arXiv：cs.LG· Yuxin Xiao, Zhiyun Lu, Jianyu Wang, Yangtian Zhang, Yang Song, Qifan Wang, Bailin Wang·· 11 小时前
ReHarness：面向智能体持续学习的循环式框架
Recurrent Harness for Agentic Continual Learning
AI 导读
研究者提出 ReHarness，一个面向智能体持续学习的循环式框架，在严格在线流式协议下，用前向读写之外的"反向修订"补充 harness 状态更新：每条实例处理完后，用其轨迹作为反馈，回改该实例检索到的反思内容，使反馈沿文本 harness 状态反向传播。在三个持续学习 benchmark 和三个语言模型骨架上，ReHarness 持续优于原始上下文基线、基于记忆与反思的方法以及提示词优化方法。
正文
Abstract:Language model agents increasingly operate over long interaction streams, requiring online adaptation to distribution shifts through continual learning. However, existing agent harnesses typically either evolve their states only through forward information flow or rely on additional replay or validation data for harness optimization. To address this, we propose ReHarness, a recurrent harness framework for agentic continual learning that complements forward read and write with backward revision under a strictly online streaming protocol. We ask what information could have been captured differently in the harness state to better inform the current instance in the stream. Specifically, after each instance, we leverage the resulting trajectory as feedback to revise the reflections retrieved for that instance, thereby propagating feedback backward through the textual harness state. Across three continual learning benchmarks and three language model backbones, ReHarness consistently outperforms strong raw-context baselines, memory- and reflection-based methods, and prompt optimization approaches. Further analysis shows that ReHarness produces source-grounded and meaningful revisions, corrects overgeneralized or outdated experience under distribution shifts, and improves both stability and plasticity during continual learning.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.12668 [cs.LG] |
| (or arXiv:2610.12668v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.12668 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yuxin Xiao [view email]
[v1]
Thu, 8 Oct 2026 19:44:43 UTC (422 KB)
来源：arXiv：cs.LG · arxiv.org