《计算机应用》唯一官方网站 ›› 2026, Vol. 46 ›› Issue (3): 683-695.DOI: 10.11772/j.issn.1001-9081.2025030294

• 人工智能 •    下一篇

面向知识图谱补全的大模型方法综述

张昊洋, 张丽萍(), 闫盛, 李娜, 张学飞   

  1. 内蒙古师范大学 计算机科学技术学院,呼和浩特 010022
  • 收稿日期:2025-03-24 修回日期:2025-06-27 接受日期:2025-06-30 发布日期:2025-07-18 出版日期:2026-03-10
  • 通讯作者: 张丽萍
  • 作者简介:张昊洋(2000—),男,内蒙古包头人,硕士研究生,主要研究方向:教育数据挖掘、智慧教育
    闫盛(1984—),男,内蒙古包头人,讲师,硕士,CCF会员,主要研究方向:计算机教育
    李娜(1980—),女,湖北黄陂人,副教授,硕士,主要研究方向:教育信息化、信息技术教育、数字化学习资源设计与开发
    张学飞(1997—),男,内蒙古乌兰察布人,硕士研究生,CCF会员,主要研究方向:计算机教育。
  • 基金资助:
    国家自然科学基金资助项目(61462071);内蒙古自然科学基金资助项目(2023LHMS06009);内蒙古自然科学基金资助项目(2024MS06020);内蒙古自治区教育科学研究“十四五”规划2023年度课题(2023NGHZXZH119);内蒙古自治区教育科学研究“十四五”规划2023年度课题(NGJGH2023234);内蒙古师范大学基本科研业务费专项资金资助项目(2022JBQN108);内蒙古师范大学基本科研业务费专项资金资助项目(2022JBQN008)

Review of large language model methods for knowledge graph completion

Haoyang ZHANG, Liping ZHANG(), Sheng YAN, Na LI, Xuefei ZHANG   

  1. College of Computer Science and Technology,Inner Mongolia Normal University,Hohhot Inner Mongolia 010022,China
  • Received:2025-03-24 Revised:2025-06-27 Accepted:2025-06-30 Online:2025-07-18 Published:2026-03-10
  • Contact: Liping ZHANG
  • About author:ZHANG Haoyang, born in 2000, M. S. candidate. His research interests include educational data mining, smart education.
    YAN Sheng, born in 1984, M. S., lecturer. His research interests include computer education.
    LI Na, born in 1980, M. S., associate professor. Her research interests include education informatization, information technology education, digital learning resources design and development.
    ZHANG Xuefei, born in 1997, M. S. candidate. His research interests include computer education.
  • Supported by:
    National Natural Science Foundation of China(61462071);Natural Science Foundation of Inner Mongolia(2023LHMS06009);2023 Project of “the 14th Five-Year Plan” for Educational Science Research of Inner Mongolia Autonomous Region(2023NGHZXZH119);Fundamental Research Funds of Inner Mongolia Normal University(2022JBQN108)

摘要:

知识图谱(KG)可从海量数据中提取并结构化表示先验知识,在智能系统的构建与应用中发挥着关键作用。知识图谱补全(KGC)旨在预测KG中缺失的三元组以提升完整性和可用性,通常涵盖编码环节与预测环节。然而,传统的KGC方法在编码环节存在难以有效利用额外信息与语义信息的问题,而在预测环节存在知识覆盖不完全及封闭世界问题,且先编码后预测的框架会受到嵌入表示形式和计算效率的限制。大语言模型(LLM)凭借丰富的知识和强大的理解力能够解决这些问题。因此,对面向知识图谱补全的大模型方法进行综述。首先,概述KG与LLM的基本概念及研究现状,并阐述KGC的流程;其次,将现有基于LLM的KGC方法从将LLM作为编码器、将LLM作为生成器以及基于提示引导三方面进行总结和梳理;最后,总结模型在不同数据集上的性能表现并探讨基于LLM的KGC研究面临的问题与挑战。

关键词: 知识图谱, 知识图谱补全, 大语言模型, 链接预测, 三元组分类

Abstract:

Knowledge Graph (KG) can extract and structurally represent the prior knowledge from massive data, and plays a key role in the construction and application of intelligent systems. Knowledge Graph Completion (KGC) aims to predict missing triples in the KGs to improve integrity and usability, and usually covers encoding and prediction links. However, the traditional KGC methods have difficulties in utilizing additional information and semantic information effectively in the encoding process, the problems of incomplete knowledge coverage and closed world in the prediction process, and the framework of first encoding and then prediction will be limited by embedded representation forms and computing efficiency. Large Language Models (LLMs) can solve these problems with rich knowledge and strong understanding abilities. Therefore, LLM methods for KGC were reviewed. Firstly, the basic concepts and research status of KGs and LLMs were outlined, and the KGC process was explained. Secondly, the existing KGC methods based on LLMs were summarized and sorted out from three aspects: using LLM as an encoder, using LLM as an generator, and basing on prompt guidance. Finally, the performance of the models on different datasets was summed up and the problems and challenges faced by KGC research based on LLMs were discussed.

Key words: knowledge graph, Knowledge Graph Completion (KGC), Large Language Model (LLM), link prediction, triple classification

中图分类号: