让AI从错误中自动提炼出可复用的精准技能，提升任务成功率。
AutoRefine: Compiling Trajectories into Validated Typed Agent Artifacts
- 通过对比失败与成功轨迹，自动生成无类型依赖的修复方案。
- 在TravelPlanner上达到80.56%成功率，远超基线的50.0%。
- 适合需要长期学习和技能积累的复杂任务系统开发者。
大型语言模型代理反复遇到相似任务，但现有系统将所有经验固化为单一预定义结构，导致不同需求（如局部约束、可复用流程、委派目标）无法匹配合适表达形式。本文提出AutoRefine，将轨迹学习视为类型化产物编译。它通过对比失败与成功轨迹，生成与类型无关、基于证据的干预规范，并将其编译为首个满足运行时所有权顺序的规则、技能或有限子代理：所选结构必须拥有所有指定的观察、状态变量、依赖决策及完成条件。验证机制同样遵循此逻辑：类型特定的契约门检测生成对象是否实现其声明边界；回放门仅在改进源失败案例且不退化保留案例时才接纳。使用GPT-5.6-terra作为共享主干，AutoRefine在ALFWorld、ScienceWorld、TravelPlanner和SpreadsheetBench上取得最高成功率，与SkillCraft持平；在TravelPlanner上达80.56%，显著优于最强基线50.0%。移除边界闭合或回放验证分别损失15.00和16.11个百分点，为各项策略中最大。在纵向TravelPlanner流中，经历60次学习任务后，保留测试成功率维持在89–91%，先前任务无性能下降；冻结仓库提升全部25个源-目标对表现，域内平均提升14.20点，跨域为6.99点。
原文摘要 · Abstract (English)
Large language model agents repeatedly encounter related tasks, yet systems that learn from trajectories commit every lesson to one predefined artifact form. A local constraint, a reusable procedure, and a delegated objective require different amounts of runtime ownership, so one form either under-specifies the correction or wraps it in execution machinery it does not need. We present AutoRefine, which treats trajectory learning as typed artifact compilation. It contrasts failed against successful trajectories to derive a type-neutral, evidence-linked intervention specification, then compiles that specification into the first Rule, Skill, or bounded Subagent that closes it under a runtime-relative ownership order: the selected schema must own every specified observation, state variable, dependent decision, and completion condition. Validation is stated in the same terms: a type-specific contract gate tests whether the generated object realizes its declared boundary, and a replay gate admits it only when it improves the correction cases linked to its source failures without regression on preservation cases. With GPT-5.6-terra as the shared backbone, AutoRefine records the highest success on ALFWorld, ScienceWorld, TravelPlanner, and SpreadsheetBench, and ties the best result on SkillCraft; on TravelPlanner it reaches 80.56% success against 50.0% for the strongest baseline. Removing boundary closure or replay validation costs 15.00 and 16.11 percentage points, the two largest losses among our construction and admission policies. In a longitudinal TravelPlanner stream, the repository holds 89--91% held-out success after 60 learning tasks with no net loss on previously solved tasks, and frozen repositories improve all 25 evaluated source--target pairs, more within a domain (14.20 points on average) than across domains (6.99).
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。