Kernelization of Natural Gradient Methods for Physics Informed Neural Networks

Presentation of a kernelization framework for natural gradient methods applied to Physics-Informed Neural Networks (PINNs).

Slides