Text this: Transfer in Reinforcement Learning Domains