For an example of a faulty reward function, refer to the following link: https://blog.openai.com/faulty-reward-functions/. For more information about deep RL, refer to the following link: http://karpathy.github.io/2016/05/31/rl/.