitted from the first terminal A10 is additionally registered in the dictionary 103B by executing the processing shown in FIG.

  First, the input unit 1010B of the second terminal B11 inputs the original text to be translated according to the user's operation (step SB101). Next, the translation unit 102B of the second terminal B11 performs translation processing with reference to the dictionary 103B (step SB102). Finally, the output unit 104B of the second terminal B11 outputs the translation result by the translation unit 102B (step SB103).

  In the translation process in step SB102, the translation unit 102B translates the original text by referring to and using the translation dictionary entry received from the terminal A10 and additionally registered in the dictionary 103B in step SA106 shown in FIG.

  By executing the above processing, in this embodiment, the dictionary entry extraction unit 105A of the terminal A10 extracts, from the dictionary 103A, a dictionary entry that translates the dictionary entry used in the translation processing in the reverse direction. To work.

  FIG. 4 is an explanatory diagram illustrating an example of a dictionary entry stored in the dictionary 103A according to the present embodiment. Among these, FIG. 4A shows a dictionary entry for translating “Nakano” into “Nakano” in the Japanese → English direction. FIG. 4B is a dictionary entry that translates “Nakano” from “Nakano” in English to Japanese (in the opposite direction to the dictionary entry shown in FIG. 4A).

  In the case of the example illustrated in FIG. 4, it is assumed that the input unit 101A of the terminal A10 inputs “My name is Nakano” in step S101. Next, in step SA102, it is assumed that the translation unit 102A obtains a translation result “My name is Nakano” using the dictionary entry shown in FIG. Next, in step SA103, the output unit 104A outputs the translation generated by the translation unit 102A. After the translation is output, in step SA104, the dictionary entry extraction unit 105A selects a dictionary entry (dictionary entry shown in FIG. 4B) corresponding to the dictionary entry shown in FIG. 4A used in the translation process. Extract. In step SA105, the dictionary entry extraction unit 105 transmits the extracted dictionary entry shown in FIG. 4B to the terminal B11 via the communication line L12. Then, the dictionary entry registration unit 106B of the terminal B11 additionally registers the transmitted dictionary entry shown in FIG. 4B in the dictionary 103B in step SA106.

  Next, in step SB101, it is assumed that the input unit 101B of the second terminal B10 inputs an English sentence “Nice to meet you, Nakano san”. Next, in step SB102, the translation unit 102B performs translation processing with reference to the dictionary 103B. In this case, the dictionary 103B additionally registers a dictionary entry (see FIG. 4B) for translating “Nakano” into “Nakano” in English and Japanese in step SA106 already executed. Therefore, the translation unit 102B can obtain a translation result of “I am glad to meet you, Mr. Nakano” by using this additionally registered dictionary entry.

  As described above, even if a dictionary entry for correctly translating “Nakano” is not registered in the dictionary 103B on the terminal B11, “Nakano” → “Nakano” on the first terminal A10. Triggered by the translation of "", a necessary dictionary entry is received from the first terminal A10, and thereafter, "Nakano" → "Nakano" can be correctly translated on the second terminal B11. .

  Further, as shown above, for each dictionary entry used for translation on terminal A10, a dictionary entry that translates the dictionary entry in the reverse direction is sent to terminal B11 and used for translation on terminal B11. become able to. In other words, when a user of the terminal A10 uses a certain word, the user of the terminal B11 can also translate a sentence including the word. To be precise, the terminal B11 can perform translation in the opposite direction to that of the terminal A10. This is because a user of the terminal A10 and a user of the terminal B11, whose native languages are different languages, are connected to the respective terminals. This means that a word used by one user can be used later by another user in a situation in which a dialogue is performed via the installed translation system.

  In the present embodiment, data exchanged between terminals via the communication line L12 is limited to dictionary entries for translating the dictionary entries effectively used in the translation process in the reverse direction. This is on the same order as the number of words in the input sentence, and the amount of data transferred is much smaller than when the related technique described in Patent Document 1 is used. Therefore, there is an advantage that the communication cost is low and the deterioration of the response accompanying the communication load can be suppressed.

  As described above, according to the present embodiment, the machine translation system includes the first terminal A10 and the second terminal B11. Further, the first terminal A10 extracts from the dictionary 103A a dictionary 103A, a translation unit 102A that translates an input sentence with reference to the dictionary 103A, and a dictionary entry associated with the dictionary entry used for the translation process, And a dictionary entry extraction unit 105A for transmitting via the communication line L12. Further, the second terminal B11 receives the dictionary 103B, the dictionary registration unit 106B that receives the dictionary entry transmitted from the first terminal A10 via the communication line L12, and registers the dictionary entry in the dictionary 103B. The translation part 102B which translates a sentence is included.

  By providing the above configuration, the dictionary entry extraction unit 105A of the first terminal A10 is effectively used in the translation process when the translation process is performed in the first terminal A10, and is reflected in the translation result. For the dictionary entry, the dictionary 103A is referred to, and the dictionary entry associated with the dictionary entry is extracted and transmitted to the second terminal B11. For example, the first terminal A10 extracts a dictionary entry for backward translation corresponding to the dictionary entry effectively used in the translation process. The second terminal B11 receives the dictionary entry extracted from the first terminal A10, and the dictionary entry registration unit 106B additionally registers it in the dictionary 103B. By doing so, the dictionary entry additionally registered in the second terminal B11 can be referred to and usable when machine translation processing is performed in the second terminal B11 thereafter.

