JPH05204965A - Word meaning selecting device - Google Patents

Abstract

PURPOSE:To secure a fact that one meaning of a word is selected surely, even in the case the meaning of a word cannot be selected by only a constraint condition described in a dictionary. CONSTITUTION:In a dictionary 3, an example is added in addition to a constraint condition at every meaning of a word at the time of preparing the dictionary. In the case the meaning of a word cannot be selected by only a constraint condition, a word meaning selecting means 2 retrieves an example being the most similar to an input expression by using a similar expression retrieving means 4, and selects the meaning of a word to which its example is added, as the most appropriate meaning of a word.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a word sense selecting device in a natural language analysis system, such as a machine translation system, a natural language interface, and a revision support system, which analyzes an inputted natural language sentence and outputs a corresponding semantic expression. Regarding

[0002]

2. Description of the Related Art Generally, a natural language word expression has a plurality of word meanings, and at the time of analysis, one of the word meanings b1 to bn corresponding to the word expression a is selected according to the appearance environment of the word expression a. You need to choose the most appropriate one.

In the current natural language analysis system, it is basically said that "a restriction on a language structure for a certain expression is described in a dictionary / grammar in advance, and these are applied to an input expression to select a word meaning". The word sense selection method is adopted. FIG. 4 shows the structure of a dictionary for a certain headword a used in this word sense selection method. When the partial expression A of natural language is input and the word meaning a for the word expression a in the partial expression A is determined, this dictionary is selected with the entry word a as a key, and from the corresponding meaning groups b1 to bn, meaning bi the attendant constraints Ci1～cim _i is subexpression a satisfies is selected as appropriate meaning.

For example, in selecting the meaning of a word,
This constraint condition is mainly described as a semantic constraint for the essential case element. As an example, FIG. 5 shows a part of the dictionary for the Japanese verb "take". When the sentence "He took glasses" is input to the Japanese analysis system that has this dictionary, it matches the conditions of "the subject is a person" and "what the case wears". The meaning of "" is to select "to remove / remove what you are wearing".

On the other hand, recently, for example, in machine translation, parallel translation cases are accumulated in advance, a case similar to the input is searched using a thesaurus, and the case is imitated and translated, so that some translation result is always output. A method of case-based translation, which guarantees that it is done, has also been proposed (Sumita / Iida / example-driven machine translation, IPSJ Natural Language Processing Research Group, 82-5 (1991)).

[0006]

However, according to the word sense selection method of the above-mentioned form, "in the word sense groups b1 to bn corresponding to the word representation a in the input partial representation A, the accompanying constraint conditions are partially represented. If there are a plurality of word meanings that A satisfies, it cannot be determined which of them is most appropriate. "" Word meaning groups b1 to b corresponding to the word expression a in the input partial expression A
If none of the subexpressions A satisfies the constraint conditions associated with n, the analysis will fail at that point. " Ideally, if all the constraint conditions associated with the word sense groups b1 to bn corresponding to the word expression a are exclusive and the union set covers all the appearance environments of the word expression a, the word expression a The only meaning of any input is always corresponding to any input including, but it is actually impossible to describe the condition in this way.

Further, in the above-described case-based translation method, since the translated word for the input expression is directly determined in correspondence with each surface vocabulary, the meaning of the input expression cannot be determined. In addition, a large-scale parallel translation example database is required to realize this method, but it is difficult to collect a sufficient amount of parallel translation text for this purpose, and even if it can be collected, it will be automatically analyzed. Therefore, there is no means for structurally associating with each other, and if it is done manually, it has a drawback that it takes a lot of time and cost.

As described above, the word meaning selection method based on the constraint description is "a word meaning in which the partial expression A satisfies the accompanying constraint condition in the word meaning groups b1 to bn corresponding to the word expression a in the input partial expression A. , It is not possible to determine which one of them is the most appropriate. "" The subexpression A has all the constraint conditions associated with the meaning groups b1 to bn corresponding to the word expression a in the input subexpression A. If it is not satisfied, the analysis will fail at that point. " On the other hand, the case-based translation method has a drawback that it is difficult to create a bilingual case database and, in its original form, cannot be applied to word sense selection.

The object of the present invention is to add a constraint condition to each word sense in the dictionary and add an example so that the same meaning as the case-based translation method can be used as a compensation function when the word sense cannot be selected only by the constraint. An object of the present invention is to provide a word sense selection device that makes use of the advantages of both the word sense selection method by constraint description and the case-based analysis of selecting the most appropriate sense by using the similar expression search function.

