Co-Skill：面向技能演化的协作通信框架
Co-Skill: A Collaborative Communication Framework for Skill Evolution
浏览论文内容
中文总结 AI 辅助
提出协作通信框架CCF，通过前缀合并轨迹字典树、渐进式技能树及分离演化方案，解决混合演化中云端与边缘端盲目通信问题，显著降低令牌消耗并提升任务成功率。
中文摘要 AI 辅助
通过技能实现智能体演化，对于基于大语言模型（LLM）的智能体迭代提升任务成功率至关重要。混合演化是一种成本高效的范式，其中云端大语言模型（LLM）负责分析和生成技能，而边缘端小语言模型（SLM）负责执行和内化这些技能。然而，现有的混合方法（如SkillRL）仍然面临成功率低和令牌（token）消耗高的问题。我们发现这源于盲目通信：云端无法感知边缘端的执行能力，而边缘端也不理解云端的分析需求。为此，我们提出了协作通信框架（Collaborative Communication Framework, CCF），以实现高效的边缘-云端演化。CCF通过三种技术实现：（1）一种云端感知的前缀合并轨迹字典树（prefix-merged trajectory trie），边缘端通过合并共享前缀并定位分歧点来压缩轨迹，以便云端高效分析；（2）一种边缘感知的渐进式技能树（progressive skill tree），云端逐步构建分层技能树，以匹配边缘端SLM的执行能力；（3）一种基于这两棵树的协作技能演化方案，该方案以分离的方式演化云端LLM和边缘端SLM，共同提升任务成功率。在ALFWorld和WebShop上的实验表明，与最先进的混合方法相比，CCF将LLM+SLM的令牌消耗降低了15.6%至41.9%，同时持续将任务成功率提升了25.8%至76.4%。
英文摘要
Agent evolution through skills becomes critical for LLM-based agents to iteratively improve task success rate. Hybrid evolution is a cost-efficient paradigm where a cloud LLM analyzes and generates skills while an edge SLM executes and internalizes them. However, existing hybrid methods, such as SkillRL, still suffer from low success rate and high token usage. We find this stems from blind communication: the cloud cannot perceive the edge's execution capability, while the edge does not understand the cloud's analysis needs. We thus propose the Collaborative Communication Framework (CCF) to achieve effective edge-cloud evolution. CCF is realized via three techniques: (1) a cloud-aware prefix-merged trajectory trie where the edge compresses trajectories by merging shared prefixes and pinpointing divergence points for efficient cloud analysis, (2) an edge-aware progressive skill tree where the cloud progressively builds a hierarchical skill tree to match edge SLM execution capability, and (3) a collaborative skill evolution scheme upon these two trees that evolves cloud LLM and edge SLM in a separated way to jointly improve task success rate. Experiments across ALFWorld and WebShop show that CCF reduces LLM+SLM tokens by 15.6%--41.9% over state-of-the-art hybrid methods while consistently improving 25.8%--76.4% task success rate.
