Using Unity to Help Solve Reinforcement Learning