[0010]

A word sense selecting device according to a first aspect of the present invention comprises a dictionary describing constraint conditions and examples for selecting the word sense for each word sense, a specific expression from a set of expressions, and a specific expression. A similar expression search means for searching a similar expression,
When determining the meaning of a word expression in the input expression, the constraint condition described in the meaning of the entry word that matches the word expression in the dictionary is applied to the input expression, and the accompanying constraint condition is applied to the input expression. An example in which a satisfying word sense is selected as the most appropriate word sense, and when there are multiple word senses that satisfy the accompanying constraints, the accompanying constraint conditions are respectively described in the meanings that satisfy the input phrase. And a word sense selecting unit that searches for an example most similar to an appropriate partial expression including the word expression in the input expression by the similar expression searching unit, and selects the word meaning in which the example is described as the most appropriate word meaning. , Input language expression is linguistically analyzed and decomposed into partial expressions to generate a set of partial expressions, which are sent to the word meaning selecting means and received from the word meaning selecting means. To produce a semantic representations from the set of semantic, characterized in that it comprises a natural language analysis unit for outputting.

The word sense selecting device according to the second aspect of the present invention searches a dictionary describing a constraint condition and an example for selecting the word sense for each word sense and an expression most similar to a specific expression from a set of expressions. When the meaning of a word expression in the input expression is determined by the similar expression search means, the constraint condition described in the word meaning of the entry word that matches the word expression in the dictionary is applied to the input expression, and the associated constraint is applied. When a condition that the input expression satisfies the condition is selected as the most appropriate word meaning, and if the input expression does not satisfy any constraint condition, the input expression is Is searched for by the similar expression search means, and the word meaning in which the example is described is selected as the most appropriate word meaning. And a meaning definition selecting means, which performs linguistic analysis of the input language expression and decomposes it into partial expressions to generate a set of partial expressions, which is sent to the word meaning selecting means, and a meaning expression from the meaning meaning set received from the word meaning selecting means. Generate,
It is characterized by comprising a natural language analysis means for outputting.

[0012]

In the word sense selecting device of the present invention, the natural language analyzing means receives the input expression, linguistically analyzes the input expression, decomposes it into partial expressions, generates a set of partial expressions, and sends it to the word meaning selecting means. The word meaning selecting means refers to the dictionary, determines the meaning of the word expression in the received partial expressions using the similar expression searching means if necessary, and sends it to the natural language analyzing means. The natural language analysis means generates and outputs a semantic expression from the set of word senses received.

In the dictionary, an example is added in addition to the constraint condition for each word sense when the dictionary is created.

The similar expression search means searches the set of examples for an example that is most similar to a specific subexpression. The input expression used to determine the most similar example-The method of calculating the distance between examples includes a method that uses a thesaurus similar to the translation method based on conventional cases, as well as the meaning of each expression as a pair of feature and feature value. There are various methods, such as a method of counting the number of features whose values match each other.

[0015]

EXAMPLES Next, examples of the present invention will be described with reference to FIGS.

FIG. 1 is a block diagram showing an embodiment of the first and second inventions, and FIG. 2 is a diagram showing the structure of a dictionary for a certain headword a used in the word sense selecting device of the invention.
FIG. 8 is a diagram showing an example of a dictionary for the verb “take”.

In FIG. 1, the natural language analysis means 1 receives a linguistic expression as an input 11, linguistically analyzes the input expression, decomposes it into partial expressions, generates a set of partial expressions, and sends it to the word sense selecting means 2.

The word meaning selecting means 2 refers to the dictionary 3 to determine the word meaning for the word expression in the received partial expressions.

In the dictionary 3, as shown in FIG. 2, the constraint conditions ci1 to cim are applied to the word meaning bi for the headword a.
_i and examples ei1 to eil _i are described. The number of examples may be at least one. An example of the described example is shown in FIG.

When the partial expression A including the word expression a is input, the word sense selecting means 2 refers to the dictionary shown in FIG. 2 using the entry word a as a key, and uses the constraint description described in the prior art. Similar to the word sense selection method, the constraint conditions associated with the word sense groups b1 to bn corresponding to the word representation a are applied to the partial representation A. The constraint conditions that the subexpression A satisfies are ci1 to ci.
In the case of only m _i , the word meaning bi is selected as an appropriate word meaning.

In the first invention, the word expression a is included in the constraint conditions associated with the word sense groups b1 to bn corresponding to the word expression a.
When there are a plurality of constraints satisfying the partial expression A including "," the word sense selecting means 2 uses the similar expression searching means 4 to select a part from the examples added to the word sense group with the constraint condition. The example most similar to the expression A is searched, and the word meaning bj to which the example ejk is added is selected as the most appropriate word meaning.

