Increased Reinforcement Learning Performance through Transfer of Representation Learned by State Prediction Model | IEEE Conference Publication | IEEE Xplore