Learning Routines for Effective Off-Policy Reinforcement Learning

                       Edoardo Cetin 1 Oya Celiktutan 1

             Abstract                 engineering and are often quite influential on the perfor-
   The performance of reinforcement learning de-         mance (Mahmood et al., 2018). Algorithms that learn also
   pends upon designing an appropriate action space,       these additional components end-to-end would alleviate
   where the effect of each action is measurable, y ...