Intrinsically Motivated Reinforcement Learning: An Evolutionary Perspective