东北大学学报:自然科学版 ›› 2018, Vol. 39 ›› Issue (11): 1535-1539.DOI: 10.12068/j.issn.1005-3026.2018.11.004

• 信息与控制 • 上一篇    下一篇

表示学习知识图谱的实体对齐算法

朱继召, 乔建忠, 林树宽   

  1. (东北大学 计算机科学与工程学院, 辽宁 沈阳110169)
  • 收稿日期:2017-08-01 修回日期:2017-08-01 出版日期:2018-11-15 发布日期:2018-11-09
  • 通讯作者: 朱继召
  • 作者简介:朱继召(1986-),男,山东菏泽人,东北大学博士研究生; 乔建忠(1964-),男,辽宁兴城人,东北大学教授,博士生导师; 林树宽(1966-),女,吉林长春人,东北大学教授,博士生导师.冯明杰(1971-), 男, 河南禹州人, 东北大学副教授; 王恩刚(1962-), 男, 辽宁沈阳人, 东北大学教授,博士生导师.
  • 基金资助:
    国家自然科学基金资助项目(61272177).

Entity Alignment Algorithm for Knowledge Graph of Representation Learning

ZHU Ji-zhao, QIAO Jian-zhong, LIN Shu-kuan   

  1. School of Computer Science & Engineering, Northeastern University, Shenyang 110169, China.
  • Received:2017-08-01 Revised:2017-08-01 Online:2018-11-15 Published:2018-11-09
  • Contact: QIAO Jian-zhong
  • About author:-
  • Supported by:
    -

摘要: 与现有的根据知识图谱的结构信息或实体属性特征进行相似度匹配的实体对齐的方法不同,提出了一种基于表示学习的知识图谱实体对齐方法.首先,在低维向量空间下,通过机器学习方法学得实体和关系的语义表示,这种表示形式蕴含了知识图谱的内在结构信息及实体属性特征;其次,将人工标注的实体对作为先验知识,学习知识图谱间实体对的映射关系.经实验验证表明:与基于特征匹配的方法SiGMa相比,本文方法能够有效提高知识图谱实体对齐的精确率,同时保持较高的F1值.

关键词: 机器学习, 表示学习, 知识图谱, 知识融合, 实体对齐

Abstract: A novel supervised method for knowledge graph entity alignment based on representation learning was proposed, which is different from the existing methods due to the similarity of structural information or attributive characters. First, the method automatically learns the semantic representations for the entities and relations of a knowledge graph in a low-dimensional vector space was proposed, and these embeddings contain the intrinsically structural information of a knowledge graph and the attributive features of entities. Afterwards, taking the manually aligned entity pairs as prior knowledge, the cross-KG mapping relationship between entities could be learned, which will be used for predicting entity alignment. Experiments conducted on real datasets demonstrated that our method can effectively improve the precision of knowledge graph entity alignment while keeping a high F1 score, when compared with the feature matching based method SiGMa.

Key words: machine learning, representation learning, knowledge graph, knowledge fusion, entity alignment

中图分类号: