State Space Decomposition and Subgoal Creation for Transfer in Deep Reinforcement Learning | Read Paper on Bytez