Accelerating Reinforcement Learning through Implicit Imitation