arXiv cs.MA· Xinwen Liu, Zhuocheng Pan, Isabella Zhu, Jawei Zhang, Xudong Liu, Tianyu Wo·· 10 天前AI 评分29
通信拓扑如何诊断多智能体 LLM 故障：MAScope 两阶段框架
Know the Shape, Find the Fault: Topology-Conditioned Diagnosis of Multi-Agent LLM Failures
AI 导读
研究者提出 MAScope 两阶段框架，从无拓扑标签的执行轨迹中恢复通信拓扑并据此诊断多智能体 LLM 的协调故障。在 851 条 MAST-clean 轨迹上，真实拓扑上下文将 gpt-mini 的 Macro-F1 从 0.173 提升至 0.350，预测拓扑达 0.346，接近仅用轨迹的 gpt-5.4 基线 0.372。
正文 · 原文
Abstract:Multi-agent LLM systems coordinate task execution through exchanges of information among agents. When coordination breaks down, similar symptoms in execution traces can reflect different problems in how information is passed, used, or verified. Communication topology captures how agents exchange information and provides structural cues for distinguishing coordination failure modes. Using these cues for diagnosis requires establishing how topology relates to failure patterns and recovering the relevant structure from execution traces that lack explicit topology labels. We analyze the relationship between communication topology and failure patterns and introduce MAScope, a two-stage framework for topology-conditioned diagnosis. Its Trace Structural Extractor TSE recovers communication topology from heterogeneous execution traces by grounding an interaction graph in message evidence. The Topology-Conditioned Judge TC-Judge then classifies failures using the trace, predicted topology, an empirical failure prior estimated from separate labeled traces, and a short description of topology-specific failure patterns. Under a fixed orchestration structure, the recovered topology can be reused across executions. Experimental results show a statistically significant association between communication topology and failure type, with $\chi^2 = 409.9$ and $p = 1.2 \times 10^{-70}$. On the \num{851} MAST-clean traces, ground-truth topology context raises gpt-mini's Macro-F1 from $0.173$ to $0.350$. With predicted topology, the pipeline achieves $0.346$, approaching the trace-only gpt-5.4 baseline of $0.372$. For \num{1000} traces under a fixed orchestration structure, the projected pipeline cost, including one topology extraction, is approximately $6\%$ of repeated gpt-5.4 diagnosis cost. These results show that topology-conditioned context improves failure diagnosis and supports lower-cost deployment.
| Subjects: | Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.10126 [cs.MA] |
| (or arXiv:2610.10126v1 [cs.MA] for this version) | |
| https://doi.org/10.48550/arXiv.2610.10126 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Zhuocheng Pan [view email]
[v1]
Wed, 7 Oct 2026 14:07:48 UTC (1,968 KB)
来源：arXiv cs.MA · arxiv.org