Hacker News· gmays·· 6 天前AI 评分29
Harness Learning：通过执行反馈实现可泛化的测试时自适应
Harness Learning Enables Generalizable Test-Time Adaptation
AI 导读
研究者提出 harness learning，训练 proposer 模型利用执行反馈修改 solver 的 harness（组织模型调用、工具使用与信息流的可执行程序），并将该过程形式化为对可执行程序的元学习，harness 修订相当于梯度适应中的权重更新。
正文 · 原文
Abstract:A language-model agent is jointly defined by its model and its harness, the executable program that organizes model calls, tool use, and information flow. Because different tasks call for different ways of organizing these operations, the harness needs to be adapted using feedback from the task at hand. We introduce harness learning, which trains a proposer model to revise a solver's harness using execution feedback. We formulate this process as meta-learning over executable programs, with harness revisions playing the role of weight updates in gradient-based adaptation. We train the proposer with reinforcement learning, using the task performance of revised harnesses as the reward. At test time, the proposer uses feedback from successive executions on a new task to refine the harness, without performing any parameter-space update. Experiments on reasoning and multi-hop question answering show that harness learning improves revision quality and that the ability to adapt at test time transfers to unseen tasks. Policies trained on individual revisions can continue improving harnesses over multiple rounds, while the benefits of training on revision sequences vary across settings. These findings suggest a path towards continually learning agents that turn accumulated experience into generalizable improvements.
| Subjects: | Computation and Language (cs.CL); Machine Learning (cs.LG) |
| Cite as: | arXiv:2609.35738 [cs.CL] |
| (or arXiv:2609.35738v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2609.35738 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yuda Song [view email]
[v1]
Mon, 28 Sep 2026 17:52:54 UTC (768 KB)
来源：Hacker News · arxiv.org