Reinforcement-learning algorithms 1,2 are inspired by our understanding of decision making in humans and other animals in which learning is supervised through the use of reward signals in response to ...