Beyond Target Networks: Improving Deep $Q$-learning with Functional Regularization

Open in new window