基于大模型智能体的安全风险态势感知框架构建  

Construction of a Security Risk Situational Awareness Framework Based on LLM-based Agents

作　　者：王明程 王高开 李勇男[1] Wang Mingcheng;Wang Gaokai;Li Yongnan(School of National Security,People’s Public Security University of China,Beijing 100038;School of Economics and Management,Shanxi University,Shanxi Taiyuan 030000)

机构地区：[1]中国人民公安大学国家安全学院,北京100038 [2]山西大学经济与管理学院,山西太原030000

出　　处：《情报理论与实践》2024年第7期190-198,共9页Information Studies:Theory & Application

基　　金：某省部级重点项目“题目略”(项目编号:2023JSZ04);中国人民公安大学基本科研业务费项目“总体国家安全观视域下生物安全情报体系建设研究”(项目编号:2023JKF01ZK04)的成果。

摘　　要：[目的/意义]基于大模型智能体构建安全风险态势感知框架,可以将大模型智能体嵌入安全风险态势感知,为安全情报智能化转型提供参考,为安全风险监测预警效能提升提供支撑。[方法/过程]首先,运用规范研究法,对大模型智能体的研究进展及其赋能下的安全风险态势感知进行概述;其次,在系统论、博弈论、风险评估理论等的指导下,对大模型智能体驱动的安全风险态势感知运作机理进行剖析;最后,构建基于大模型智能体的安全风险态势感知框架并对其内涵进行解析。[结果/结论]构建的安全风险态势感知框架以大模型智能体为基础,智能感知风险态势要素;以多智能体博弈协作为依托,智慧评估安全风险态势;以仿真模拟技术为切入点,智能分析应用风险态势情报。以此系统、精准地刻画安全风险态势,为大模型智能体在安全情报领域的实践应用奠定基础。[Purpose/significance]This study develops a framework for integrating LLM-based agents into the domain of security risk situation awareness.This initiative aims to serve as a benchmark for facilitating the intelligent transformation of security intelligence operations,thereby supporting the advancement of national capabilities in security risk monitoring and early warning.[Method/process]Employing normative research methodologies,this article initially addresses the research progress of LLM-based agents and their empowerment of enhancing security risk situational awareness.Following this,guided by the principles of systems theory,game theory,and risk assessment theory,it delves into the operational mechanisms behind security risk situational awareness as propelled by LLM-based agents.Conclusively,it establishes a framework for security risk situational awareness predicated on LLM-based agents and elucidates its inherent meanings.[Result/conclusion]The framework,grounded on LLM-based agents,intelligently perceives elements of risk situations.It relies on the collaboration and game theory of multiple agents to intelligently evaluate the security risk situation.Using simulation and modeling techniques as an entry point,it intelligently analyzes and applies risk situation intelligence,aiming for a systematic and precise depiction of security risk situations.This lays the foundation for the practical application of LLM-based agents in the field of security intelligence.

关 键 词：大模型智能体 安全情报 情报工作 态势感知 风险监测 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] D526[自动化与计算机技术—控制科学与工程]