An Overview of Entity Alignment: Cross-lingual and Graph Approaches

Knowledge Graph is a technology for more effective knowledge sharing and makes the web smarter. By building Knowledge Graphs, information resources can be more easily calculated, understood and evaluated, effectively and accurately contributing to further development in various fields. Research on Knowledge Graph fusion aims to generate new knowledge by using models to fuse knowledge graphs from various domains, in which entity alignment is one of the fundamental researches in this field. For state-of-the-art studies, the hotspot of entity alignment primarily focus on cross-lingual and graph learning approaches. In this paper, we first introduce the main problems faced in the field of entity alignment research. Then the acquisition and usage of relevant models of entity alignment for cross-lingual and graph learning approaches are introduced. Next, the most frequently used evaluation indexes of entity alignment are presented. Finally, the challenges and future research directions of entity alignment in these two fields are further discussed.