  By executing the above processing, if a word is input and translated at the first terminal A10, then the word can be translated in the reverse direction at the second terminal B11.

  For example, it is assumed that a Japanese user has a terminal A10 equipped with a Japanese-English machine translation system, and an American user has a terminal B11 equipped with an English-Japanese machine translation system. A situation is assumed in which texts are exchanged with the other party by translating Japanese → English and English → Japanese on their own terminals. In such a situation, a Japanese user maintains a user dictionary and registers his / her name. Specifically, translation dictionary entries “Nakano” → “Nakano” (Japanese → English) and “Nakano” → “Nakano” (English → Japanese) are stored in advance in the user dictionary of the terminal A. Shall. Also, it is assumed that these Japanese-English and English-Japanese dictionary entries are associated with each other as a reverse translation relationship.

  When a Japanese user inputs “My name is Nakano” at the terminal A10, this dictionary entry is referred to and correctly translated as “My name is Nakano”. In the translation process, the dictionary entry “Nakano” → “Nakano” is used and reflected in the result, so that the dictionary entry extraction unit 105A changes the corresponding dictionary entry “Nakano” → “Nakano” for backward translation. Extracted and transmitted to terminal B11. The transmitted dictionary entry “Nakano” → “Nakano” is registered in the dictionary 103B of the second terminal B11.

  Next, when an American user inputs “My name is Nakano” on the terminal B11, since the necessary dictionary entry is registered in the dictionary 103B of the terminal B11, “Nice to meet you, Nakano san” Can be translated correctly as “I am glad to meet you, Mr. Nakano”.

  Further, according to the present embodiment, the data transferred via the communication line L12 is the number of dictionary entries that is about the same as the number of dictionary entries that are effectively used at the time of translation processing in the first terminal A10. . In other words, the number of dictionary entries equal to the number of words in the original sentence / translation sentence is only transmitted. Therefore, there is an effect that the data transfer amount is very small as compared with the case of using the related technique described in Patent Document 1.

Embodiment 2. FIG.
Next, a second embodiment of the present invention will be described. In the present embodiment, the configuration of the machine translation system is the same as that of the first embodiment shown in FIG. 1 except that the dictionary entry extraction unit 105A transmits the dictionary entry itself used for translation processing via the communication line L12. The configuration is the same as that shown in FIG.

  In the present embodiment, for example, as in the example shown in FIG. 4, the terminal A10 inputs “My name is Nakano” in step SA101, and the translation unit 102A proceeds with the translation processing in step SA102. Assume that a translation “My name is Nakano” is generated using the dictionary entry shown in FIG. In the present embodiment, since the dictionary entry shown in FIG. 4A is used for the translation process, the dictionary entry extraction unit 105A extracts the entry itself used for the translation process from the dictionary 103A. Then, the dictionary entry extraction unit 105A sends (transmits) the extracted dictionary entry to the second terminal B11 via the communication line L12.

  The dictionary entry registration unit 106B of the second terminal B11 additionally registers the received (received) dictionary entry in the dictionary 103B.

  By doing the above, when the Japanese sentence “My name is Nakano” is entered on the terminal B11, the dictionary entry can be used to correctly translate “My name is Nakano”. . In this case, even if the headword “Nakano” is not originally registered in the dictionary 103B, the same dictionary entry can be used also in the terminal B11 as a result of being used in the terminal A10 as described above. is there.

  Further, in the present embodiment, what is sent (transmitted) via the communication line L12 is limited to the dictionary entry used for translation processing at the terminal A10. Therefore, as compared with the case where the related technique described in Patent Document 1 is used, the advantage that the communication amount between terminals is extremely small is obtained as in the first embodiment.

Embodiment 3. FIG.
Next, a third embodiment of the present invention will be described with reference to the drawings. In this embodiment, the configuration of the machine translation system is the same as that of FIG. The configuration is the same as that shown in the first embodiment shown in FIG.

  FIG. 5 is an explanatory diagram illustrating an example of contents (dictionary entries) stored in the dictionary 103A according to the third embodiment. In FIG. 5, FIG. 5A is an example of a dictionary entry for Japanese-English translation that translates “wine” into “wine”. In the example shown in FIG. 5A, a number (JE00000001) for uniquely identifying this entry is described below <id>. Also, <in> and below indicate source language blocks, and <out> and below indicate target language blocks.

  In addition, the description [J] of the source language block indicates that this block is a description related to Japanese. The word heading is “wine” on the line following [J], the part of speech is a noun and a detail. This indicates that the part of speech is a common noun. In addition, the description [E] of the target language block indicates that this block is a description about English, and the line [J] following [E] has the word heading “wine” and its part of speech is NOUN (noun ) And the fine part of speech is cmn (general noun).

  The part starting with <related> is a list of dictionary entry IDs related to this entry. In the example shown in FIG. 5A, four dictionary entries having JE00000010, EJ00000010, JE00000011, and EJ00000011 as IDs are associated with this dictionary entry.

