How To Implement Differential Reinforcement