On the Convergence of Smooth Regularized Approximate Value Iteration Schemes

Despite the widespread use, the impact of these core techniques on the convergence of RL algorithms is not yet fully understood.