arXiv：cs.LG· Lecheng Kong, Like Hui, Nikos Kanakaris, Prithwish Jana, Sahika Genc, Narayanan Sadagopan·· 4 小时前AI 评分50
FreeEvolve：让智能体进化器自动学习优化流程
FreeEvolve: Learning to Evolve Beyond Fixed Loops
AI 导读
论文提出 FreeEvolve，将语言模型智能体进化器中原本手工设计的固定搜索循环也交给进化器自主决定，包括测试什么、收集多少证据、保留哪些候选和何时停止。进化决策遵循一个可编辑的进化技能，并通过元进化在新生成的目标智能体上评分改进。
正文
Abstract:Agent evolvers automate the design of the prompts, skills and workflows around language model agents, yet the optimization process they follow is still designed by hand: a fixed search loop decides how candidates are evaluated, which are kept and when the search stops. We propose FREEEVOLVE, which automates this process as well. An environment specifies the goal, target agent, evaluator, data and resource limits; within these limits, the evolver itself decides what to test, how much evidence to collect, which candidates to pursue and when to stop. These decisions follow an editable evolution skill, which we improve through meta-evolution by scoring each candidate skill on the fresh target agent it produces. The optimization process thus becomes a capability learned from experience rather than a loop engineered in advance. On tau3-bench, ARC-AGI-2, ARC-AGI-3 and Terminal-Bench 2.1, FREEEVOLVE controls the evolution campaign by itself, yet improves the primary held-out metric by 13.6 points on average and matches or exceeds hand-designed evolvers. The learned process keeps improving with experience: meta-evolved skills add 6.9 points over the seed skill on fresh target agents, demonstrating transferability across environments.
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.09197 [cs.AI] |
| (or arXiv:2610.09197v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.09197 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Lecheng Kong [view email]
[v1]
Tue, 6 Oct 2026 22:47:19 UTC (223 KB)
来源：arXiv：cs.LG · arxiv.org