Accelerating Reinforcement Learning through GPU Atari Emulation Steven Dalton