Multiscale Physics‐Informed Neural Networks for the Inverse Design of Hyperuniform Optical Materials
Advanced Optical Materials · 2025-06 · 13 (16)
DOI: 10.1002/adom.202403304 ↗Kurz gefasst
Fachartikel in der Zeitschrift Advanced Optical Materials, online erschienen am 19.05.2025. Verfasst von Roberto Riganti und 4 weiteren Autor:innen. Die Arbeit wurde bisher 14-mal zitiert (Crossref) und nennt 71 Literaturangaben. Gefördert von: Army Research Office, National Science Foundation.
Abstract (Originalsprache)
In this study, multiscale physics‐informed neural networks (MscalePINNs) are employed for the inverse design of finite‐size photonic materials with stealthy hyperuniform (SHU) disordered geometries. Specifically, MscalePINNs are shown to capture the fast spatial variations of complex fields scattered by arrays of dielectric nanocylinders arranged according to isotropic SHU point patterns, thus enabling a systematic methodology to inversely retrieve their effective dielectric profiles. This approach extends the recently developed high‐frequency homogenization theory of hyperuniform media and retrieves more general permittivity profiles for applications‐relevant finite‐size SHU and optical systems, unveiling unique features related to their isotropic nature. In particular, the existence of a transparency region beyond the long‐wavelength approximation is numerically corroborated, enabling the retrieval of effective and isotropic locally homogeneous media even without disorder‐averaging, in contrast to the case of uncorrelated Poisson random patterns. The flexible multiscale network approach introduced here enables the efficient inverse design of more general effective media and finite‐size optical metamaterials with isotropic electromagnetic responses beyond the limitations of traditional homogenization theories.
Förderung
- Army Research Office
- National Science Foundation
Zitieren
APA
Riganti, R., Zhu, Y., Cai, W., Torquato, S., & Negro, L. D. (2025). Multiscale Physics‐Informed Neural Networks for the Inverse Design of Hyperuniform Optical Materials. Advanced Optical Materials, 13(16). https://doi.org/10.1002/adom.202403304
Vancouver
Riganti R, Zhu Y, Cai W, Torquato S, Negro LD. Multiscale Physics‐Informed Neural Networks for the Inverse Design of Hyperuniform Optical Materials. Advanced Optical Materials. 2025;13(16). doi:10.1002/adom.202403304
BibTeX
@article{Riganti2025,
author = {Riganti, Roberto and Zhu, Yilin and Cai, Wei and Torquato, Salvatore and Negro, Luca Dal},
title = {Multiscale Physics‐Informed Neural Networks for the Inverse Design of Hyperuniform Optical Materials},
journal = {Advanced Optical Materials},
year = {2025},
volume = {13},
number = {16},
doi = {10.1002/adom.202403304}
}
EUVIMED