Info
This is the tutorial I gave at the IAIFI summer school 2022 as a supplement to Yasaman Bahri’s lecture on “Deep learning in the large-width regime”.
I explain the notion of kernel in Neural Tangent Kernel (NTK), which is central to the characterization of neural networks in the infinite-width regime. The video below is part 1 of my tutorial, where I walk through the analytic calculation which reveals the kernel in a feature regression model. I have learned about this approach from Mihai Nica - check out his brilliant videos and lecture notes on his website.