DNNsolver: Accurate and Efficient Deep Learning Modeling of Polarization-Sensitive Diffractive Neural Networks
Electromagnetics
DNNsolver proposes a deep learning approach for accurate and efficient modeling of polarization-sensitive diffractive neural networks, addressing limitations of thin element approximation in capturing intra-layer electromagnetic coupling.
See LICENSE file Diffractive Neural NetworksDeep Learning ModelingPolarization-Sensitive Optics