Deep vs Shallow Reinforcement Learning for low dimensional continuous control tasks