摘要
随着知识图谱(KG)的广泛应用,其构建过程中出现的知识缺失问题日益引起关注,因此知识图谱补全(KGC)成为当下的关键研究方向之一。知识图谱补全旨在自动识别并填补图谱中缺失的实体和关系,提升知识图谱的完整性和精准性。本文系统梳理了知识图谱补全的主要方法,包括传统方法、基于深度学习、图神经网络、强化学习及元学习的补全策略,重点分析了各自的特点与适用场景。同时,探讨了知识图谱补全技术在中医药典籍挖掘中的创新应用,通过补全技术整合分散的典籍资源,推进中医药学理论的现代化解读,融合多模态知识实现中医古籍的智能化发展与传承。该研究为中医药典籍知识的整理与现代化应用提供了理论支持与技术保障。
With the widespread application of knowledge graphs(KGs),the challenge of knowledge incompleteness during their construction has garnered significant research attention,positioning knowledge graph completion(KGC)as a pivotal frontier in artificial intelligence research.KGC methodologies automatically identify and supplement missing entities and relationships to enhance KG integrity and precision.This paper systematically categorizes mainstream KGC approaches,including traditional statistical models,deep learning architectures,graph neural networks,reinforcement learning frameworks,and meta-learning strategies,with critical analyses of their technical characteristics and scenario-specific applicability.Furthermore,we pioneer the exploration of KGC's transformative role in traditional Chinese medicine(TCM)Classics mining.The proposed intelligent completion mechanism synergistically integrates fragmented herbal knowledge from historical archives,facilitates modern reinterpretation of TCM theories through cognitive computing,and enables cross-modal knowledge fusion for intelligent inheritance of medical wisdom.Our research establishes a methodological foundation for systematizing TCM knowledge while providing an AI-driven paradigm for revitalizing ancient medical heritage.
作者
于佳田
佟旭
佟琳
唐竹慧
张宇鹏
张华敏
YU Jiatian;TONG Xu;TONG Lin;TANG Zhuhui;ZHANG Yupeng;ZHANG Huamin(Institute of Basic Theory for Traditional Chinese Medicine,China Academy of Chinese Medical Sciences,Beijing 100700,China;Institute of Traditional Chinese Medicine Information,China Academy of Chinese Medical Sciences,Beijing 100700,China)
出处
《中国中医基础医学杂志》
2025年第8期1339-1346,共8页
JOURNAL OF BASIC CHINESE MEDICINE
基金
国家重点研发计划项目(2023YFC3502900)
国家自然科学基金项目(82305439)
中国中医科学院基本科研业务费优秀青年科技人才培养专项(ZZ16-YQ-053)
中央级公益性科研院所基本科研业务费专项(YZX-202422)。
关键词
知识图谱
知识图谱补全
中医药
典籍挖掘
Knowledge graph
Knowledge graph completion
Traditional Chinese medicine
Classical text mining
作者简介
于佳田,在读硕士研究生,从事基于智能化技术的中医理论创新研究;通信作者:佟旭,副研究员,从事基于智能化技术的中医理论创新研究;张华敏,研究员,从事经典名方考证及其现代科学内涵解析研究。
