Exploring reinforcement learning dynamics in Iterated Prisoner's Dilemma and Chicken using Deep Q-Networks.