Title: Monitoring LLM-based Multi-Agent Systems Against Corruptions via Node Evaluation Show Chinese title

Title: 通过节点评估监控基于大语言模型的多智能体系统中的损坏情况

Abstract: Large Language Model (LLM)-based Multi-Agent Systems (MAS) have become a popular paradigm of AI applications. However, trustworthiness issues in MAS remain a critical concern. Unlike challenges in single-agent systems, MAS involve more complex communication processes, making them susceptible to corruption attacks. To mitigate this issue, several defense mechanisms have been developed based on the graph representation of MAS, where agents represent nodes and communications form edges. Nevertheless, these methods predominantly focus on static graph defense, attempting to either detect attacks in a fixed graph structure or optimize a static topology with certain defensive capabilities. To address this limitation, we propose a dynamic defense paradigm for MAS graph structures, which continuously monitors communication within the MAS graph, then dynamically adjusts the graph topology, accurately disrupts malicious communications, and effectively defends against evolving and diverse dynamic attacks. Experimental results in increasingly complex and dynamic MAS environments demonstrate that our method significantly outperforms existing MAS defense mechanisms, contributing an effective guardrail for their trustworthy applications. Our code is available at https://github.com/ChengcanWu/Monitoring-LLM-Based-Multi-Agent-Systems. Show Chinese abstract

Abstract: 基于大型语言模型（LLM）的多智能体系统（MAS）已成为人工智能应用的一种流行范式。 然而，MAS中的可信度问题仍然是一个关键关注点。 与单智能体系统中的挑战不同，MAS涉及更复杂的通信过程，使其容易受到腐蚀攻击。 为了缓解这一问题，已经开发了几种基于MAS图表示的防御机制，其中智能体代表节点，通信形成边。 然而，这些方法主要集中在静态图防御上，试图在固定图结构中检测攻击或优化具有特定防御能力的静态拓扑。 为了解决这一限制，我们提出了一种针对MAS图结构的动态防御范式，该范式持续监控MAS图中的通信，然后动态调整图拓扑，准确地破坏恶意通信，并有效防御不断演变和多样化的动态攻击。 在日益复杂和动态的MAS环境中进行的实验结果表明，我们的方法显著优于现有的MAS防御机制，为它们的可信应用提供了一个有效的保障措施。 我们的代码可在 https://github.com/ChengcanWu/Monitoring-LLM-Based-Multi-Agent-Systems 获取。