arXiv cs.CL· Zixuan Zhu, Yitong Hu, Yong Dai, Junfeng Fang, Chunyang Jiang, Senkang Hu, Yuzhi Zhao·· 14 小时前AI 评分38
Unified Context Evolution（UCE）：让 LLM 智能体经验跨任务持续进化的无梯度框架
Unified Context Evolution for LLM Agents
AI 导读
研究者提出 Unified Context Evolution（UCE），一个无梯度框架，将 LLM 智能体的经验外化为由 Memory、Strategy、Workflow、Skill 四类可进化上下文单元（ECU）组成的持续演化库，按类型生成、检索、依使用结果评分并淘汰失效单元。
正文 · 原文
Abstract:LLM-based agents can solve multi-step interactive tasks by combining reasoning with environment feedback, yet each episode starts from the same fixed context and any useful strategy discovered along the way is lost once the task ends. Existing approaches either limit learning to the current task or pool all experience into a single untyped store, without distinguishing knowledge types, tracking quality through use, or balancing what the library still lacks. We introduce Unified Context Evolution (UCE), a gradient-free framework that externalizes agent experience into an evolving library of typed Evolvable Context Units (ECUs). UCE decomposes experience into four complementary types (Memory, Strategy, Workflow, and Skill), each generated from trajectories under type-specific conditions, retrieved at decision time, scored through repeated usage outcomes, and pruned when no longer valuable. A scheduling module allocates each cycle's generation budget toward the types where the library is weakest. Across two interactive benchmarks, UCE raises ALFWorld success from 75.4% to 96.3% and WebShop task score from 45.1% to 61.3%, and remains effective across alternative actor reasoning backbones without model retraining.
| Comments: | Updated to the EMNLP 2026 camera-ready version |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2606.02304 [cs.CL] |
| (or arXiv:2606.02304v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2606.02304 arXiv-issued DOI via DataCite |
Submission history
From: Zixuan Zhu [view email]
[v1]
Mon, 1 Jun 2026 14:25:29 UTC (434 KB)
[v2]
Sat, 24 Oct 2026 20:44:22 UTC (1,226 KB)
来源：arXiv cs.CL · arxiv.org