JPS6383868A - Automatic translation system for natural language - Google Patents

Abstract

PURPOSE:To automatically select proper equivalent words by means of a connection value expression dictionary and a bilingual dictionary and in the simultaneous occurrence relation of words, by adding the meaning categories to both dictionaries. CONSTITUTION:An expression (connection value expression) dictionary 10 uses verbs, adjectives, etc., as keys to define the positions of other words, additional words, the meaning categories, etc. Then a bilingual dictionary 11 which contains the meaning categories in response to each equivalent word of an object language is used in addition to said dictionary 10. For instance, a keyword 'TABERU' (to eat) is inputted to a category detector 7 together with a subject word 'NIWATORI' (chicken). Thus a category 'SHOKURYO' (food) is detected out of the dictionary 10. The 'NIWATORI' is inputted to a category equivalent word detector 8, the pairs of categories (cook, bird) and (chicken, meat) are detected out of the dictionary 11. A category comparison/ equivalent word deciding device 9 searches an equivalent word having a category equal to or lower than 'SHOKURYO' in a meaning category system. Then an equivalent word of chicken is selected and outputted owing to a fact the meat is set at a rank lower than 'SHOKURYO'.

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical field to which the invention pertains) The present invention relates to an automatic natural language translation method that automatically translates a natural language read from an input device into another native language.

(Prior art) Conventionally, the selection of translation words for nouns, etc. has been carried out either by human intervention during the translation process and by selecting from candidate translations, or by looking at the translation results and making corrections. It was getting worse.

These methods have the drawbacks of requiring human intervention, which increases translation time and costs. To solve this problem, the objective is to provide a method for automatically selecting translated words.

(Structure of the Invention) The main feature of the present invention is that by assigning semantic categories to a bond expression dictionary and a bilingual dictionary, translation words are automatically selected based on the co-occurrence relationship of words using these dictionaries. shall be.

(Embodiment) FIG. 1 is a diagram showing an example of the overall basic configuration of the automatic translation system of the present invention, in which 1 is an input device for inputting the source language, and 2 is an analysis/conversion device for analyzing and converting the source language. 3 is a generation device that generates a target language based on the output of the analysis/conversion device 2;
4 is an output device that outputs the generated target language, and 5 is an analysis/
6 is a conversion dictionary and a generation dictionary.

Next, an example of the operation in which a translated word such as a noun is selected according to the present invention will be explained using FIGS. 2 to 5.

FIG. 2 shows an example of the configuration of a device that selects translated words such as nouns.

Reference numeral 7 indicates a category detection device from the combined value expression dictionary, 8 indicates a category/translation detection device from the bilingual dictionary, 9 indicates a category comparison/translation determining device, 10 indicates a combined value expression dictionary, and 11 indicates a bilingual dictionary.

Here, the category detection device 7 is positioned as the analysis/conversion device 2 in FIG. 1, and the category/translation detection device 8. The category comparison/translation determination device 9 is positioned in the generation device 3.

FIG. 3 shows an example of the structure of the combination value expression dictionary 10, and FIG. 4 shows an example of the structure of the bilingual dictionary 11.

Below, in the sentence "eat chicken", "chicken"
The operation will be explained using an example of selecting a translation word.

First, when a key word (eat) and a target word (chicken) are input into the category detection device 7, the combination value expression dictionary 10
Detects the category "food".

Next, when "chicken" is input into the category/translation detection device 8, the bilingual dictionary 11 selects two pairs of translations and categories.
Detect one.

That is, (cock, bird) and (chicken, meat).

Next, when the output of the category detection device 7 and the output of the category/translation detection device 8 are input to the category comparison/translation determination bag @9, the category comparison/translation determination device 9 selects the fifth
In the semantic category system as shown in the example in the figure,
Search for translated words that have the same or subordinate categories as "food".

First, we take ``bird'' and compare it with ``food,'' and reject the translation of cock because ``bird'' is not a subordinate of ``food.''

Next, since "meat" is listed under "food", a translated word for "chicken" is selected and output.

(Effects of the Invention) As described above, according to the present invention, an appropriate translation of a noun or the like can be automatically selected without human intervention.

[Brief explanation of the drawing]

Figure 1 is a diagram showing an example of the overall basic configuration of the automatic translation system of the present invention, Figure 2 is an example of the configuration of a translation word selection device, Figure 3 is an example of the configuration of a combination value expression dictionary, and Figure 4 is a diagram of a bilingual dictionary. Example of Structure: FIG. 5 is an example of the structure of a semantic category system. 1...Input device, 2...Analysis/conversion device, 3
...Generation device, 4...Output device, 5...
Analysis/conversion dictionary, 6... Generation dictionary, 7... Category detection device, 8... Category/translation detection device, 9... Category comparison/translation determination device, 10...
Combined value expression dictionary, 11... bilingual dictionary. Patent Applicant Nippon Telegraph and Telephone Corporation Figure 1 Figure 2 Figure 3 Figure 4 Funeral Figure 5 Abstract A Body # Department

Claims (1)

[Claims] From one natural language (source language) to another natural language (target language)
In a system that automatically performs translation, when there are multiple translations in the target language that correspond to nouns, etc. in the source language, it is possible to use verbs, adjectives, etc. as keys to determine the position of other words, attached words, and semantics. Expression determined by category, etc. (bond value expression)
Using a dictionary and a bilingual dictionary in which semantic categories are assigned to each translation word in the target language, the semantic categories of nouns, etc. in the source language are determined by the co-occurrence relationship of key words using a bond valence expression dictionary, Comparing Semantic Category 1 defined there with Semantic Category 2 assigned to the bilingual dictionary, it is determined whether Semantic Category 1 is subordinate to Semantic Category 2 or A natural language automatic translation method characterized in that if the words are the same, a translated word corresponding to the semantic category 2 is selected.