Entity Linking Method Based on Integrating Knowledge Graph Subgraph Information and Entity Information

Abstract

The invention relates to an entity linking method based on sub-graph information and entity information of a knowledge graph. The method includes the following steps: acquiring question text, feeding a trained mention recognition model, and identifying predicted mention words; Mentions are recalled to obtain target mentions with higher accuracy. Obtain all candidate entities corresponding to the mentioned words from the knowledge base according to the target mentioned words; for each candidate entity, retrieve its knowledge graph subgraph information; integrate the knowledge graph subgraph information into the candidate entity to obtain the candidate entity text, The entity boundary label is introduced into the question text and candidate entity text to construct an entity disambiguation dataset; the trained entity disambiguation model is fed, and the knowledge base entities corresponding to the mentioned words are obtained by pooling, and the entity linking is completed. The invention can effectively improve the recall rate of mentioned word recognition, improve the sorting accuracy rate of candidate entities, reduce error transmission, and effectively solve the reasoning and unseen entities in the training set.

Description

technical field [0001] The invention relates to the field of natural language processing of artificial intelligence, in particular to an entity linking method based on sub-graph information and entity information of a knowledge map. Background technique [0002] Knowledge base question answering (KBQA) refers to given a natural language question, through semantic understanding and analysis of the question, and then use the knowledge base to query and reason to get the answer. There are two types of methods: methods based on semantic analysis (SP) and methods based on information extraction (IE). Among them, the information extraction method has a relatively fixed model framework and good generalization performance. This type of method generally adopts a pipelined architecture, first performing (1) entity linking, that is, linking the mention in the question to the entity of the knowledge base. Then carry out (2) relation identification, that is, map the intent of the quest...

AI Technical Summary

Problems solved by technology

[0004] Based on this, in order to solve the problem of small amount of information in the upper text and blurred boundaries of Chinese named entities, the present invention provides an entity link method based on the integration of knowledge map subgraph information and entity information, by improving the performance of KBQA entity link tasks, and then Further improve the performance of the whole knowledge base question answering task to address the problems mentioned in the background above

Images