Switching the Loss Reduces the Cost in Batch Reinforcement Learning

Open in new window