Policy Evaluation Networks

Jean Harb, Tom Schaul, Doina Precup +1

Many reinforcement learning algorithms use value functions to guide the search for better policies. These methods estimate the value of a single policy while generalizing across ma…