s3：用协作多智能体强化学习实现群体简单扩展
原标题：s$^3$: Simple Scaling of Swarms via Cooperative Multi-agent RL
AI 摘要
论文提出 s3，一种基于协作式多智能体强化学习的方法，让语言模型智能体自行学习何时独立工作、何时共享发现或读取队友消息，无需预设组织架构或角色。在训练与推理预算匹配的条件下，s3 群体在协调任务上以两倍于训练规模的智能体数量解决了超过最强基线 25 倍的问题，在探索任务中减少近 50% 重复工作，并比 best-of-N 高出最多 9.4 个百分点。该方法表明多智能体组织可以从共享目标中涌现，而非由设计者预先指定。
正文节选
s3: Simple Scaling of Swarms via Cooperative Multi-agent RL Abstract Groups can achieve more than individuals in the same amount of time, but only if their members coordinate: they must decide who works on what and what to share. However, this organization in multi-agent systems today usually comes from arbitrary, static harnesses designed a priori without adapting to agent behaviors or task specifications. We ask whether agents can instead learn to organize themselves. We introduce s3, the firs