LearningTwo-PlayerMixtureMarkovGames: KernelFunctionApproximationandCorrelated Equilibrium

Neural Information Processing Systems 

We propose anovel online learning algorithm to find aNash equilibrium by minimizing the dualitygap.