Emphatic Algorithms for Deep Reinforcement Learning

               Ray Jiang 1 Tom Zahavy 1 Zhongwen Xu 1 Adam White 1 2
                Matteo Hessel 1 Charles Blundell 1 Hado van Hasselt 1

              Abstract                Many reinforcement learning (RL) agents learn off-policy
                                  to some extent, to learn about the greedy policy while ex-
   Off-policy learning allows us to learn about pos-       ploring (Watkins, 1989), to make predictions about policies
...