Reinforcement learning (RL) in real world applications faces major hurdles - the foremost being safety of the physical system controlled by the learning agent and the varying environment conditions in which the autonomous ...