C Gagne, P Dayan - Advances in Neural Information …, 2021 - proceedings.neurips.cc
Distributional reinforcement learning (RL)–in which agents learn about all the possible long-
term consequences of their actions, and not just the expected value–is of great recent …