Entity linking method based on context semantic relation and document consistency constraint

Abstract

The invention provides an entity linking method based on context semantic relation and document consistency constraint in order to solve the problems of cost and the like caused by the fact that a large amount of manual annotation is needed in a traditional entity linking method. The method comprises three steps of data preprocessing, candidate entity generation and candidate entity disambiguation. In the data preprocessing stage, the noise problem existing in the data is solved; in the candidate entity generation stage, a candidate entity set with a high recall rate is obtained by using a filtering technology of a Wikipedia connection graph; in the candidate entity disambiguation stage, a candidate entity set is used as a weak supervision constraint, the relation between entities and local contexts and coherence information between entities in a document are considered, candidate entity disambiguation is carried out through a neural network, a final entity link result is obtained, and candidate entities correspond to a knowledge graph.

Description

technical field [0001] The present invention relates to the field of knowledge map entity linking, in particular, to an entity linking method based on contextual semantic relations and document consistency constraints. Background technique [0002] In recent years, with the rapid development of artificial intelligence technology, how to enable machines to understand human language through natural language processing technology (NLP) has attracted the attention of many scholars. How to train the computer to recognize the entities in the text and link the entities to the database correctly is a key step for the computer to understand human language. Entity Linking (EntityLinking, EL) refers to mining potential entity relationships in human language texts, and linking them to the knowledge map entities where they are located, to solve the ambiguity (that is, polysemy of one word) and diversity (that is, multiple words and one word) between entities. righteousness) task. It is...

AI Technical Summary

Problems solved by technology

[0005] A. Data preprocessing: Select unlabeled documents and Wikipedia data in the experimental data set to preprocess, because when using multiple data sets, there will be the following two problems: the data organization format of multiple data sets is not uniform; The entity annotations of some datasets are incomplete;

Images