arXiv cs.MA· Kunal Jha, Gon\c{c}alo Faria, Shuyue Stella Li, Scott Geng, Faeze Brahman, Noah A. Smith, Yulia Tsvetkov, Max Kleiman-Weiner, Natasha Jaques·· 11 小时前AI 评分48
s³：用协作式多智能体强化学习实现群体规模的简单扩展
s$^3$: Simple Scaling of Swarms via Cooperative Multi-agent RL
AI 导读
研究者提出 s³，首个让智能体自行决定独自工作时长、何时共享发现或读取队友消息的多智能体强化学习（MARL）方法，无需预定义组织或角色。在训练与推理预算匹配下，用两倍于训练规模的智能体群体评估时，s³ 解决的协作问题数量超过最强基线 25 倍；在并行探索任务中，通信智能体将重复工作减少近 50%，并比 best-of-N 高出最多 9.4 个百分点。
正文 · 原文
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 s$^3$, the first multi-agent reinforcement learning method in which agents decide for themselves how long to work alone and when to share findings or read teammates' messages, with no predefined organization or roles. Our s$^3$ agents learn to take different approaches, read each other's work, exhibit emergent roles that complement each other in the swarm, and generalize to bigger teams. We study both coordination, where agents must combine private information, and parallel exploration, where sharing discoveries can make search more efficient. With matched training and inference budgets, s$^3$, swarms evaluated with twice as many agents as in training solve over 25$\times$ more coordination problems than the strongest baseline, and in exploration, communicating agents reduce repeated work by nearly 50% and outperform best-of-N by up to 9.4 percentage points. Rather than being designed in advance, s$^3$ enables organization to emerge from a shared goal. This opens the door to capable systems built from many models that learn to work together.
| Subjects: | Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.08615 [cs.MA] |
| (or arXiv:2611.08615v1 [cs.MA] for this version) | |
| https://doi.org/10.48550/arXiv.2611.08615 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Kunal Jha [view email]
[v1]
Thu, 5 Nov 2026 19:07:15 UTC (11,552 KB)
来源：arXiv cs.MA · arxiv.org