arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2401.05133cs.AIcs.MA
收藏

超越对称零和博弈的神经种群学习

Neural Population Learning beyond Symmetric Zero-sum Games

Google DeepMind(谷歌DeepMind);University College London(伦敦大学学院)

Siqi Liu, Luke Marris, Marc Lanctot, Georgios Piliouras, Joel Z. Leibo, Nicolas Heess

AI总结:

本文提出NeuPL-JPSRO算法,通过技能迁移学习在n人一般和博弈中高效寻找均衡,并在OpenSpiel、MuJoCo和夺旗战中验证了其收敛性和可扩展性。

摘要:

我们研究了在n人一般和博弈中寻找均衡的计算高效方法,特别是那些涉及复杂视觉运动技能的博弈。我们展示了现有方法如何在这种设置中挣扎,无论是在计算上还是在理论上。然后我们介绍了NeuPL-JPSRO,这是一种神经种群学习算法,它受益于技能的迁移学习,并收敛到博弈的粗相关均衡(CCE)。我们展示了在OpenSpiel游戏套件中的经验收敛性,并通过精确博弈求解器进行了严格验证。然后我们将NeuPL-JPSRO部署到复杂领域,我们的方法在MuJoCo控制域中实现了自适应协调,并在夺旗战中实现了技能迁移。我们的工作表明,均衡收敛的种群学习可以大规模且通用化地实现,为解决具有混合动机的异构玩家之间的现实世界博弈铺平了道路。

英文摘要:

We study computationally efficient methods for finding equilibria in n-player general-sum games, specifically ones that afford complex visuomotor skills. We show how existing methods would struggle in this setting, either computationally or in theory. We then introduce NeuPL-JPSRO, a neural population learning algorithm that benefits from transfer learning of skills and converges to a Coarse Correlated Equilibrium (CCE) of the game. We show empirical convergence in a suite of OpenSpiel games, validated rigorously by exact game solvers. We then deploy NeuPL-JPSRO to complex domains, where our approach enables adaptive coordination in a MuJoCo control domain and skill transfer in capture-the-flag. Our work shows that equilibrium convergent population learning can be implemented at scale and in generality, paving the way towards solving real-world games between heterogeneous players with mixed motives.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