XDO: ADoubleOracleAlgorithmfor Extensive-FormGames

Policy Space Response Oracles (PSRO) is a reinforcement learning (RL) algorithm for two-player zero-sum games that has been empirically shown to find approximate Nash equilibria in large games.