Provably Feedback-Efficient Reinforcement Learning via Active Reward Learning

Open in new window