Physics-informed neural networks with hard constraints for inverse design

L Lu, R Pestourie, W Yao, Z Wang, F Verdugo… - SIAM Journal on …, 2021 - SIAM
Inverse design arises in a variety of areas in engineering such as acoustic, mechanics,
thermal/electronic transport, electromagnetism, and optics. Topology optimization is an …

Deep convolutional neural networks to predict mutual coupling effects in metasurfaces

S An, B Zheng, MY Shalaginov, H Tang… - Advanced Optical …, 2022 - Wiley Online Library
Metasurfaces have provided a novel and promising platform for realizing compact and high‐
performance optical devices. The conventional metasurface design approach assumes …

Active learning of deep surrogates for PDEs: application to metasurface design

R Pestourie, Y Mroueh, TV Nguyen, P Das… - npj Computational …, 2020 - nature.com
Surrogate models for partial differential equations are widely used in the design of
metamaterials to rapidly evaluate the behavior of composable components. However, the …

Inverse designed extended depth of focus meta-optics for broadband imaging in the visible

E Bayati, R Pestourie, S Colburn, Z Lin, SG Johnson… - …, 2022 - degruyter.com
We report an inverse-designed, high numerical aperture (∼ 0.44), extended depth of focus
(EDOF) meta-optic, which exhibits a lens-like point spread function (PSF). The EDOF meta …

Physics-enhanced deep surrogates for partial differential equations

R Pestourie, Y Mroueh, C Rackauckas, P Das… - Nature Machine …, 2023 - nature.com
Many physics and engineering applications demand partial differential equations (PDE)
property evaluations that are traditionally computed with resource-intensive high-fidelity …

Efficient inverse design of large-area metasurfaces for incoherent light

R Pestourie, W Yao, B Kanté, SG Johnson - Acs Photonics, 2022 - ACS Publications
Incoherent light is ubiquitous, yet designing optical devices that can handle its random
nature is very challenging, since directly averaging over many incoherent incident beams …

Physics-enhanced deep surrogates for PDEs

R Pestourie, Y Mroueh, C Rackauckas, P Das… - 2023 - dspace.mit.edu
Many physics and engineering applications demand Partial Differential Equations (PDE)
property evaluations that are traditionally computed with resource-intensive high-fidelity …

Active learning of deep surrogates for PDEs: application to metasurface design

P Raphaël, M Youssef, TV Nguyen… - NPJ Computational …, 2020 - search.proquest.com
Surrogate models for partial differential equations are widely used in the design of
metamaterials to rapidly evaluate the behavior of composable components. However, the …

[图书][B] Design and Characterization of Optical Metasurface Systems

E Bayati - 2022 - search.proquest.com
The importance of optics is increasing in our daily lives, from medical devices to the cameras
that facilitate global communications. Because of their universal use, the size, scale and …

Data-Efficient Training with Physics-Enhanced Deep Surrogates

R Pestourie, Y Mroueh, CV Rackauckas… - AAAI 2022 Workshop …, 2022 - openreview.net
We present a “physics-enhanced deep-surrogate”(“PEDS”) approach to fast surrogate
models for complex physical systems described by partial differential equations (PDEs) and …