Learning Two-Player Mixture Markov Games: Kernel Function Approximation and Correlated Equilibrium
This addresses the challenge of exploration in high-dimensional function spaces for game theory and reinforcement learning, offering a novel method with theoretical guarantees.
The paper tackles learning Nash equilibria in two-player zero-sum Markov Games with nonlinear function approximation using RKHS, proposing an online algorithm that achieves an O(√T) regret with polynomial computational complexity under mild assumptions.
We consider learning Nash equilibria in two-player zero-sum Markov Games with nonlinear function approximation, where the action-value function is approximated by a function in a Reproducing Kernel Hilbert Space (RKHS). The key challenge is how to do exploration in the high-dimensional function space. We propose a novel online learning algorithm to find a Nash equilibrium by minimizing the duality gap. At the core of our algorithms are upper and lower confidence bounds that are derived based on the principle of optimism in the face of uncertainty. We prove that our algorithm is able to attain an $O(\sqrt{T})$ regret with polynomial computational complexity, under very mild assumptions on the reward function and the underlying dynamic of the Markov Games. We also propose several extensions of our algorithm, including an algorithm with Bernstein-type bonus that can achieve a tighter regret bound, and another algorithm for model misspecification that can be applied to neural function approximation.