网络安全知识图谱研究综述

A comprehensive review of network security knowledge graph

摘要:

随着互联网规模不断增大,产生了越来越多的网络安全数据,这些数据存在多源异构、数据缺失、噪声、不一致等问题,严重影响网络安全数据的质量,知识图谱具有数据统一、 可解释、可融合推理等特性,可有效应对网络安全数据的这些问题。本文分析了网络安全领域知识图谱的发展和研究现状,围绕知识实体识别、关系抽取和知识图谱补全等知识图谱构建技术,从智能渗透、舆情监测和威胁感知3个方面系统总结了目前的具体应用,给出了下一步研究的方向。在网络空间安全领域,有效的网络空间安全领域知识图谱技术体系,为应对强对抗、高动态环境下的攻防博弈提供知识要素与智能推理提供支撑,同时也是网络空间高级、持续、威胁感知的基础。

Abstract:

As the scale of the Internet continues to expand, more and more network security data are generated. These data are characterized by multi-source heterogeneity, missing data, noise, inconsistency, etc., which severely affect the quality of network security data. Knowl edge graphs possess characteristics such as data unification, interpretability, and fusion rea soning, which can effectively address these issues in network security data. This paper ana lyzed the development and current research status of knowledge graphs in the field of network security. It focused on knowledge graph construction techniques such as knowledge entity recognition, relationship extraction, and knowledge graph completion. From three aspects: intelligent penetration, public sentiment detection, and threat perception, it systematically summarized the current specific applications and provides directions for future research. In the field of cybersecurity, an effective knowledge graph technology system in cyberspace pro vides support for knowledge elements and intelligent reasoning to deal with adversarial at tacks and defense games in high-dynamic environments. It also serves as the foundation for advanced,continuous,andthreat-awarecyberspaceoperations.