让大模型学会持续学习并复用技能，提升长任务决策能力。
Co-Evolving LLM Decision and Skill Bank Agents for Long-Horizon Tasks
- 双代理协同：决策代理从可学习技能库调用技能，技能库代理自动挖掘新技能。
- 在6个游戏上，8B模型平均奖励提升超25.1%，优于4个前沿基线。
- 适合研究长序列决策、技能复用的AI系统开发者和游戏智能体研究者。
长时序交互环境是评估智能体技能使用能力的理想场景，要求多步推理、跨时间步的技能链式调用，以及在延迟奖励和部分可观测条件下的稳健决策。游戏是评估智能体技能应用的优质测试平台。尽管大语言模型（LLM）作为游戏智能体具有潜力，但其在长时序任务中常因缺乏发现、保留和重用结构化技能的机制而表现不稳定。我们提出COSPLAY框架，其中一个LLM决策代理从可学习的技能库中检索技能以指导动作，同时一个由智能体管理的技能流水线从无标签的轨迹中自动发现可复用技能，构建技能库。该框架使决策代理更优地学习技能检索与动作生成，同时技能库代理持续提取、优化并更新技能及其契约。在六个游戏环境上的实验表明，采用8B基础模型的COSPLAY在单人游戏基准上相比四个前沿基线平均奖励提升超过25.1%，并在多人社交推理游戏中保持竞争力。
原文摘要 · Abstract (English)
Long horizon interactive environments are a testbed for evaluating agents skill usage abilities. These environments demand multi step reasoning, the chaining of multiple skills over many timesteps, and robust decision making under delayed rewards and partial observability. Games are a good testbed for evaluating agent skill usage in environments. Large Language Models (LLMs) offer a promising alternative as game playing agents, but they often struggle with consistent long horizon decision making because they lack a mechanism to discover, retain, and reuse structured skills across episodes. We present COSPLAY, a co evolution framework in which an LLM decision agent retrieves skills from a learnable skill bank to guide action taking, while an agent managed skill pipeline discovers reusable skills from the agents unlabeled rollouts to form a skill bank. Our framework improves both the decision agent to learn better skill retrieval and action generation, while the skill bank agent continually extracts, refines, and updates skills together with their contracts. Experiments across six game environments show that COSPLAY with an 8B base model achieves over 25.1 percent average reward improvement against four frontier LLM baselines on single player game benchmarks while remaining competitive on multi player social reasoning games.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。