自适应自动框架：面向开放式任务流的智能体系统部署的持续自我改进
Adaptive Auto-Harness: Sustained Self-Improvement for Agentic System Deployment on Open-Ended Task Streams
AI总结：
提出自适应自动框架（Adaptive Auto-Harness），通过状态化多智能体进化器、带求解时路由的框架树和人工引导机制，解决开放式任务流中自动框架性能退化问题，在多个流上超越现有基线。
摘要：
自动框架系统（如A-Evolve、GEPA和Meta-Harness）通过从执行反馈中优化提示、技能、工具、记忆和支持基础设施来改进LLM智能体，但它们通常在固定的离线基准上进行评估。实际部署中呈现的是开放式任务流：历史记录无固定终点增长，异构任务需要不同的框架，问题分布随时间变化。这些挑战使得单一反复密集更新的框架变得脆弱，导致性能退化，准确率早期达到峰值后下降。这激发了具有任务自适应性的持续框架构建。我们引入了自适应自动框架（Adaptive Auto-Harness），一个针对此类流的框架和系统。该框架将到 oracle 框架的差距分解为进化损失和适应损失。系统通过状态化多智能体进化器、带求解时路由的框架树以及针对历史缺乏所需信号情况的人工引导钩子来解决这些损失。在预测市场、安全竞赛和事件预测流中，自适应自动框架优于五个现有的自动框架基线，消融实验将收益归因于更好的构建、路由或针对性的人工引导。代码可在 https://github.com/A-EVO-Lab/AdaptiveHarness 获取。
英文摘要：
Auto-harness systems such as A-Evolve, GEPA, and Meta-Harness improve LLM agents by optimizing prompts, skills, tools, memories, and supporting infrastructure from execution feedback, but they are typically evaluated on fixed offline benchmarks. Real deployments instead present open-ended task streams: histories grow without a fixed endpoint, heterogeneous tasks require different harnesses, and problem distributions shift over time. These challenges make a single repeatedly and densely updated harness brittle, causing performance degradation as accuracy peaks early and then declines. This motivates sustained harness construction with task-wise adaptation. We introduce Adaptive Auto-Harness, a framework and system for such streams. The framework decomposes the gap to an oracle harness into evolution loss and adaptation loss. The system addresses these losses with a stateful multi-agent evolver, a harness tree with solve-time routing, and human-steering hooks for cases where history lacks the needed signal. Across prediction-market, security-competition, and event-forecasting streams, Adaptive Auto-Harness outperforms five existing auto-harness baselines and ablations attribute gains to better construction, routing, or targeted human steering. Code is available in \href{https://github.com/A-EVO-Lab/a-evolve/tree/release/adaptive-auto-harness}{Link}.
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。
