Distributional Reinforcement Learning

Idea from jason https://arxiv.org/abs/2107.06106