知识图谱构建技术综述

知识图谱在医疗、金融、农业等领域得到快速发展与广泛应用，其可以高效整合海量数据的有效信息，为实现语义智能化搜索以及知识互联打下基础。随着深度学习的发展，传统基于规则和模板的知识图谱构建技术已经逐渐被深度学习所替代。梳理知识抽取、知识融合、知识推理3类知识图谱构建技术的发展历程，重点分析基于卷积神经网络、循环神经网络等深度学习的知识图谱构建方法，并归纳现有方法的优劣性与发展思路。此外，深度学习虽然在自然语言处理、计算机视觉等领域取得了较大成果，但自身存在依赖大规模样本、缺乏推理性与可解释性等缺陷，限制了其进一步发展。为此，对知识图谱应用于深度学习以改善深度学习自身缺陷的相关方法进行整理，分析深度学习的可解释性、指导性以及因果推理性，归纳知识图谱的优势以及发展的必要性。在此基础上，对知识图谱构建技术以及知识图谱应用于深度学习所面临的困难和挑战进行梳理和分析，并对该领域的发展前景加以展望。

Survey of Knowledge Graph Construction Techniques

Knowledge graph has been rapidly developed and widely used in the medical,financial,agricultural,and other fields. It can efficiently integrate the effective information of massive data and lay the foundation for semantic intelligent search and knowledge interconnection. With the development of deep learning,the traditional knowledge graph construction technology based on rules and templates has been gradually replaced by deep learning. This paper studies the development process of three types of knowledge graph construction technologies:knowledge extraction,knowledge fusion,and knowledge reasoning;focuses on knowledge graph construction methods based on deep learning such as Convolutional Neural Network(CNN)and Recurrent Neural Network(RNN);and summarizes the advantages and disadvantages of existing methods and development ideas. In addition,although deep learning has made great achievements in Natural Language Processing(NLP),computer vision,and other fields,its own defects such as reliance on large-scale samples,lack of reasoning,and interpretability limit its further development.Therefore,this paper sorts out the relevant methods for applying knowledge graph to deep learning to address the defects of the latter;analyzes the interpretability,guidance,and causal reasoning of deep learning;and summarizes the advantages of knowledge graph and the necessity of development. On this basis,this paper studies and analyzes the construction technology of knowledge graph and the difficulties and challenges faced by the application of knowledge graph in deep learning and looks forward to the development prospect of this field.