Regularizing Reinforcement Learning with State Abstraction | IEEE Conference Publication | IEEE Xplore