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

We transform reinforcement learning (RL) into a form of supervised learning (SL) by turning traditional RL on its head, calling this Upside Down RL (UDRL). RL predicts rewards, while UDRL instead uses rewards as task-defining, together with representations of time horizons and other computable of historic and desired future data. UDRL learns to interpret these observations as commands, mapping them to actions (or action) through SL on past (possibly accidental) experience. UDRL to achieve high rewards or other goals, through input commands such: get lots of reward within at most so much time! A separate paper 63 on experiments with UDRL shows that even a pilot version of UDRL can traditional baseline algorithms on certain challenging RL problems. also also conceptually simplify an approach 60 for teaching a robot to humans. First videotape humans imitating the robot's current behaviors, let the robot learn through SL to map the videos (as input commands) to behaviors, then let it generalize and imitate videos of humans executing unknown behavior. This Imitate-Imitator concept may actually explain biological evolution has resulted in parents who imitate the babbling of babies.