Cross-language entity alignment method based on knowledge graph multi-view information

Abstract

The invention discloses a cross-language entity alignment method based on knowledge graph multi-view information. The method comprises the following steps: firstly, according to triples and entity description texts of two language knowledge graphs, respectively extracting information to construct a structure graph and a text graph, and encoding vector representation on an entity structure and vector representation on a text by using a double-layer graph convolutional network; according to the entity description texts and a cross-language corpus, encoding vector representation on entity description by using a bidirectional long-short-term memory network; and calculating a final cross-language aligned entity pair by combining a weighting mode with vector distances of paired entities under three perspectives. According to the method, cross-language entity alignment of the knowledge graph is realized, entity vector representation is optimized based on multi-view information of the structure and the text, and the cross-language entity alignment accuracy is improved.

Description

technical field [0001] The present invention relates to a method for aligning cross-language entities based on knowledge map multi-view information, and in particular to a technology for realizing cross-language entity alignment using convolutional neural networks based on knowledge map structures and text information. Background technique [0002] Due to the rapid development of the Internet and the explosive growth of Internet information, people need to structure the information for further analysis and utilization, and serve various tasks and scenarios, thus knowledge graphs came into being. The knowledge graph is essentially a large-scale semantic network and a structured knowledge base, which formally describes the things in the objective world and the relationship between them. Entity alignment is to determine whether entities with different names or entities from different sources point to unique objects in the real world. In multilingual knowledge graphs, there are...

AI Technical Summary

Problems solved by technology

With the introduction of deep learning and the gradual development in the field of natural language processing, entity-based embedding representations and deep neural network entity alignment methods have become mainstream. Most methods are based on structured data of knowledge graphs, usually attribute triples and The comparison and calculation of relational triples fail to effectively utilize textual information to optimize entity alignment

Images