The Robot Brains Podcast cover image

Sergey Levine explains the challenges of real world robotics

The Robot Brains Podcast

CHAPTER

Offline Reinforcement Learning

offline reinforcement learning is different from the more canonical, onland reinforcement learning. The challenge is that if we think about the active process, the one where it's like you train a dog,. If the a i agent has some idea about an action that might be good, but it doesn't know if it's good, what is it going to do? Wel is going o go and try that action, experience the outcome and adjust its understanding accordingly. In off line arl, how do you determine if particular conclusion you draw from the data is actually accurate? Counterfactual quarry: Is this other inference actually accurate or is it simply a delusion caused by incomplete data?"

00:00
Transcript
Play full episode

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
App store bannerPlay store banner