Text implication recognition method and device

Abstract

The invention belongs to the technical field of natural language processing, in particular to a text implication recognition method and a device. The present invention aims at solving the problems ofcomplex feature description and feature extraction of sentences in the prior art. The text implication identification method of the invention comprises the following steps: the source sentence and thetarget sentence of the text implication pair to be identified are subjected to dependency syntactic analysis; according to the analysis result, the source sentence dependency triple set and the target sentence dependency triple set are obtained; the semantic relationship between the source sentence dependent triple set and the target sentence dependent triple set is compared and the comparison result is obtained. The semantic implication relationship between source sentences and target sentences is predicted according to the comparison results. The method of the invention can improve the comparison precision between sentences from the whole angle and the alignment precision between the dependent triples of the sentences, thereby improving the accuracy of the prediction of the semantic implication relationship. The device of the invention also has the above-mentioned beneficial effects.

Description

technical field [0001] The invention belongs to the technical field of natural language processing, and in particular relates to a method and device for recognizing text implication. Background technique [0002] At present, the research on textual entailment recognition has received extensive attention in the field of natural language, and it has shifted from the initial pure theoretical discussion to larger-scale theoretical research and empirical platform construction. Currently, the methods of textual entailment recognition mainly include the following four types: [0003] First: the classification method. This method formalizes textual entailment as a classification problem. According to the labeled training examples, learn the features and build classifiers, most of which are based on SVM, and the features used for learning are mainly lexical-syntactic and semantic features based on lexical overlap or first-order logic. [0004] Second: deep semantic analysis and rea...

AI Technical Summary

Problems solved by technology

[0011] In order to solve the above-mentioned problems in the prior art, that is, in order to solve the problem that the prior art needs to perform complex feature description and feature extraction on sentences, the first aspect of the present invention provides a method for identifying textual implications, including:

Images