aiqbench analysis
This research applies physics-constrained neural networks to accelerate the inverse design of diffractive waveguide combiners, directly addressing efficiency and accuracy in augmented reality optics. By enforcing energy conservation through manifold projection, the method creates a differentiable surrogate that bridges rigorous coupled-wave analysis with gradient-based optimization. This approach represents a significant advancement in machine learning applications for solving complex problems in quantum-optical physics and photonic device engineering.
Physics-constrained neural networks for surrogate modeling of lossless periodic structures
arXiv:2606.28119v2 Announce Type: replace-cross Abstract: We introduce a physics-constrained neural network for the rapid prediction of rigorous coupled-wave analysis outputs in the form of Jones matrices. Starting from energy conservation in lossless layered periodic structures, we use the fact that the scattering outputs lie on a Stiefel manifold. This energy constraint is enforced as a hard condition by projecting onto the manifold using differentiable symmetric orthogonalization. The resulting surrogate enforces energy conservation by construction while preserving differentiability for gradient-based inverse design. The performance and generality of the proposed approach are demonstrated through the inverse design of a diffractive waveguide combiner for augmented reality glasses.
Abstract:We introduce a physics-constrained neural network for the rapid prediction of rigorous coupled-wave analysis outputs in the form of Jones matrices. Starting from energy conservation in lossless layered periodic structures, we use the fact that the scattering outputs lie on a Stiefel manifold. This energy constraint is enforced as a hard condition by projecting onto the manifold using differentiable symmetric orthogonalization. The resulting surrogate enforces energy conservation by construction while preserving differentiability for gradient-based inverse design. The performance and generality of the proposed approach are demonstrated through the inverse design of a diffractive waveguide combiner for augmented reality glasses.
11 pages, 5 figures. Supporting Information Document (PDF) and Video S1 (MP4) are provided as ancillary files
Optics (physics.optics); Machine Learning (cs.LG)
arXiv:2606.28119 [physics.optics]
(or arXiv:2606.28119v2 [physics.optics] for this version)
https://doi.org/10.48550/arXiv.2606.28119 Focus to learn more arXiv-issued DOI via DataCite
Supporting_Information.pdf
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 211 words
Related coverage
Polyconvex neural-network surrogates with learned paired directional features for anisotropic hyperelastic homogenization
arXiv · Oct 12, 2026
QiT: Quantum-Inspired Transformer for Visual Recognition Task
arXiv · Sep 17, 2026
Physics-Constrained Neural Surrogate for Domain Growth Prediction in Systems with Conserved Kinetics
arXiv · Sep 15, 2026
[AINews] High Taste Testers Ep. 1
Latent Space · Oct 17, 2026