Is Exploration or Optimization the Problem for Deep Reinforcement Learning?

Glen Berseth

In the era of deep reinforcement learning, making progress is more complex, as the collected experience must be compressed into a deep model for future exploitation and sampling. M…