Prefer a chat interface with context about you and your work?
Smoothed functional-based gradient algorithms for off-policy reinforcement learning: A non-asymptotic viewpoint