面向知识图谱的知识推理研究进展

近年来，随着互联网技术和应用模式的迅猛发展，引发了互联网数据规模的爆炸式增长，其中包含大量有价值的知识.如何组织和表达这些知识，并对其进行深入计算和分析，备受关注.知识图谱作为丰富直观的知识表达方式应运而生.面向知识图谱的知识推理是知识图谱的研究热点之一，已在垂直搜索、智能问答等应用领域发挥了重要作用.面向知识图谱的知识推理旨在根据已有的知识推理出新的知识或识别错误的知识.不同于传统知识推理，由于知识图谱中知识表达形式的简洁直观、灵活丰富，面向知识图谱的知识推理方法也更加多样化.本文将从知识推理的基本概念出发，介绍近年来面向知识图谱知识推理方法的最新研究进展.具体地，本文根据推理类型划分，将面向知识图谱的知识推理分为单步推理和多步推理，根据方法的不同，每类又包括基于规则的推理、基于分布式表示的推理、基于神经网络的推理以及混合推理.本文详细总结这些方法，并探讨和展望面向知识图谱知识推理的未来研究方向和前景.

Knowledge Reasoning Over Knowledge Graph: A Survey

In recent years, the rapid development of Internet technology and Web applications has triggered the explosion of various dataon the Internet, which contains a large amount of valuable knowledge. How to organize, represent and analyze these knowledge has attracted much attention. Knowledge graph was thus developed to organize these knowledge in a semantical and visualized manner. Knowledge graph oriented knowledge inference then comes into being one of the hot research topics, which has played an important role in vertical search, intelligent question-answer, etc. The goal of knowledge graph oriented knowledge inferenceis to infer new facts or identify erroneous factsaccording to existing ones. Unlike traditional knowledge inference, knowledge graph oriented knowledge inferenceis more diversified, due to the simpleness, intuition, flexibility, and richness of knowledge representation inknowledge graph. Starting with the basic conceptof knowledge inference, this paper presents a survey on the recently developed methods for knowledge graph oriented knowledge inference. Specifically, the research progress is reviewed in detail from two aspects:one-step inference andmulti-step inference, each including rulebased inference, distributed embedding based inference, neural network based inference and hybrid inference. Finally, future research directions and outlook of knowledge graph oriented knowledge inference are discussed and prospected.