JP3315221B2 - Conversation sentence translator - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for translating a conversation sentence in a source language into a conversation sentence in a target language efficiently.

[0002]

2. Description of the Related Art In example-based translation, which is attracting attention as a translation method for conversational sentences, etc., a large number of example sentences in a source language and a target language are stored in a bilingual form in advance, and an input sentence and an example sentence in the source language are stored. Translation is performed by judging the similarity.

In this method, the translation performance can be improved only by collecting bilingual translations, but on the other hand, a large number of bilingual translation examples are required, which makes retrieval difficult. Furthermore, it is difficult to calculate the similarity between the input sentence and the example sentence in the source language.

Further, a technique has been proposed in which a bilingual example is abstracted in the form of a pattern, and an input sentence in a source language is converted into a sentence in a target language based on the correspondence between the patterns. However, special expressions that cannot be made into patterns using this method need to be registered in parallel.

[0005]

Accordingly, the number of translation examples or translation patterns handled in any case becomes enormous,
It is difficult to efficiently search for the optimal pattern or bilingual example that matches the input sentence.

Further, when there is a correspondence from an example sentence or pattern of one source language to an example sentence or pattern of a plurality of target languages due to the ambiguity of the source language, an example sentence or pattern of any target language is used. Since it is difficult to judge whether it is OK or not, it is necessary to force the user to perform complicated operations such as inquiring one by one.

In example-based translation, it is necessary to replace words in an example sentence of a target language based on a difference between an input sentence and an example sentence of a source language. It is difficult to select an appropriate word from words. In the case of translation by pattern matching, the same problem occurs when selecting a target language word to be put in a space.

[0008] An object of the present invention is to provide information on scenes (such as airports, hotels, food, shopping, vehicles, etc.) in which a conversational sentence to be translated is used at the time of translation as scene information. It is an object of the present invention to solve the above-mentioned problem by adding a scene for a current conversation in advance at the time of translation and adding it to each translation.

The present invention also provides a means for eliminating the trouble of inputting scene information of a conversation for each user.

[0010]

According to the present invention, an object of the present invention is to provide an input means for inputting a spoken sentence in a source language, a word dictionary, and a method for inputting the source language by referring to the word dictionary. Analysis means for analyzing a conversation sentence, scene information input means for inputting scene information to set a conversation scene, management means for storing and managing the input scene information, and display means for displaying the input scene information. Using the bilingual database of the source language conversation sentence and the target language conversation sentence to which the scene information is added, and the input scene information and the bilingual database, the source language conversation sentence is converted into the target language conversation sentence. and converting means, and output means for outputting the conversational sentence of the transformed target language, each source language word in the word dictionary
Typical scene information to be used is added,
The conversion means cancels the initial state or scene information
If no scene information has been entered by
The current conversation scene from words in the spoken language of the source language
This is achieved by a conversational sentence translator characterized by performing a conversion .

[0011]

According to the apparatus of the present invention, scene information is input, and the conversion means converts the scene information input through the scene information input means and the bilingual database to which the scene information is added to set a conversation scene. To convert the source language conversation sentence input by the input means into the target language conversation sentence. Here, each word in the source language in the word dictionary is used
Typical scene information to be added
The means is for canceling the initial state or scene information.
If scene information has not been input, the
The current conversation scene is estimated from words in the language
Each word that composes the input translation target sentence
Can be used to estimate the current conversation scene. This
As a result, the scene information will be
Even if it is not set by
Translation is possible.

[0012]

[0013]

[0014]

[0015]

[0016]

According to a further preferred feature of the device according to the invention, if no scene information has been added to the input word or the selected translation, the current scene information is added to this word or translation, and The added scene information is used from the time of translation. It is difficult to add typical scene information to all correspondences between words in the source language of the word dictionary and translations of the target language. In that case, scene information is added only to the correspondence with the word or translation that is considered important, and for the remaining words, the set conversation scene when the word or translation is used for input or translation is added as scene information. By doing so, scene information can be automatically added to many words and translations.

According to still another preferred feature of the apparatus according to the present invention, a plurality of scene information is added to each word or each correspondence between a word and a translated word in the word dictionary, and a word or a translated word is extracted from a plurality of scenes. Selected and used,
It is possible to construct a dictionary having more accurate scene information.

According to still another preferred feature of the apparatus according to the present invention, scene frequency information is added to the input word or the selected translated word, and the conversion means performs the above-mentioned addition from the conversion at the next translation. Scene frequency information is also used. By using the frequency information when there is a plurality of pieces of scene information set in the word dictionary, evaluation based on the frequency is performed even when different scene information is set in a plurality of input words. This may allow for more accurate translation.

[0020]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below in detail with reference to preferred embodiments shown in the drawings. Here, we consider Japanese as the source language and English as the target language.

FIG. 1 is a schematic block diagram showing a conversational sentence translator according to an embodiment of the present invention.

In FIG. 1, reference numeral 1 denotes an input unit serving as a means for inputting a conversation sentence in the source language to be translated, 2 denotes an analysis unit for analyzing the input conversation sentence in the source language, and 3 denotes an analysis unit. Word dictionary with scene information mainly used
4 is a scene information input unit as means for inputting scene information, 5 is a scene information management unit for storing and managing the inputted scene information, 6 is a bilingual database to which mainly used scene information is added, 7 is A conversion unit for generating the internal structure of the conversational sentence of the target language by referring to the scene information of the scene information managing unit 5 and the bilingual database 6 based on the analysis result of the conversational sentence of the source language, and 8 converting the internal structure of the target language A generating unit for obtaining a surface expression; an output unit for outputting a conversation sentence in a target language;
Is a scene information display unit for displaying current scene information held by the scene surface management unit 5.

