Safe Policy Search for Lifelong reinforcement learning with sublinear regret

January 19th, 2021