Learning Two-Player Mixture Markov Games: Kernel Function Approximation and Correlated Equilibrium

Neural Information Processing Systems 

The case of two-player Markov Games (MG) has been of particular interest.