arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2607.07702cs.CL
收藏

从噪声轨迹到根本原因:用于智能体优化的结构轨迹分析和因果提取

From Noisy Traces to Root Causes: Structural Trajectory Analysis and Causal Extraction for Agent Optimization

Microsoft Research(微软研究院);University of Chinese Academy of Sciences(中国科学院大学)

Ying Chang, Jiahang Xu, Xuan Feng, Chenyuan Yang, Peng Cheng, Yuqing Yang

AI总结:

研究针对长期智能体优化中实际执行轨迹难直接用于优化的问题,提出STRACE框架。该框架在批次和轨迹内分别进行故障模式挖掘与因果定位,去除冗余和非因果步骤,实验表明其显著优于基线,提升了智能体优化成功率。

摘要:

长期智能体的优化越来越依赖基于反射的机制,其中大语言模型(LLM)作为优化器来诊断智能体故障并改进智能体策略。然而,实际执行轨迹难以直接用于优化:大型轨迹集合往往冗余且异构,导致优化效率低下且容易过度拟合低价值故障;同时,每个单独的轨迹也包含许多无关步骤,而诸如截断或滑动窗口等简单的上下文缩减方法可能会丢弃因果重要证据并产生误导性的优化信号。为了解决这一困境,我们引入了STRACE(结构轨迹分析和因果提取),这是一个构建高信噪比优化上下文以进行更精确和有效优化的框架。在批次级别,STRACE挖掘故障模式以过滤冗余轨迹并保留代表性故障;在每个选定的轨迹内,它在文本依赖图上执行因果定位,以去除非因果步骤并识别真正的优化根本原因模块。实证结果表明,STRACE明显优于标准上下文过滤基线。值得注意的是,在具有挑战性的形式验证任务(VeruSAGE-Bench)上,它成功地优化了人类专家设计的智能体,成功率提高了1.4倍(从42.5%提高到58.5%)。代码可在该https URL获取。

英文摘要:

The optimization of long-horizon agents increasingly relies on reflection-based mechanisms, where a large language model (LLM) acts as an optimizer to diagnose agent failures and improve agent policies. However, real execution traces are difficult to use directly for optimization: large trace collections are often redundant and heterogeneous, making optimization inefficient and prone to overfitting to low-value failures; meanwhile, each individual trajectory also contains many irrelevant steps, while naive context reduction methods such as truncation or sliding windows can discard causally important evidence and produce misleading optimization signals. To resolve this dilemma, we introduce STRACE (Structural TRajectory Analysis and Causal Extraction), a framework that constructs high signal-noise optimization contexts for more precise and effective optimization. At the batch level, STRACE mines failure patterns to filter redundant traces and retain representative failures; within each selected trace, it performs causal localization over a textual dependency graph to remove non-causal steps and identify the true root-cause module for optimization. Empirical results demonstrate that STRACE significantly outperforms standard context-filtering baselines. Notably, on a challenging formal verification task (VeruSAGE-Bench), it successfully optimizes human-expert designed agents, delivering $1.4\times$ success-rate improvement (42.5% to 58.5%). The code is available at https://github.com/moomight/STRACE .

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。

↑