大型语言模型智能体能胜过纳什（均衡）吗？测试自玩多智能体游戏中的去中心化协调
Do LLMs Beat Nash? Testing Decentralized Coordination in Self-Play Multi-Agent Games
浏览论文内容
中文总结 AI 辅助
该研究测试无通信时LLM智能体能否超越纳什均衡，构建无通信博弈基准发现部分前沿托管模型可超越纳什，开放权重模型收益有限且无法迁移至多智能体团队。
中文摘要 AI 辅助
在没有中央控制器部署的大型语言模型（LLM）智能体，通常被认为需要通信才能协调其行动。我们探究不进行通信时仍能实现什么：当同一模型的独立实例无法通信时，它们能否充分推理出对应智能体的情况，从而超越非协调博弈的标准博弈论基准（纳什均衡）？我们推出了一个一次性无通信博弈基准，其中13种语言模型的每一种仅被告知其对应智能体运行相同模型，并针对基础博弈的纳什均衡进行评估。在涵盖7种原型、每个玩家拥有2至10个行动的双人矩阵博弈中，两种前沿托管模型始终超越纳什基准，在数种原型中接近最优联合结果；而大多数开放权重模型仅实现部分收益，且收益随博弈结构差异显著变化。在拥有4个及以上可互换智能体的团队博弈中，性能大幅下降，尤其随着行动空间增大，这表明驱动双人博弈自玩收益的能力无法迁移至更大的多智能体团队。
英文摘要
Large language model agents deployed without a central controller are often assumed to require communication to coordinate their actions. We ask what remains possible without it: when independent instances of the same model cannot communicate, can they still reason about their counterparts well enough to exceed the standard game-theoretic baseline for uncoordinated play? We introduce a benchmark of one-shot, no-communication games in which each of thirteen language models is told only that its counterparts are running the same model and is evaluated against the Nash equilibrium of the underlying game. In two-player matrix games spanning seven archetypes and two to ten actions per player, two frontier-hosted models consistently exceed their Nash benchmark, approaching the optimal joint outcome in several archetypes, while most open-weight models achieve only partial gains that vary sharply by game structure. Performance degrades substantially in team-based games with four or more interchangeable agents, particularly as the action space grows, suggesting that whatever capability drives self-play gains in dyadic games does not transfer to larger multi-agent teams.
发表机构
- McGill University（麦吉尔大学）
- School of Computer Science, McGill University（麦吉尔大学计算机学院）
- Centre for Intelligent Machines（智能机器中心）
机构由 AI 辅助整理，请以论文原文为准。