  FIG. 5B is a diagram showing an example of a dictionary entry whose dictionary entry ID is JE00000010. In the dictionary entry shown in FIG. 5B, the source language is Japanese, the target language is English, and the Japanese common noun “Cabernet Sauvignon” is translated into the English common noun “Cabernet Sauvignon”. Represents.

  FIG. 5C shows a dictionary entry for translating the English common noun “Cabernet Sauvignon” into the Japanese common noun “Cabernet Sauvignon” in the opposite direction to the dictionary entry shown in FIG. 5B. The dictionary entry ID is EJ00000010.

  FIG. 5D shows a dictionary entry for translating the Japanese common noun “Pinot Noir” into the English common noun “Pinot Noir”, similar to FIGS. 5A to 5C. This indicates that the entry ID is JE00000011.

  FIG. 5E shows a dictionary entry for translating the English common noun “Pinot Noir” into the Japanese common noun “Pinot Noir”, similar to FIGS. 5A to 5D. This indicates that the entry ID is EJ00000011.

  Further, these dictionary entries shown in FIGS. 5B to 5E are described as related dictionary entries in blocks beginning with <related> in the dictionary entry shown in FIG. Here, as the related dictionary entry, when the word “wine” is input, a word that is highly likely to appear in the subsequent exchange is registered. In the example shown in FIG. 5, sentences containing “Wine” are followed by sentences containing “Cabernet Sauvignon”, “Pinot Noir”, “Cabernet Sauvignon”, and “Pinot Noir”. Is likely to require translation. Therefore, the dictionary entries shown in FIGS. 5A to 5E are registered as related dictionary entries of the dictionary entry (the dictionary entry whose ID is JE00000001) shown in FIG.

  In the terminal B11, the entries corresponding to FIGS. 5B, 5C, 5D, and 5E are not initially registered in the dictionary, and in the initial state, the words “cabernet sauvignon” and “ “Pinot Noir” is an undefined word during Japanese-English translation, and “Cabernet Sauvignon” and “Pinot Noir” are undefined words during English-Japanese translation.

  In the present embodiment, the machine translation system generally operates as follows using the dictionary 103A shown in FIG. First, it is assumed that the input unit 101 of the terminal A10 inputs a sentence “What is wine?” As an input sentence. Then, the translation unit 102A translates with reference to the dictionary 103A.

  During the translation process, the translation unit 102A uses the dictionary entry shown in FIG. 5A to translate “wine” → “wine”, and so on to say “What would you like for wine?” Suppose you get translation results.

  Next, the dictionary entry extraction unit 105A refers to the information (translation result information) obtained from the translation unit 102A and the dictionary 103A, and extracts a dictionary entry to be transmitted to the terminal B11 from the dictionary 103A.

  The dictionary entry extraction unit 105 generally extracts dictionary entries as follows. First, the dictionary entry extraction unit 105 identifies a dictionary entry used to obtain a translation result. Here, the dictionary entry whose ID is JE00000001 shown in FIG. 5A is used to translate “wine” into “wine”. Therefore, the dictionary entry extraction unit 105A specifies the dictionary entry shown in FIG. 5A as one of the dictionary entries to be identified.

  Next, the dictionary entry extraction unit 105 extracts, for each identified entry, a dictionary entry related to the identified entry (that is, a dictionary entry corresponding to the entry ID described in the <related> block. FIG. 5). In the example shown, the <related> block of the dictionary entry shown in Fig. 5A describes four IDs JE00000010, EJ00000010, JE00000011, and EJ00000011. Therefore, the dictionary entry extraction unit 105A Four dictionary entries corresponding to the IDs (specifically, dictionary entries corresponding to FIGS. 5B, 5C, 5D, and 5E) are extracted from the dictionary 103A.

  Then, the dictionary entry extraction unit 105A transmits the extracted dictionary entry to the terminal B11 via the communication line L12. In the example shown in FIG. 5, the dictionary entry extraction unit 105 </ b> A is combined with the dictionary entries extracted corresponding to the other dictionary entries used in the translation, as shown in FIGS. The dictionary entry shown in e) is transmitted to the terminal B11.

  The dictionary entry registration unit 106B of the terminal B11 additionally registers the dictionary entry transmitted from the terminal A10 in the dictionary 103B on the terminal. In the example shown in FIG. 5, the dictionary entry registration unit 106B additionally registers the dictionary entries shown in FIGS. 5 (b), (c), (d), and (e).

  Next, it is assumed that the terminal B11 inputs “My favorite is Pinot Noir.” In accordance with a user operation. In this example, translation is performed using the dictionary entry additionally registered in the dictionary 103B by the processing described above, and translation is performed appropriately as “Pino Noir is a preference”.

  For example, in the dictionary 103B, no dictionary entry is registered for “Pinot Noir” in the initial state, so this word is an unregistered word in the initial state, and appropriate translation cannot be performed. However, in accordance with the processing described above, the word “Wine” is translated on the terminal A10, and the dictionary entry for Japanese-English translation corresponding to the related words “Cabernet Sauvignon” and “Pinot Noir” ( English-Japanese translation dictionary entries (see FIGS. 5D and 5E) for “Cabernet Sauvignon” and “Pinot Noir” are extracted. Then, it is transmitted to the terminal B11 and additionally registered in the dictionary 103B, so that the terminal B11 can correctly translate sentences including these words.

