Stanford CS224R Deep Reinforcement Learning | Spring 2025 | Lecture 8: Reward Learning

Transcript

All right, let's get started. Great. So for today, we're planning to finish offline reinforcement learning. And along the reward learning front, the key goals today are to figure out why is task specification hard. Why do naive methods for trying to specify rewards fail? And what are some methods for learning reward functions from human supervision... Read More