Speeding Up Reinforcement Learning by Exploiting Causality in Reward Sequences | IEEE Conference Publication | IEEE Xplore