  As described above, according to the present embodiment, when the user of the terminal A10 inputs and translates “What is the wine?”, The terminal A10 transfers the information to the terminal B11, as shown in FIGS. ) Is transmitted, and is additionally registered in the translation dictionary on the terminal B11. Then, a response using the related dictionary entry (for example, “My favorite is Pinot Noir.”) Can be appropriately translated on the terminal B11.

  According to the present embodiment, when a certain user uses a certain word in a situation where a plurality of users are interacting with each other via the translation system installed in each terminal, the word related to the word is displayed. The dictionary entry is transmitted to another terminal, and can be used for translation processing at that terminal. Therefore, even if the topic spreads with clues from words uttered by someone, the dictionary entries of the related words that are necessary there are additionally registered in the dictionary of each terminal, so dialogue after the topic spreads Translated correctly.

  Compared with the case where the related technique described in Patent Document 1 is used, the amount of data transfer is small, and there is an advantage that an increase in communication cost and response deterioration can be prevented. It is the same as the case of the form.

  Next, embodiments of the present invention will be described with reference to the drawings. FIG. 6 is a block diagram illustrating a more specific configuration example of the machine translation system in the present embodiment. The machine translation system shown in the present example corresponds to the machine translation system shown in the first embodiment.

  As shown in FIG. 6, in this embodiment, the machine translation system includes a first interpreting terminal A 1301a and a second interpreting terminal B 1301b that perform translation by voice input. Interpreter terminal A 1301a and interpreter terminal B 1301b are connected via a wireless LAN 1318 that operates as a communication line connecting the terminals. As shown in FIG. 6, the first interpreting terminal A 1301a and the second interpreting terminal B 1301b have the same configuration regarding hardware and software.

  First, as shown in FIG. 6, the first interpretation terminal A 1301a includes a CPU 1302a, a RAM 1303a, a flash ROM 1304a, a microphone 1314a, a speaker 1315a, an A / D / D / A converter 1305a, and a wireless LAN control. Part 1306a. The CPU 1302a operates by program control. The RAM 1303a is a readable / writable memory. The flash ROM 1304a is a nonvolatile read / write memory. The microphone 1314a is a device for inputting sound. The speaker 1315a is a device that outputs sound.

  The A / D / D / A conversion unit 1305a converts the sound input from the microphone 1314a into digital data. The A / D / D / A converter 1305a converts the digital data into analog data and sends (outputs) the analog data to the speaker 1315a. The wireless LAN control unit 1306a has a network connection function using a wireless LAN.

  The CPU 1302a, the RAM 1303a, the flash ROM 1304a, the microphone 1314a, the speaker 1315a, the A / D / D / A conversion unit 1305a, and the wireless LAN control unit 1306a connect the units in the interpretation terminal A 1301a to exchange data within the terminal. They are connected via an address bus / data bus 1317a which is a communication path to be enabled.

  The RAM 1303a is a machine translation subprogram 1308a that implements a machine translation function, a dictionary entry extraction subprogram 1309a, a dictionary entry registration subprogram 1310a, a speech recognition synthesis subprogram 1311a, and an overall control that controls the entire process. The subprogram 1307a is stored. The flash ROM 1304a stores a system dictionary 1312a and a user dictionary 1313a used in machine translation processing.

The second interpreting terminal B 1301b has the same configuration as the interpreting terminal A 1301a.
That is, the communication terminal B 1301b includes a CPU 1302b, a RAM 1303b, a flash ROM 1304b, a microphone 1314b, a speaker 1315b, an A / D / D / A converter 1305b, and a wireless LAN controller 1306b. The functions of these components are the same as the functions of the CPU 1302a, RAM 1303a, flash ROM 1304a, microphone 1314a, speaker 1315a, A / D / D / A converter 1305a, and wireless LAN controller 1306a in the communication terminal A 1301a. The CPU 1302b, the RAM 1303b, the flash ROM 1304b, the microphone 1314b, the speaker 1315b, the A / D / D / A conversion unit 1305b, and the wireless LAN control unit 1306b are connected via an address bus / data bus 1317b.

  In interpreter terminal A 1301a, CPU 1302a executes processing in accordance with the control of overall control subprogram 1307a. First, the CPU 1302a controls the A / D / D / A conversion unit 1305a to acquire (input) sound from the microphone 1314a and execute A / D conversion. The CPU 1302a temporarily stores the audio data digitized by A / D conversion in the RAM 1303a via the address bus / data bus 1317a.

  Next, the CPU 1302a activates the speech recognition / synthesis subprogram 1311a, and executes speech recognition processing using the speech data temporarily stored in the RAM 1303a as input under the control of the speech recognition / synthesis subprogram 1311a.

  Next, the CPU 1302a starts the machine translation subprogram 1308a with the text of the speech recognition result as an input. The CPU 1302a executes machine translation processing with reference to the system dictionary 1312a and the user dictionary 1313a according to the control of the machine translation subprogram 1308a, and obtains a machine translation result.

