Q-Learning for Continuous Actions with Cross-Entropy Guided Policies

Preprint English OPEN
Simmons-Edler, Riley; Eisner, Ben; Mitchell, Eric; Seung, Sebastian; Lee, Daniel;
  • Subject: Computer Science - Artificial Intelligence

Off-Policy reinforcement learning (RL) is an important class of methods for many problem domains, such as robotics, where the cost of collecting data is high and on-policy methods are consequently intractable. Standard methods for applying Q-learning to continuous-value... View more
Share - Bookmark