Episode 25: Nicklas Hansen, UCSD, on long-horizon planning and why algorithms don't drive research progress
Generally Intelligent
00:00
Is It Possible to Generate a Neural Network Model?
The policy really depends on the model, both in like the world model and also the neural network model. And we don't expect them to generalize super well from even to like to do it, for example. If you have a bunch of policies for at least one specific robot, you could reuse within that robot. But who knows what you can do if you have a population of models. Maybe it's still make something that is sensible.
Transcript
Play full episode
