AI RESEARCH
Data-Augmented Game Starts for Accelerating Self-Play Exploration in Imperfect Information Games
arXiv CS.LG
•
ArXi:2605.14379v1 Announce Type: new Finding approximate equilibria for large-scale imperfect-information competitive games such as StarCraft, Dota, and CounterStrike remains computationally infeasible due to sparse rewards and challenging exploration over long horizons. In this paper, we propose a multi-agent starting-state sampling strategy designed to substantially accelerate online exploration in regularized policy-gradient game methods for two-player zero-sum (2p0s) games.