Near-optimal Reinforcement Learning in Factored MDPs | Read Paper on Bytez