  Next, the CPU 1302a starts the speech recognition synthesis subprogram 1311a with the obtained machine translation result text as an input. Then, the CPU 1302a executes speech synthesis processing under the control of the speech recognition / synthesis subprogram 1311a to obtain speech data.

  Next, the CPU 1302a sends the obtained audio data to the A / D / D / A conversion unit 1305a via the address bus / data bus 1317a under the control according to the instruction of the overall control subprogram 1307a (output) To do). Then, the A / D / D / A converter 1305a performs D / A conversion on the audio data sent from the CPU 1302a, and sends (outputs) the analog data as analog data to the speaker 1315a. The speaker 1315a converts the analog voice data sent from the A / D / D / A converter 1305a into air vibrations and outputs it as voice.

  Further, after executing the machine translation process, the CPU 1302a activates the dictionary entry extraction subprogram 1309a in parallel with the speech synthesis / output process described above. Under the control of the dictionary entry extraction subprogram 1309a, the CPU 1302a extracts a reverse translation dictionary entry corresponding to the dictionary entry used for translation from the system dictionary 1312a and the user dictionary 1313a.

  Next, the CPU 1302a sends (outputs) the dictionary entry extracted in accordance with the dictionary entry extraction subprogram 1309a to the wireless LAN control unit 1306a. The wireless LAN LAN control unit 1306a transmits the dictionary entry extracted by the CPU 1302a to the interpreting terminal B1301b via the network connection by the wireless LAN 1318 established with the interpreting terminal B1301b via the antenna 1316a.

  Interpretation terminal B 1301b receives dictionary entry data using wireless LAN control unit 1306b. Then, the CPU 1302b executes processing for registering the received dictionary entry data in the user dictionary 1313b in accordance with the control of the dictionary entry registration subprogram 1310b.

  Thereafter, the CPU 1302b refers to and uses the additionally registered dictionary entry when executing the interpreting process according to the speech recognition synthesis subprogram 1311b and the machine translation subprogram 1308b of the interpreting terminal B1301b.

  Next, the operation of the machine translation system in this embodiment will be described using a specific input example. In this example, it is assumed that the user A of the interpretation terminal A 1301a is Japanese and the user B of the interpretation terminal B 1301b is American. The interpreter terminal A 1301a is set to perform an interpreter process (speech translation process) from Japanese to English, and the interpreter terminal B 1301b is set to perform an interpreter process from English to Japanese. To do.

  In addition, it is assumed that the Japanese user A has previously registered the entries shown in FIGS. 7A and 7B in the user dictionary. In this example, the dictionary entry in FIG. 7A has JE00000001 as an ID, and the dictionary entry shown in FIG. 7B has EJ00000001 as an ID. In the dictionary entry shown in FIG. 7, based on the dictionary entry for text translation shown in FIG. 4, the pronunciation information in Japanese and English is hiragana in the case of Japanese, and the phonetic symbol in the case of English. Are additionally described as the value of the “read” field. These pronunciation information is used when executing speech recognition processing for the source language and speech synthesis processing for the target language.

  Each dictionary entry indicates the ID of a reverse translation dictionary entry corresponding to the dictionary entry as a value of <inv>. The reverse translation dictionary entry corresponding to the dictionary entry in FIG. 7A is a dictionary entry whose ID is EJ00000001 (that is, the dictionary entry shown in FIG. 7B). Further, the reverse translation dictionary entry corresponding to the dictionary entry in FIG. 7B is a dictionary entry having an ID of JE00000001 (that is, the dictionary entry shown in FIG. 7A).

  Next, how the process is executed using the dictionary shown in FIG. 7 in this embodiment will be described. First, it is assumed that the Japanese user A speaks to the microphone 1314a of the interpreter terminal A1301a, "I am a name". The interpretation terminal A 1301a performs A / D conversion on the voice picked up (input) from the microphone 1314a, and uses it as an input for processing executed in accordance with the voice recognition synthesis subprogram 1311a. Perform recognition processing. Then, the interpretation terminal A 1301a converts the text “My name is Nakano” as a result of the speech recognition. Interpretation terminal A 1301a uses the reading information of the user dictionary entry shown in FIG. 7A in speech recognition processing to convert the utterance “Naka” into the text “Nakano”.

  Next, the CPU 1302a of the interpretation terminal A 1301a receives the text “My name is Nakano” as the speech recognition result, and refers to the user dictionary entry shown in FIG. 7A according to the control of the machine translation subprogram 1308a. Proceed (execute) the translation process and obtain the translation result “My name is Nakano”. The CPU 1302a uses the English text of the translation result as an input of processing executed in accordance with the speech recognition synthesis subprogram 1311a, and performs speech synthesis processing with reference to the system dictionary 1312a and the user dictionary 1313a. Interpretation terminal A 1301a performs D / A conversion on the speech synthesis result in A / D / D / A conversion section 1305a and outputs the result from speaker 1315a.

  Next, after performing the translation process, the CPU 1302a extracts a reverse dictionary entry corresponding to the dictionary entry used for translation under the control of the dictionary entry extraction subprogram 1309a. Specifically, the CPU 1302a extracts (extracts) a number described as a value corresponding to <inv> from the dictionary entry used for translation, and extracts a dictionary entry having the extracted number as <id>. To do.

