LLM代理在共识任务中常失效，需用经典容错算法增强可靠性。
Resilient Consensus in Agentic AI
- 将LLM代理视为拜占庭共识博弈，测试其在不同通信拓扑下的协作能力。
- 即使理论保证收敛，提示后的LLM代理仍无法达成共识，且不受温度与推理长度影响。
- 引入经典容错滤波器可显著提升共识成功率，效果依赖网络拓扑鲁棒性。
大型语言模型（LLM）代理正被广泛部署于多代理系统中，需协调并就共同决策达成一致。本文探讨经典的容错共识理论（适用于确定性代理）是否适用于可能表现出敌意行为的LLM代理。将LLM协议达成视为拜占庭共识博弈，在完全图和一般通信图上进行受控实验。结果表明，提示过的LLM代理无法实现理论上可达到的一致性：即使经典理论保证存在收敛算法，共识仍会失败，且该失败现象在不同温度和推理深度下持续存在。同时，通过经典容错共识滤波器对代理进行包装，可显著改善一致性表现。滤波器带来的收益取决于底层拓扑本身提供的鲁棒性程度。研究结果表明，经典容错共识理论为智能体人工智能的安全性提供了有价值的分析视角。
原文摘要 · Abstract (English)
Large language model (LLM) agents are increasingly deployed in multi-agent systems where they must coordinate and agree on shared decisions. We ask whether classical resilient consensus theory, developed for deterministic agents, transfers to LLM agents that may behave adversarially. Framing LLM agreement as a Byzantine consensus game, we run controlled experiments on complete and general communication graphs. We find that prompted LLM agents fail to reach agreement that is achievable in principle: consensus can fail even in settings where classical theory guarantees that a convergent algorithm exists, and this failure persists across temperatures and horizons. At the same time, wrapping the agents with classical resilient consensus filters improves agreement. The benefit of filtering depends on how much robustness the underlying topology already provides. Our results suggest that classical resilient consensus theory is a useful lens for the safety of agentic AI.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。