Last summer, I supervised a summer intern, Cédric Odin, student at Ecole Normale in Ker Lann, France, on Dynamic Programming in Distributional Reinforcement Learning. A state-of-the-art is now available online https://hal.archives-ouvertes.fr/hal-03168889
The classic approach to reinforcement learning is limited in that it only predicts the expected return. The specialized literature has long tried to remedy this problem by studying risk-sensitive models, but the distributional approach will not emerge until 2017. Since the seminal article Bellemare, Dabney, and Munos 2017 and the state-of-the-art performance of the C51 algorithm in the ATARI 2600 suite of benchmark tasks (Bellemare, Naddaf, et al. 2013), research has focused on understanding the behaviour of distributional algorithms. In this paper we place Bellemare’s original results in distributional dynamic programming in parallel with the classic results.