  In the example shown above, since the dictionary entry shown in FIG. 7A is used for translation, the CPU 1302a stores the value corresponding to <inv> in the dictionary entry shown in FIG. 7A. Referring to EJ00000001, a dictionary entry having the same <id> value is extracted from the dictionary (system dictionary 1312a and user dictionary 1313a). In the example shown in FIG. 7, since the <id> in the dictionary entry shown in FIG. 7B, which is one entry in the system dictionary, is EJ00000001, the CPU 1302a extracts this dictionary entry together with other entries.

  Next, the CPU 1302a controls the wireless LAN control unit 1306a to transmit the extracted dictionary entry to the interpretation terminal B 1301b via the network connection by the wireless LAN 1318.

  Next, when the CPU 1302b of the interpretation terminal B 1301b receives (receives) the dictionary entry shown in FIG. 7B or the like, it additionally registers the received dictionary entry in the user dictionary 1313b according to the control of the dictionary entry registration subprogram 1310b. Process.

  Thereafter, it is assumed that the American user B utters “Nice to meet you, Nakano san” toward the microphone 1314b of the interpretation terminal B 1301b. Interpretation terminal 1301b inputs user B's utterance from microphone 1314b and performs A / D conversion using A / D / D / A conversion section 1305b. Next, the CPU 1302b performs voice recognition processing on the input voice data in accordance with the control of the voice recognition subprogram 1311b. When executing this voice recognition process, the CPU 1302b refers to the user dictionary 1313b and uses the reading information of “Nakano” received and registered from the interpretation terminal A 1301a.

  Next, when the CPU 1302b obtains the text “Nice to meet you, Nakano san” as the speech recognition result, it performs English-Japanese machine translation processing according to the control of the machine translation subprogram 1308b. In the machine translation process, the CPU 1302b uses the dictionary entry received and registered from the interpretation terminal A 1301a (see FIG. 7B) to appropriately translate “I am glad to meet you, Mr. Nakano”. be able to.

  Thereafter, the CPU 1302b performs speech synthesis processing on the translation result according to the control of the speech recognition synthesis subprogram 1311b. Interpretation terminal B 1301b converts the synthesized speech into analog data by D / A conversion, and outputs the analog data from speaker 1315b.

  As described above, in this embodiment, the data sent through the network connection by the wireless LAN is only a part of the dictionary entry of the translation dictionary. Further, as can be seen from the example of dictionary entries including FIG. 7, the amount of data per dictionary entry is relatively small. In addition, since the number of dictionary entries to be transmitted (transmitted) is approximately the same as the number of words in the translated sentence, the amount of transmitted data is relatively small as a whole.

  For example, if the average data amount per dictionary entry is 100 bytes, the translation contains 20 words, and 20 dictionary entries are transmitted, the total amount of data transmitted at one time is 20 ×. 100 = only 2000 bytes. For this reason, even if the network is very slow, for example, there is only a communication speed of 64 Kbps (bit per second), the transmission of the dictionary entry can be completed in about 0.3 seconds.

  On the other hand, when a technique for transmitting the entire dictionary file via a network connection is used, it is necessary to transmit the entire dictionary having a normal size of several hundred KB to several tens of MB. Therefore, the amount of data to be transmitted is much larger than that of the machine translation system shown in this embodiment. Accordingly, the communication cost is high and the response is deteriorated.

  In addition, the amount of data flowing on the network is much smaller in the machine translation system shown in the present embodiment than in the case of using the technique of directly accessing the dictionary through the network connection during the analysis process. . As described above, in this embodiment, since the amount of data flowing on the network is small as compared with the related technology described in Patent Document 1, it is possible to suppress communication costs and avoid adverse effects on the response of the system. effective.

  As described above, it is apparent that the same effect as the dictionary sharing effect that can be expected in the related art described in Patent Document 1 can be obtained.

  In the present embodiment, the case of performing speech translation for translating speech into speech has been described as an example. However, the machine translation system can also be applied according to similar processing in the case of text translation for translating text into text. Specifically, each terminal is equipped with a text input / output function such as a keyboard and a display device, and instead of (or even with) a microphone, a speaker, and an A / D / D / A converter. Do not use). Further, each terminal does not need to start a speech recognition synthesis program, and may perform translation processing according to a machine translation subprogram directly on the input text and output the translation result as text. The operation in which the CPU of each terminal executes processing according to the dictionary entry extraction subprogram and the dictionary entry registration subprogram is the same as the operation shown in the present embodiment. In this case, the pronunciation information of the source language and the target language is not necessary as the information content included in each dictionary entry stored in the dictionary.

  In the present embodiment, the case where the source language is Japanese and the target language is English has been described. However, the language to be translated is not limited to these, and the machine translation system uses any language. You may make it perform a translation process as object.

  In this embodiment, it is assumed that the input information is text or speech. However, for example, a machine translation system can be easily realized by modifying text input on paper as input. . In this case, the machine translation system may be configured, for example, to perform character recognition from an image obtained by reading a printed text with a camera and perform a translation process on the obtained text. Similarly, regarding the output, the machine translation system may display the text on a CRT display, or may output the text as a speech. In addition, for example, a machine translation system can be easily configured to output in various formats such as printing on paper and presenting the output.

