Jiang, N. PAC Reinforcement Learning With an Imperfect Model. AAAI 2018, 32.