[10] Chelsea Finn - Learning to Learn with Gradients
The Thesis Review
00:00
The Importance of Meta-Learning in Robotics
In the initial section of my thesis we did experiments on like regression and then these other tasks like omniglot. And I really like doing a mix of work in kind of core machine learning with very simple simpler experiments and also trying to apply that to real robots. So after that I was actually fairly confident that it should work well for some robotics problems, but still an open problem whether this could be applied to them.
Play episode from 23:50
Transcript