  In addition, the machine translation system may be modified so that translation from the source language text into a plurality of target languages is performed simultaneously, and dictionary entries are extracted corresponding to each translation result.

  Moreover, although the case where the machine translation system includes two terminals has been described in the present embodiment, the machine translation system may include three or more terminals. For example, a dictionary entry extracted at a certain terminal may be modified to be transmitted to all other terminals. In this case, the machine translation system may extract a reverse translation entry as shown in the first embodiment as a dictionary entry to be extracted. In addition, as shown in the second embodiment, the machine translation system may extract the dictionary entry itself used for translation. Further, as shown in the third embodiment, the machine translation system may be configured to extract and transmit related entries of dictionary entries used for translation.

  As described above, the machine translation system can be variously modified and configured without changing the spirit of the machine translation system.

  Next, the minimum configuration of the machine translation system according to the present invention will be described. FIG. 8 is a block diagram illustrating a minimum configuration example of the machine translation system. As shown in FIG. 8, the machine translation system includes a first terminal A10 and a second terminal B11 as the minimum components. Further, the first terminal A10 includes a mass storage device 603A that stores the dictionary 103A, a translation unit 102A, and a dictionary entry extraction unit 105A. The second terminal B11 includes a mass storage device 603B that stores the dictionary 103B and a dictionary entry registration unit 106B.

  In the machine translation system with the minimum configuration shown in FIG. 8, the large-capacity storage device 603A stores a dictionary 103A used for translation. The translation unit 102A has a function of translating an input sentence using the dictionary 103A stored in the mass storage device 603A. The dictionary entry extraction unit 105A uses the dictionary 103A stored in the large-capacity storage device 603A, and uses the dictionary 103A stored in the large-capacity storage device 603A to store dictionary entries associated with the dictionary entries used by the translation unit 102A in the translation process. A function is provided for extracting and transmitting the extracted dictionary entry to the second terminal B11 via the communication line L12.

  The mass storage device 603B stores a dictionary 103B used for translation. The dictionary entry registration unit 106B has a function of receiving a dictionary entry transmitted from the first terminal A10 via the communication line L12 and registering the received dictionary entry in the dictionary 103B stored in the mass storage device 603B.

  According to the machine translation system with the minimum configuration shown in FIG. 8, it is possible to use dictionary information on another translation system during translation processing, and the amount of data transferred between translation systems is small. Can be done.

  In each of the above embodiments and examples, characteristic configurations of the machine translation system as shown in the following (1) to (6) are shown.

  (1) The machine translation system includes a first translation system (for example, terminal A10) and a second translation system (for example, terminal B11), and the first translation system uses a dictionary (for example, for example) First translation for translating an input sentence using first dictionary storage means (for example, realized by mass storage device 603A) for storing dictionary 103A) and dictionary stored in first dictionary storage means First, a dictionary entry associated with the dictionary entry used by the first translation means for the translation process is stored in the first dictionary by means (for example, realized by the translation unit 102A) and the dictionary stored in the first dictionary storage means. The dictionary entry extracting means (for example, a dictionary entry extracting means for transmitting the extracted dictionary entry to the second translation system via the communication line (for example, communication line L12)). The second translation system is realized by second dictionary storage means (for example, large capacity storage means 603B) for storing a dictionary (for example, dictionary 103B) used for translation. A dictionary entry means for receiving a dictionary entry transmitted from the first translation system via the communication line and registering the received dictionary entry in a dictionary stored in the second dictionary storage means (for example, a dictionary) And is realized by the entry registration unit 106B).

  (2) In the machine translation system, the second translation system uses a dictionary entry included in the dictionary stored in the second dictionary storage unit to use a second translation unit that translates the input sentence (for example, the translation unit 102B). May be included).

  (3) In the machine translation system, the first translation means translates the input sentence from the first language (for example, Japanese) into the second language (for example, English), and the dictionary entry extraction means has the first The dictionary entry for translating from the second language to the first language is extracted from the dictionary stored in the first dictionary storage means with respect to the dictionary entry used by the translation means of It may be configured.

  (4) In the machine translation system, the first translation unit translates the input sentence from the first language into the second language, and the dictionary entry extraction unit includes the dictionary entry used by the first translation unit for the translation process. It may be configured to extract itself from the dictionary stored in the first dictionary storage means.

  (5) In the machine translation system, the second dictionary storage means stores a system dictionary (for example, the system dictionary 1312b) that is a dictionary storing general vocabulary and a user dictionary (for example, a system dictionary, for example) 1313b), and the second translation unit may be configured to translate the input sentence using the system dictionary and the user dictionary stored in the second dictionary storage unit.

  (6) In the machine translation system, the dictionary registration unit may be configured to register the dictionary entry received from the first translation system in a user dictionary stored in the second dictionary storage unit.

  The present invention is applicable to uses such as a machine translation device that translates a first language into a second language, a program for installing a machine translation function in a computer, and the like. Further, the present invention can be applied to uses such as an automatic interpretation terminal equipped with a speech translation function and a program for installing a speech translation function in a portable terminal. The present invention also provides a translation server or interpreter server having a function of receiving text or speech from a mobile terminal, performing text translation or speech translation, and returning it to the mobile terminal, or a program for mounting such a function on a server computer. It is applicable to uses such as.

