Entity alignment method, device and equipment

Abstract

The invention discloses an entity alignment method, device and equipment. The method comprises steps of generating a heterogeneous sequence according to a sampling strategy based on degree perception;obtaining an entity alignment scheme represented by a target entity on the basis of correlation between nodes in the heterogeneous sequence, specifically, according to a preset sampling strategy based on degree perception, performing sequence sampling on each relation triple in the combined knowledge graph, and generating a heterogeneous sequence in which entity nodes and relation nodes appear interactively; and obtaining a target entity representation based on the positions of the entity nodes and the relationship nodes in the heterogeneous sequence and the correlation between the relationship nodes and the entity nodes. Because the target entity representation is obtained based on the positions of the entities and the relationships in the heterogeneous sequence and the correlation between the entities, the dependency between the entities can be enhanced, the long-distance interdependence characteristics between the entities can be captured conveniently, the representation capabilityfor the entities is improved, the alignment information is effectively propagated, and the entity alignment effect is improved.

Description

technical field [0001] The present invention relates to the technical field of knowledge graphs, in particular to an entity alignment method, device and equipment. Background technique [0002] In recent years, Knowledge Graph (KG) has played a pivotal role in many researches and applications of artificial intelligence. As the technical cornerstone of question answering and recommendation systems, it has received widespread attention. The knowledge map mainly describes the attributes of various entities (such as China, Jack Ma, etc.) and the relationship between them, usually expressed in the form of triples: (subject, predicate, object) can be abbreviated as SPO (subject, predicate, object) , the triplet describing the relationship between various entities is called a relational triplet: (entity, relation, entity). [0003] Knowledge graphs are constructed from unstructured information. The existing knowledge graphs are all extracted from major encyclopedia websites and un...

AI Technical Summary

Problems solved by technology

[0004] Due to the large differences in the expression of various aspects of entity information in different knowledge graphs, the existing methods for entity alignment mostly propagate alignment information through the learning of the unique triple structure of knowledge graphs, which have low expressiveness and poor information dissemination. Sufficiently efficient problem, it is difficult to efficiently obtain dependencies between two distant entities

Images