/MachineLearning Dec 9
Reinforcement Learning Upside Down: Don't Predict Rewards -- Just Map Them to Actions