计算机工程与应用2025，Vol.61Issue(14)：37-53,17.DOI:10.3778/j.issn.1002-8331.2409-0253
时态知识图谱表示学习研究综述
Survey on Temporal Knowledge Graph Representation Learning
摘要
Abstract
Knowledge graph(KG)representation learning aims to convert the original symbolic knowledge representa-tion into numerical knowledge representation for better knowledge-driven applications.Temporal knowledge graph(TKG)representation learning technology makes full use of time information in KG and achieves performance improvement.This paper systematically reviews TKG representation learning methods from four aspects:(1)It briefly introduces the related concepts,typical tasks of TKG representation learning and traditional static methods.(2)It summarizes two kinds of methods of TKG representation learning,namely,interpolation task oriented method and extrapolation task oriented method,and introduces the typical models of the two kinds of methods respectively.(3)Eight benchmark datasets for TKG representation learning and evaluation results of several typical models on the benchmark datasets are sorted out.(4)The current technical challenges and opportunities are analyzed.关键词
知识图谱(KG)/时态知识图谱(TKG)/表示学习/知识表示Key words
knowledge graph(KG)/temporal knowledge graph(TKG)/representation learning/knowledge representation分类
信息技术与安全科学引用本文复制引用
何鹏,姚瑶,刘秋菊..时态知识图谱表示学习研究综述[J].计算机工程与应用,2025,61(14):37-53,17.基金项目
河南省高等教育教学改革研究与实践重点项目(2024SJGLX0206) （2024SJGLX0206）
河南省科技攻关项目(222102210014). （222102210014）
