arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2510.10035cs.AI
收藏

失败驱动的工作流优化

Failure-Driven Workflow Refinement

Sun Yat-sen University(中山大学);X-Era AI Lab(X-Era人工智能实验室)

Jusheng Zhang, Kaitong Cai, Qinglin Zeng, Ningyuan Liu, Stephen Fan, Ziliang Chen, Keze Wang

AI总结:

本文指出现有LLM工作流优化存在信息坍缩缺陷,提出将优化目标转为最小化高维失败特征空间中的预期失败质量,并引入CE-Graph框架通过反例近似失败分布并进行算子约束的图编辑,以更低成本实现了更高的鲁棒性。

摘要:

优化基于LLM的工作流通常被表述为全局搜索,其中候选工作流基于标量指标进行评估。然而,这种范式存在一个关键缺陷:信息坍缩。通过将丰富的多步执行轨迹简化为简单的成功/失败信号,现有方法对失败的底层结构视而不见,从根本上阻碍了它们对工作流失败分布的建模。我们将这一挑战重新概念化为一个分布问题。我们提出了一种新范式,其优化目标不是最大化标量分数,而是直接最小化工作流的预期失败质量,即定义在高维失败特征空间(FSS)上的失败概率密度函数的积分。这种分布视角使我们能够从低效的零阶优化转向对失败景观本身进行有原则的类梯度下降。我们引入了CE-Graph,这是一个通过新颖的失败驱动优化过程来实施该范式的框架。CE-Graph从反例池中近似失败分布,将其最密集区域识别为重复失败模式,并通过提议与验证机制应用有针对性的算子约束图编辑,以贪婪地减少失败质量。在数学、代码和QA基准测试中,我们的CE-Graph以显著低于强基线的成本实现了更高的鲁棒性。这表明系统的可靠性并非源于避免失败,而是源于系统性地学习和重塑其失败分布的几何结构。

英文摘要:

Optimizing LLM-based workflows is typically formulated as a global search, where candidate workflows are evaluated based on a scalar metric. This paradigm, however, suffers from a critical flaw: information collapse. By reducing rich, multi-step execution traces to simple success/failure signals, existing methods are rendered blind to the underlying structure of failures, fundamentally preventing them from modeling the workflow's failure distribution. We reconceptualize this challenge as a distributional problem. We propose a new paradigm where the optimization goal is not to maximize a scalar score, but to directly minimize a workflow's Expected Failure Mass, i.e., the integral of its failure probability density function defined over a high-dimensional Failure Signature Space (FSS). This distributional lens allows us to move from inefficient, zero-order optimization to a principled, gradient-like descent on the failure landscape itself. We introduce CE-Graph, a framework that operationalizes this paradigm through a novel, failure-driven refinement process. CE-Graph approximates the failure distribution from a pool of counterexamples, identifies its densest regions as recurring failure modes, and applies targeted, operator-constrained graph edits via a Propose-and-Verify mechanism to greedily reduce the failure mass. On math, code, and QA benchmarks, our CE-Graph achieves higher robustness at a significantly lower cost than strong baselines. This suggests that a system's reliability emerges not from avoiding failures, but from systematically learning and reshaping the geometric structure of its failure distributions.

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。

↑