It is a block diagram which shows the structural example of the machine translation system by this invention. It is a flowchart which shows an example of the process which extracts and registers a dictionary entry. It is a flowchart which shows an example of the translation process in a 2nd terminal. It is explanatory drawing which shows an example of the dictionary entry which a dictionary stores. It is explanatory drawing which shows an example of the dictionary entry which the dictionary in 3rd Embodiment stores. It is a block diagram which shows the more specific structural example of a machine translation system. It is explanatory drawing which shows the example of the dictionary content in an Example. It is a block diagram which shows the minimum structural example of a machine translation system.

Explanation of symbols

10 Terminal A
11 Terminal B
101A, 101B Input unit 102A, 102B Translation unit 103A, 103B Dictionary 104A, 104B Output unit 105A Dictionary entry extraction unit 106B Dictionary entry registration unit

Claims (14)

A first translation system and a second translation system;
The first translation system includes:
First dictionary storage means for storing a dictionary used for translation;
First translation means for translating an input sentence using a dictionary stored in the first dictionary storage means;
Using the dictionary stored in the first dictionary storage means, the dictionary entry associated with the dictionary entry used in the translation process by the first translation means is extracted from the dictionary stored in the first dictionary storage means. Dictionary entry extraction means for transmitting the extracted dictionary entry to the second translation system via a communication line,
The second translation system includes:
A second dictionary storage means for storing a dictionary used for translation;
A dictionary registration unit that receives a dictionary entry transmitted from the first translation system via a communication line and registers the received dictionary entry in a dictionary stored in the second dictionary storage unit. A machine translation system.   2. The machine translation system according to claim 1, wherein the second translation system includes second translation means for translating an input sentence using a dictionary entry included in a dictionary stored in the second dictionary storage means. The first translation means translates the input sentence from the first language to the second language,
The dictionary entry extraction unit is configured to convert a dictionary entry for translating from the second language to the first language in the reverse direction with respect to the dictionary entry used by the first translation unit for translation processing. The machine translation system according to claim 1 or 2, wherein the dictionary is extracted from a dictionary stored in the dictionary storage means. The first translation means translates the input sentence from the first language to the second language,
The dictionary entry extraction unit extracts the dictionary entry itself used for the translation process by the first translation unit from a dictionary stored in the first dictionary storage unit. The machine translation system described in 1. The second dictionary storage means stores a system dictionary, which is a dictionary storing general vocabulary, and stores a user dictionary separately from the system dictionary,
5. The machine translation according to claim 1, wherein the second translation unit translates the input sentence using a system dictionary and a user dictionary stored in the second dictionary storage unit. system.   6. The machine translation system according to claim 5, wherein the dictionary registration means registers the dictionary entry received from the first translation system in a user dictionary stored in the second dictionary storage means. A first translation system translating an input sentence using a dictionary stored in a first dictionary storage means;
The first translation system uses the dictionary stored in the first dictionary storage unit to extract a dictionary entry associated with the dictionary entry used in the translation process from the dictionary stored in the first dictionary storage unit. And steps to
The first translation system transmitting the extracted dictionary entry to the second translation system via a communication line;
A second translation system receiving a dictionary entry transmitted from the first translation system via a communication line, and registering the received dictionary entry in a dictionary stored in a second dictionary storage means; A machine translation method comprising:   The machine translation method according to claim 7, wherein the second translation system includes a step of translating the input sentence using the dictionary entry included in the dictionary stored in the second dictionary storage unit. A first translation system translating the input sentence from the first language to the second language;
The first dictionary storage means stores a dictionary entry for the first translation system to translate the dictionary entry used for translation processing from the second language to the first language in the reverse direction. The machine translation method according to claim 7, further comprising a step of extracting from a dictionary to be performed. A first translation system translating the input sentence from the first language to the second language;
The said 1st translation system includes the step which extracts the dictionary entry itself used for the translation process from the dictionary which a 1st dictionary memory | storage means memorize | stores. Machine translation method. In a computer having a dictionary storage means for storing a dictionary used for translation,
A process of translating an input sentence using a dictionary stored in the dictionary storage means;
Using the dictionary stored in the dictionary storage means, extracting a dictionary entry associated with the dictionary entry used in the translation process from the dictionary stored in the dictionary storage means;
A machine translation program for executing a process of transmitting the extracted dictionary entry to another computer via a communication line. In a computer having a dictionary storage means for storing a dictionary used for translation,
Processing for registering a dictionary entry transmitted from another computer via a communication line in a dictionary stored in the dictionary storage means;
A machine translation program for executing a process of translating an input sentence using a dictionary entry included in a dictionary stored in the dictionary storage means. On the computer,
A process of translating the input sentence from the first language into the second language;
The dictionary entry used for the translation process is subjected to a process of extracting a dictionary entry for translating from the second language to the first language in the reverse direction from the dictionary stored in the dictionary storage means. 11. The machine translation program according to 11. On the computer,
A process of translating the input sentence from the first language into the second language;
The machine translation program according to claim 11 or 12, wherein a dictionary entry used for translation processing is extracted from a dictionary stored in a dictionary storage means.

Images