Expediting Reinforcement Learning by Incorporating Knowledge About Temporal Causality in the Environment