Forward electromagnetic modeling and inverse design of etch cylindrical holes using physics-informed neural network

This paper discusses the forward modeling and inverse design of the electromagnetic scattering characteristics of two-dimensional etched holes in the nanometre band based on the physics-informed neural network (PINNs). This study focuses on the electromagnetic behavior of randomly etched holes at micro- and nanoscales. It presents a novel approach for accurately calculating and optimizing the electromagnetic scattering characteristics of etched holes. This is achieved by skillfully integrating the simplified Finite-Difference-Frequency-Domain (FDFD) difference matrix and PINNs technology. The findings demonstrate the powerful potential and innovative applications of PINNs in solving complex electromagnetic problems.