Gradient flow in the kernel learning problem

Yang Li, Feng Ruan

This is a sequel to our paper `On the kernel learning problem'. We identify a canonical choice of Riemannian gradient flow, to find the stationary points in the kernel learning pro…