In the second aspect of the invention, when the partial expressions A including the word expression a do not satisfy the constraint conditions associated with the word meaning groups b1 to bn corresponding to the word expression a, the word meaning selecting means 2 is similar. Using the expression search means 4, from the examples added to the word sense groups b1 to bn corresponding to the word expression a,
Find an example that is most similar to the subexpression A, and use that example ej
The word meaning bj with k added is selected as the most appropriate word meaning.

The determined word meaning is sent from the word meaning selecting means 2 to the natural language analyzing means 1, and the natural language analyzing means 1 generates a meaning expression from the received set of word meanings and outputs it.

The first invention and the second invention are applied to different cases, and both can be used in combination freely. The examples attached to the dictionary can be commonly used in both inventions.

[0025]

As described above, according to the present invention, it is possible to realize a word sense selecting device that makes use of the advantages of both the conventional word sense selecting method using constraint description and the case-based analysis.
That is, by adding an example to the dictionary, it is possible to guarantee that any word sense is selected, and it is possible to cover the drawbacks of the word sense selection method by the constraint description. In addition, in order to utilize the idea of the case-based translation method as it is in word sense selection, incorporate the advantage of the case-based translation method that "it is possible to give reliability to the translation result in the form of distance from the used case". Is also possible.

On the other hand, the case-based translation has a drawback that it is "difficult to create a parallel translation case database". However, in the word sense selection device of the present invention, at least one example need only be added when creating a dictionary. Creating a dictionary in this form is not difficult. In particular, when creating a dictionary by referring to an existing dictionary such as a Japanese dictionary or an English-English dictionary, the examples in the dictionary may be used as they are. Further, even if the user is required to add at least one example when registering a new word or a new meaning for an existing word in the dictionary, the user is likely to register based on the specific example, so the burden on the user The increase is not large. In addition to this, since this example has an example described in the dictionary,
It also has the effect that the dictionary is easy for humans to see.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is a diagram showing a configuration of a dictionary for a certain headword a used in the present invention.

FIG. 3 is a diagram showing an example of the configuration of a dictionary used in the present invention in the form of a dictionary for the verb “take”.

FIG. 4 is a diagram showing a configuration of a dictionary for a certain headword a, which is used in the word sense selection method based on the constraint description of the related art.

FIG. 5 is a diagram showing an example of a configuration of a dictionary used in a word sense selection method based on a constraint description of the related art, in the form of a dictionary for a verb “take”.

[Explanation of symbols]

 1 natural language analysis means 2 word sense selection means 3 dictionary 4 similar expression search means 11 input 12 output

Claims (2)

[Claims] 1. A dictionary describing constraint conditions and examples for selecting the word meaning for each word meaning, a similar expression search means for searching an expression most similar to a specific expression, and an input expression. When determining the meaning of a certain word expression in the dictionary, the constraint condition described in the meaning of the entry word that matches the word expression in the dictionary is applied to the input expression, and the accompanying constraint condition is satisfied by the input expression. While selecting the word meaning as the most appropriate word meaning, when there are a plurality of word meanings that the input expression satisfies the accompanying constraint conditions, the accompanying constraint conditions are respectively described from the example described in the word meaning that the input expression satisfies, An example that is most similar to an appropriate partial expression including the word expression in the input expression is searched by the similar expression search means, and the word meaning in which the example is described is the most appropriate word. And meaning selection means for selecting as to generate a set of partial representation by decomposing the input language expression to language analysis part representation, and sends to the meaning selection means,
A word sense selecting device comprising natural language analyzing means for generating and outputting a semantic expression from a set of word senses received from the word sense selecting means. 2. A dictionary describing constraint conditions and examples for selecting the word sense for each word sense, a similar expression search means for searching an expression most similar to a certain expression from a set of expressions, and an input expression. When determining the meaning of a certain word expression in the dictionary, the constraint condition described in the meaning of the entry word that matches the word expression in the dictionary is applied to the input expression, and the accompanying constraint condition is satisfied by the input expression. When the word meaning is selected as the most appropriate word meaning, and when the input expression does not satisfy any constraint condition, an appropriate word expression including the word expression in the input expression is selected from the examples described in the word meaning for the headword. A word sense selecting unit that searches an example most similar to the partial expression by the similar expression searching unit and selects the word meaning in which the example is described as the most appropriate meaning. A linguistic expression is linguistically analyzed and decomposed into partial expressions to generate a set of partial expressions, which are sent to the word sense selecting means,
A word sense selecting device comprising natural language analyzing means for generating and outputting a semantic expression from a set of word senses received from the word sense selecting means.