Fetching the paper…

DR3: Value-Based Deep Reinforcement Learning Requires Explicit Regularization · Around