PE-EK-PINN: Physics Embedding with Evolving Kernel for Scalable Physics-Informed Neural Networks
My rating
This paper addresses scalable physics-informed neural networks for oscillatory electromagnetic wave problems. It explores architectural physics embedding with an evolving kernel as an alternative to relying only on governing-equation loss residuals, with practical radiation settings involving singular excitations, absorbing boundaries, and wave fields spanning tens of wavelengths.
Tags: