All of sudden, Sergey Levin’s paper Control as Inference makes a lot of sense to me. Reinforcement learning, in its essense, is estimation. The difference is that the true value spans cross time. So the inferencing is done with time.
All of sudden, Sergey Levin’s paper Control as Inference makes a lot of sense to me. Reinforcement learning, in its essense, is estimation. The difference is that the true value spans cross time. So the inferencing is done with time.