Multi-heterogeneous neighborhood-aware for Knowledge Graphs alignment

作者：

Highlights：

• Multi-source information of aligned entities and their heterogeneous neighbors is used to solve the entity alignment problem.

• A variable attention mechanism based on heterogeneous graphs is designed.

• Three well-known benchmark datasets that are rarely verified simultaneously in previous methods are used to evaluate this method.

摘要

•Multi-source information of aligned entities and their heterogeneous neighbors is used to solve the entity alignment problem.•A variable attention mechanism based on heterogeneous graphs is designed.•Three well-known benchmark datasets that are rarely verified simultaneously in previous methods are used to evaluate this method.

论文关键词：Entity alignment,Knowledge Graphs,Attribute structure,Attention mechanism,Heterogeneous graph attention

论文评审过程：Received 8 April 2021, Revised 14 August 2021, Accepted 4 October 2021, Available online 29 October 2021, Version of Record 29 October 2021.

论文官网地址：https://doi.org/10.1016/j.ipm.2021.102790