Policy Optimization finds Nash Equilibrium in Regularized General-Sum LQ Games