NLP and Reinforcement Learning