Entity Alignment for Heterogeneous Knowledge Graphs using Summary and Attribute Embeddings

Abstract

Knowledge Graph (KG) is a well known way of representing facts about the real world in the form of entities, where nodes and edges represent the entities and their respective relations. However, many KGs have been constructed inde- pendently for different purpose. Therefore, very limited number of the entities stored in different KGs are aligned. This paper presents an embedding-based en- tity alignment method. We propose a joint method of summary and attribute embeddings for entity alignment task. Our model learns the representations of entities by using relational triples, attribute triples and description as well. When entities have less number of attributes or when the relational structure couldn’t capture the meaningful representation of the entities, entity summary embedding can be useful. We perform experiments on real-world datasets and the results indicate that the proposed approach significantly outperformed the state-of-the- art models for entity alignment.