Computer Engineering and Applications ›› 2023, Vol. 59 ›› Issue (24): 70-77.DOI: 10.3778/j.issn.1002-8331.2302-0087

• Theory, Research and Development • Previous Articles     Next Articles

Joint Triple Embedding for Entity Alignment

LI Fengying, LI Jiapeng   

  1. Guangxi Key Laboratory of Trusted Software, Guilin University of Electronic Technology, Guilin, Guangxi 541004, China
  • Online:2023-12-15 Published:2023-12-15

联合三元组嵌入的实体对齐

李凤英,黎家鹏   

  1. 桂林电子科技大学 广西可信软件重点实验室,广西 桂林 541004

Abstract: Entity alignment is the task of identifying entities from different knowledge graphs that point to the same item and is important for KG fusion. Most of the existing approaches are based on graph neural networks that learn the entities embedding by modeling the neighborhood information of the entities. However, graph neural networks-based approaches have difficulty learning triples information in knowledge graphs, the triples information is not sufficiently utilized. In order to solve this problem, an entity alignment model with joint triple embedding is proposed in this paper. The proposed model computes a triple embedding for each entity and then uses this triple embedding for entity alignment. In addition, considering that the relations in the knowledge graphs have different types, in order to exploit these relation types, a relation type-aware calculation method of triple embedding is proposed. Meanwhile, constraints based on relation types are added to this model, to learn the mapping properties of relations. Experiments conducted on three real-world datasets show that this approach outperforms state-of-the-art methods, the effectiveness of the proposed method is verified.

Key words: entity alignment, graph convolutional networks, triple, knowledge graph

摘要: 实体对齐是从不同知识图谱中识别出指向相同实体的任务,对于知识图谱融合十分重要。现有的实体对齐方法大多基于图神经网络,通过建模实体的邻域信息,学习实体的嵌入表示。该方法难以学习知识图谱中的三元组信息,三元组信息利用不充分。为了解决该问题,提出了联合三元组嵌入的实体对齐模型。该模型先计算出每个实体的三元组嵌入,之后使用此三元组嵌入进行实体对齐。考虑到知识图谱中的关系具有不同的类型,为利用这些关系类型,提出了一种关系类型感知的三元组嵌入计算方法;同时在模型中加入了基于关系类型的约束,以学习关系映射属性。通过在三个公开的数据集上进行实验,实验数据优于同类实体对齐方法,验证了提出的方法的有效性。

关键词: 实体对齐, 图卷积网络, 三元组, 知识图谱