
01:07:39
In the norm definition on the LHS, is the kernel implicit in the definition of a(\eta)?

01:07:59
My mic is broken at the moment actually.

01:31:29
Thanks for the interesting talk! In which paper can we find the discussion of the minimum RKHS norm solution?

01:32:55
I think blog is this https://francisbach.com/gradient-descent-neural-networks-global-convergence/

01:45:43
Hi and thanks for the talk, is there any insight except number of neurons on why neural nets where able to identify the true features in the noisy experiment setting? Any ideas if that would work for spurious features?

01:48:02
Thanks!

01:48:08
thanks everyone