The operation of the embodiment of the present invention will now be described with reference to FIG.
This will be described with reference to the flowchart of FIG.

In step S1, a prompt for prompting the user to input is displayed on the input unit 1, and an input of a conversation sentence in the source language to be translated or current scene information is accepted. In step S2, it is determined whether the character string input by the user is an input of a conversation sentence in the source language or scene information, and the processing is sorted. Whether the input is a source language conversation sentence or scene information input is determined from the state set by the mode switching switch or the like.

If the input is scene information, the information is set in the scene information management section 5 (step S3) and the contents are displayed on the scene information display section 10 (step S4). For example, if the current mode is scene information input and the user inputs "shopping", the system determines that the input is scene information input, sends the data to the scene information management unit 5, and sends the scene data to the scene information management unit 5. "Shopping" is displayed on the information display unit 10.

On the other hand, when a conversation sentence in the source language is input, the analysis unit 2 analyzes the sentence in step S5 with reference to the word dictionary, and converts the character string in the source language into a series of words.

In step S6, the presence / absence of scene information is checked. If there is scene information, the conversion unit 7 first searches the bilingual database 6 only for the database belonging to the set scene information. (Step S7).

It is assumed that the following bilingual data and scene information exist in the bilingual database 6.

<Japanese><English><Sceneinformation> How much do you charge for it? <Shopping> How much is the fare (to LOC)? <Vehicle> At this time, if the scene information is “shopping” in response to the input “how much”, the target language expression “How”
much do you charge forit
? ”, And similarly in the information scene“ Vehicles ”
"What is the Fare (to LOC)
? Is searched.

The conversation scene information is set to reflect the "scene" where the conversation is actually used, such as "airport", "hotel", "meal", "shopping", and "vehicle".

It is not realistic to add scene information to all bilingual data. This is because there are expressions that are effective in all situations, such as greetings. In that case, no scene information is added as described below.

[0032] <Japanese><English><sceneinformation> Hello: Hello. <> Thank you: Thank you. <> If the conversation scene information has not been set by the user, a search is performed from the beginning on the entire bilingual database (step S8).

In step S11, if the search is successful, the search result is sent to the generating unit 8, and if the search is unsuccessful, the end of step S14 is determined. If the search fails as a result of searching only the scenes, the search is again performed for all databases (step S9). If the search is successful, the search result is sent to the generation unit 8.

As a result of the search, if the structure of the conversation sentence in the target language is obtained, the conversation in the target language is generated by the generation unit 8 in step S10, and the result is output to the output unit 9. If the structure of the search result is patterned, a word (translated word) in the target language corresponding to the word in the source language is determined using the scene information.

For example, when the word dictionary 3 is set as follows and a translation of the word "fishing" is determined, <Japanese><English><Sceneinformation> Fishing: change <Shopping>: fishing <Hobby> If the current scene information is “shopping,” the translated word “chan”
“ge” and “fishing” for “hobby”.

Subsequently, it is determined whether the bilingual data used for translation or the scene information associated with the word or translation in the word dictionary is different from the currently set scene (step S12). S13
Changes the setting of the scene information.

If there is no need to change, step S14
To determine whether or not to end the translation work, and if it is to be continued, return to step S1 and repeat the above algorithm.

FIG. 3 shows a screen image diagram of the conversational sentence translator according to the embodiment of the present invention.

[0039]

As described above, according to the conversational sentence translator of the present invention , each word in the source language in the word dictionary is
Typical scene information to be used is added,
The conversion means cancels the initial state or scene information
If no scene information has been entered,
The current conversation scene from words in the spoken language of the source language
To perform the conversion, so that each of the
It is possible to estimate the current conversation scene from words.
You. This allows the scene information to be
Scene information even if not set by the user
Is possible.

[Brief description of the drawings]

FIG. 1 is a schematic block diagram of a translation apparatus according to an embodiment of the present invention.

FIG. 2 is a flowchart for explaining the operation in the embodiment.

FIG. 3 is a screen image diagram of the conversational sentence translator in the embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Input part 2 Analysis part 3 Word dictionary 4 Scene information input part 5 Scene information management part 6 Conversion part 7 Bilingual database 8 Generating part 9 Output part 10 Scene information display part

Claims (4)

(57) [Claims] An input means for inputting a conversation sentence of a source language,
A word dictionary, analysis means for analyzing the input source language conversation sentence with reference to the word dictionary, and scene information input means for inputting scene information to set a conversation scene,
Management means for storing and managing the input scene information;
A display unit that displays the input scene information, a bilingual database of conversational sentences in the source language and the target language to which the scene information is added, and the input scene information and the bilingual database. A conversion unit that converts a conversation sentence of the source language into a conversation sentence of the target language; and an output unit that outputs the converted conversation sentence of the target language. The conversion unit is used for each word of the source language in the word dictionary. Be
Typical scene information is added, and the conversion means
Is canceled due to cancellation of the initial state or scene information.
If no pattern information has been entered, the
The current conversation scene is estimated from words in the language
An apparatus for translating a conversational sentence, comprising: 2. A method for inputting a word or selecting a translated word.
If no word information is added to the word or translation,
The current scene information is added, and the
Apparatus according to claim 1, wherein scene information is used . 3. Each word or single word in the word dictionary.
Multiple pieces of scene information are added to each correspondence between words and translated words.
The apparatus of claim 1 that. 4. A method for inputting a word or selecting a translated word.
Conversion frequency information is added, and the conversion means
The added scene frequency information is also used from the translation conversion.
The apparatus of claim 1 that.