The Value Function Polytope in Reinforcement Learning
https://arxiv.org/abs/1901.11524v3
Marc G. Bellemare
,
Google