Reinforcement Learning Deep Dive