DNNsolver: Accurate and Efficient Deep Learning Modeling of Polarization-Sensitive Diffractive Neural Networks
Abstract
Diffractive neural networks (DNNs), composed of cascaded diffractive layers, offer a promising platform for high-speed optical computing. However, accurate and efficient modeling of high-density, polarization-sensitive DNNs remains challenging. Conventional model based on thin element approximation (TEA) fails to capture electromagnetic coupling within layers, while full-wave simulations such as finite-difference time-domain (FDTD) methods are computationally expensive. Here, we propose DNNsolver, a deep learning-based electromagnetic surrogate model that learns the scattering response of diffractive layers. By predicting the scattering matrix rather than source-dependent output fields, DNNsolver decouples the incident wavefront from the model and could be applied to arbitrary input fields. A local scattering kernel and sliding-window strategy further allow efficient modeling of DNNs with arbitrary spatial sizes. For single diffractive layer under random speckle illumination, DNNsolver achieves an average mean squared error (MSE) of 0.0095 and about 10^6-fold speedup compared with FDTD simulations. DNNsolver is further applied to the optimization of a trilayer polarization-multiplexed image classifier, which exhibits better agreement with FDTD simulations than its TEA-based counterpart. Our work provides an accurate and efficient framework for the design and optimization of high-density, polarization-sensitive DNNs.
Related
- Cascaded Holographic-Diffractive Neural Network with Dual-Polarization Multiplexing for High-Security Optical Encryption
- Accurate Wichmann-Kroll vacuum polarization calculations with Gaussian basis set
- Accurate and efficient calculation of atomic forces in solids with non-self-consistent hybrid functionals
- AES-Debye: an Accurate, Efficient, and Scalable Engine for Debye Scattering Calculations
- Physics-Guided Multi-Channel Restoration for Efficient and Accurate Atomic Force Microscopy Imaging.
BibTeX
@misc{wang2026dnnsolver,
title = {{DNNsolver: Accurate and Efficient Deep Learning Modeling of Polarization-Sensitive Diffractive Neural Networks}},
author = {Shuyi Wang and Hong-Bo Sun and Linhan Lin},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.27323},
eprint = {2609.27323},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.27323},
}
RIS
TY - UNPB
TI - DNNsolver: Accurate and Efficient Deep Learning Modeling of Polarization-Sensitive Diffractive Neural Networks
AU - Shuyi Wang
AU - Hong-Bo Sun
AU - Linhan Lin
PY - 2026
DA - 2026/09/23
PB - arXiv
DO - 10.48550/arxiv.2609.27323
UR - https://arxiv.org/abs/2609.27323
AB - Diffractive neural networks (DNNs), composed of cascaded diffractive layers, offer a promising platform for high-speed optical computing. However, accurate and efficient modeling of high-density, polarization-sensitive DNNs remains challenging. Conventional model based on thin element approximation (TEA) fails to capture electromagnetic coupling within layers, while full-wave simulations such as finite-difference time-domain (FDTD) methods are computationally expensive. Here, we propose DNNsolver, a deep learning-based electromagnetic surrogate model that learns the scattering response of diffractive layers. By predicting the scattering matrix rather than source-dependent output fields, DNNsolver decouples the incident wavefront from the model and could be applied to arbitrary input fields. A local scattering kernel and sliding-window strategy further allow efficient modeling of DNNs with arbitrary spatial sizes. For single diffractive layer under random speckle illumination, DNNsolver achieves an average mean squared error (MSE) of 0.0095 and about 10^6-fold speedup compared with FDTD simulations. DNNsolver is further applied to the optimization of a trilayer polarization-multiplexed image classifier, which exhibits better agreement with FDTD simulations than its TEA-based counterpart. Our work provides an accurate and efficient framework for the design and optimization of high-density, polarization-sensitive DNNs.
ER -