Episode 22: Archit Sharma, Stanford, on unsupervised and autonomous reinforcement learning
Generally Intelligent
00:00
Is Model Based Learning Compatible With Model Based RL?
The question is like, what model do you learn? So I think that's the question I'm trying to answer. The last thing model based RL is a very promising way here, because first of all, they can transfer to different tasks much better. But it also allows me to hybridize the way that we're what is an opt-in with policy.
Transcript
Play full episode
