arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.03053cs.AIcs.CLcs.MA
收藏

MAS-ProVe: 理解多智能体系统的进程验证

MAS-ProVe: Understanding the Process Verification of Multi-Agent Systems

Rutgers University（罗杰斯大学）；Salesforce AI Research（Salesforce人工智能研究）

Vishal Venkataramani, Haizhou Shi, Zixuan Ke, Austin Xu, Xiaoxiao He, Yingbo Zhou, Semih Yavuz, Hao Wang, Shafiq Joty

备注：
ICML 2026 Regular
阅读 PDF
arXiv 页面 ↗HTML 原文 ↗

AI总结：

MAS-ProVe研究多智能体系统进程验证的有效性，发现LLM作为法官表现优于奖励方法，但验证过程仍面临上下文长度与性能的权衡问题。

摘要：

多智能体系统（MAS）基于大语言模型（LLMs）构建时，其推理轨迹往往表现出高度的变异性。进程验证，即评估轨迹中的中间步骤，已在一般推理设置中显示出潜力，并被建议作为指导MAS协调的潜在工具；然而，其在MAS中的实际有效性仍不清楚。为填补这一空白，我们提出了MAS-ProVe，对多智能体系统的进程验证进行了系统性的实证研究。我们的研究涵盖了三种验证范式（LLM作为法官、奖励模型和进程奖励模型），在两个验证粒度级别（智能体级别和迭代级别）上进行评估。我们进一步检验了五个代表性的验证器和四种上下文管理策略，并在多个推理基准上对六个多智能体系统框架进行了实验。我们发现，进程级验证并不总是能提高性能，且经常表现出高度的变异性，突显了可靠评估部分多智能体轨迹的难度。在所研究的方法中，LLM作为法官通常优于基于奖励的方法，训练过的法官优于通用的LLMs。我们进一步观察到LLMs作为法官与作为单个智能体之间存在小的性能差距，并识别出验证中的上下文长度与性能之间的权衡。总体而言，我们的结果表明，有效的、稳健的多智能体系统进程验证仍是一个开放性挑战，需要在当前范式之外进一步取得进展。代码可在https://github.com/Wang-ML-Lab/MAS-ProVe上获得。

英文摘要：

Multi-Agent Systems (MAS) built on Large Language Models (LLMs) often exhibit high variance in their reasoning trajectories. Process verification, which evaluates intermediate steps in trajectories, has shown promise in general reasoning settings, and has been suggested as a potential tool for guiding coordination of MAS; however, its actual effectiveness in MAS remains unclear. To fill this gap, we present MAS-ProVe, a systematic empirical study of process verification for multi-agent systems (MAS). Our study spans three verification paradigms (LLM-as-a-Judge, reward models, and process reward models), evaluated across two levels of verification granularity (agent-level and iteration-level). We further examine five representative verifiers and four context management strategies, and conduct experiments over six diverse MAS frameworks on multiple reasoning benchmarks. We find that process-level verification does not consistently improve performance and frequently exhibits high variance, highlighting the difficulty of reliably evaluating partial multi-agent trajectories. Among the methods studied, LLM-as-a-Judge generally outperforms reward-based approaches, with trained judges surpassing general-purpose LLMs. We further observe a small performance gap between LLMs acting as judges and as single agents, and identify a context-length-performance trade-off in verification. Overall, our results suggest that effective and robust process verification for MAS remains an open challenge, requiring further advances beyond current paradigms. Code is available at https://github.com/Wang-ML-Lab/MAS-ProVe.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

↑