arXiv cs.AI· Zhen Bi, Zhizhi Peng, Yan Wang, Xueshu Chen, Haofei Zhang, Yaobo Zhang, Zeyu Yang, Qing Shen, Jungang Lou·· 14 小时前AI 评分38
HaMo：面向长时程持续学习的 Harness-Model 协同框架
Harness-Model Collaboration for Long-Horizon Continual Learning
AI 导读
研究者提出 HaMo，一个结合快速非参数经验积累与慢速参数适应的 Harness-Model 协同框架，将轨迹蒸馏为策略级经验卡片供即时检索，并用经验条件持续学习目标训练模型理解与执行检索到的策略。实验显示 HaMo 在域内与跨域迁移下均优于仅 harness 和仅模型适应方案，同时提升保持与适应能力。
正文 · 原文
Abstract:Long-horizon agents continuously encounter reusable procedural knowledge, yet existing systems typically rely either on model adaptation or external memory, trading persistent learning for rapid reuse. This limitation is especially important for Web agents that repeatedly execute related multi-step workflows. We propose HaMo, a Harness-Model collaborative framework that combines fast non-parametric experience accumulation with slow parametric adaptation. HaMo distills trajectories into strategy-level experience cards for immediate retrieval, while an experience-conditioned continual learning objective trains the model to interpret, ground, and execute retrieved strategies across changing contexts. A retrieval-coverage-based selection mechanism further identifies compact and complementary trajectories for parameter updates. Experiments show that HaMo outperforms harness-only and model-only adaptation under within-domain and cross-domain shifts, improving both retention and adaptation. These results highlight the importance of jointly preserving procedural experience and internalizing the competence to transfer and execute it across tasks.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.34724 [cs.AI] |
| (or arXiv:2610.34724v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.34724 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Zhen Bi [view email]
[v1]
Mon, 26 Oct 2026 02:21:20 UTC (1,880 KB)
来源：arXiv cs.AI · arxiv.org