Provably Efficient Exploration in Reinforcement Learning: An Optimistic Approach Zhuoran Yang, Princeton University