Value Function in Frequency Domain and the Characteristic Value Iteration Algorithm

This paper considers the problem of estimating the distribution of returns in reinforcement learning, i.e., distributional RL problem.