基于多信息图注意力网络的双向迭代实体对齐  

Bi-directional iterative entity alignment based on multi-information graph attention network

作　　者：许智宏 刘梓艺[1] 王利琴 董永峰 XU Zhi-hong;LIU Zi-yi;WANG Li-qin;DONG Yong-feng(School of Artificial Intelligence,Hebei University of Technology,Tianjin 300401,China;Hebei Key Laboratory of Big Data Computing,Hebei University of Technology,Tianjin 300401,China;Hebei Data Driven Industrial Intelligent Engineering Research Center,Hebei University of Technology,Tianjin 300401,China)

机构地区：[1]河北工业大学人工智能与数据科学学院,天津300401 [2]河北工业大学河北省大数据计算重点实验室,天津300401 [3]河北工业大学河北省数据驱动工业智能工程研究中心,天津300401

出　　处：《计算机工程与设计》2023年第6期1836-1843,共8页Computer Engineering and Design

基　　金：国家重点研发计划基金项目(2019YFC1904601);国家青年科学基金项目(61902106);天津市自然科学基金项目(19JCZDJC40000)。

摘　　要：针对当前实体对齐方法无法充分利用知识图谱实体的多种信息和缺乏先验对齐种子对限制实体对齐效果的问题,提出一种基于多信息图注意力网络的双向迭代实体对齐模型。使用具有高速公路网络机制的图注意力网络,充分利用实体的多种信息获得实体的嵌入表示,提出一种双向迭代策略,设置阈值筛选高置信度实体对加入训练集中,达到扩展先验对齐种子对的目的。在3个跨语言实体对齐数据集上进行的实验结果表明,该模型有效提高了评估指标Hits@1,Hits@10的性能。Aiming at the problems that current entity alignment methods are unable to fully utilize multiple information of know-ledge graph entities and lack of priori alignment seeds to limit the effect of entity alignment,a bi-directional iterative entity alignment model based on multi-information graph attention network was proposed.A graph attention network with a highway network mechanism was used,and multiple information of entities was used to obtain the embedded representation of the entities.A bi-directional iterative strategy was proposed,in which a threshold was set to screen high confidence entity pairs into the training set to achieve the purpose of expanding prior alignment seed pairs.Result of experiments on three cross-lingual entity alignment datasets show that the model effectively improves the performance of the evaluation indicators Hits@1,Hits@10.

关 键 词：知识图谱 实体对齐 多信息 图注意力网络 高速公路网络 表示学习 双向迭代 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]