MURAL: Meta-Learning Uncertainty-Aware Rewards for Outcome-Driven
            Reinforcement Learning

Kevin Li * 1 Abhishek Gupta * 1 Vitchyr Pong 1 Ashwin Reddy 1 Aurick Zhou 1 Justin Yu 1 Sergey Levine 1

             Abstract
   Exploration in reinforcement learning is, in gen-
   eral, a challenging problem. A common technique
   to make learning easier is providing demonstra-
   tions from a human supervisor, but such demon-
   strations can be expensive and time-consuming to
   acquire. I ...