Simulator-based training of generative neural networks for the inverse design of metasurfaces

Abstract Metasurfaces are subwavelength-structured artificial media that can shape and localize electromagnetic waves in unique ways. The inverse design of these devices is a non-convex optimization problem in a high dimensional space, making global optimization a major challenge. We present a new type of population-based global optimization algorithm for metasurfaces that is enabled by the training of a generative neural network. The loss function used for backpropagation depends on the generated pattern layouts, their efficiencies, and efficiency gradients, which are calculated by the adjoint variables method using forward and adjoint electromagnetic simulations. We observe that the distribution of devices generated by the network continuously shifts towards high performance design space regions over the course of optimization. Upon training completion, the best generated devices have efficiencies comparable to or exceeding the best devices designed using standard topology optimization. Our proposed global optimization algorithm can generally apply to other gradient-based optimization problems in optics, mechanics, and electronics.

[1]  Jelena Vucković,et al.  Inverse design in nanophotonics , 2018, Nature Photonics.

[2]  Jaakko Lehtinen,et al.  Progressive Growing of GANs for Improved Quality, Stability, and Variation , 2017, ICLR.

[3]  Shanhui Fan,et al.  Adjoint Method and Inverse Design for Nonlinear Nanophotonic Devices , 2018, ACS Photonics.

[4]  Jason Hickey,et al.  Data-driven metasurface discovery , 2018, ACS nano.

[5]  Yongmin Liu,et al.  Deep-Learning-Enabled On-Demand Design of Chiral Metamaterials. , 2018, ACS nano.

[6]  Federico Capasso,et al.  Topology-Optimized Multilayered Metaoptics , 2017, 1706.06715.

[7]  Ole Sigmund,et al.  On the Design of Compliant Mechanisms Using Topology Optimization , 1997 .

[8]  David Sell,et al.  Large-Angle, Multifunctional Metagratings Based on Freeform Multimode Geometries. , 2017, Nano letters.

[9]  W. Cai,et al.  A Generative Model for Inverse Design of Metamaterials , 2018, Nano letters.

[10]  Paulo Henrique da F Silva,et al.  Blending PSO and ANN for Optimal Design of FSS Filters With Koch Island Patch Elements , 2010, IEEE Transactions on Magnetics.

[11]  David Sell,et al.  Periodic Dielectric Metasurfaces with High‐Efficiency, Multiwavelength Functionalities , 2017 .

[12]  Yi Yang,et al.  Nanophotonic particle simulation and inverse design using artificial neural networks , 2018, Science Advances.

[13]  Jianji Yang,et al.  Analysis of material selection on dielectric metasurface performance. , 2017, Optics express.

[14]  Martin A. Green,et al.  Self-consistent optical parameters of intrinsic silicon at 300 K including temperature coefficients , 2008 .

[15]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[16]  Jiajun Wu,et al.  Visual Object Networks: Image Generation with Disentangled 3D Representations , 2018, NeurIPS.

[17]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[18]  Jianji Yang,et al.  High-efficiency, large-area, topology-optimized metasurfaces , 2019, Light: Science & Applications.

[19]  Philippe Lalanne,et al.  RETICOLO software for grating analysis , 2021 .

[20]  Jakob S. Jensen,et al.  Systematic design of photonic crystal structures using topology optimization: Low-loss waveguide bends , 2004 .

[21]  Jeff Donahue,et al.  Large Scale GAN Training for High Fidelity Natural Image Synthesis , 2018, ICLR.

[22]  P. Genevet,et al.  Recent advances in planar optics: from plasmonic to dielectric metasurfaces , 2017 .

[23]  B. Jalali,et al.  Silicon Photonics , 2006, Journal of Lightwave Technology.

[24]  Alexander Y. Piggott,et al.  Inverse design and demonstration of a compact and broadband on-chip wavelength demultiplexer , 2015, Nature Photonics.

[25]  Jonathan A. Fan,et al.  Global optimization of dielectric metasurfaces using a physics-driven neural network , 2019, Nano letters.

[26]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[27]  Jianji Yang,et al.  Topology-optimized metasurfaces: impact of initial geometric layout. , 2017, Optics letters.

[28]  Soumith Chintala,et al.  Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks , 2015, ICLR.

[29]  Junsuk Rho,et al.  Designing nanophotonic structures using conditional deep convolutional generative adversarial networks , 2019, Nanophotonics.

[30]  D. R. Luna,et al.  Using natural optimization algorithms and artificial neural networks in the design of effective permittivity of metamaterials , 2013, 2013 SBMO/IEEE MTT-S International Microwave & Optoelectronics Conference (IMOC).

[31]  Jonathan A. Fan,et al.  Review of numerical optimization techniques for meta-device design [Invited] , 2019, Optical Materials Express.

[32]  Ole Sigmund,et al.  Topology optimization for nano‐photonics , 2011 .

[33]  David Sell,et al.  Freeform Metagratings Based on Complex Light Scattering Dynamics for Extreme, High Efficiency Beam Steering , 2017, 1709.05019.

[34]  O. Sigmund,et al.  Topology optimization approaches , 2013, Structural and Multidisciplinary Optimization.

[35]  Zongfu Yu,et al.  Training Deep Neural Networks for the Inverse Design of Nanophotonic Structures , 2017, 2019 Conference on Lasers and Electro-Optics (CLEO).

[36]  拓海 杉山,et al.  “Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks”の学習報告 , 2017 .

[37]  Qi-Jun Zhang,et al.  Artificial neural networks for RF and microwave design - from theory to practice , 2003 .

[38]  Eli Yablonovitch,et al.  Adjoint shape optimization applied to electromagnetic design. , 2013, Optics express.

[39]  Hugo E. Hernandez-Figueroa,et al.  Prediction of Dispersion Relation and PBGs in 2-D PCs by Using Artificial Neural Networks , 2012, IEEE Photonics Technology Letters.

[40]  Jacob Scheuer,et al.  Genetically optimized all-dielectric metasurfaces. , 2017, Optics express.

[41]  David Sell,et al.  Robust design of topology-optimized metasurfaces , 2019, Optical Materials Express.

[42]  David Sell,et al.  Ultra-High-Efficiency Anomalous Refraction with Dielectric Metasurfaces , 2018 .

[43]  Sandeep Inampudi,et al.  Neural network based design of metagratings , 2018, Applied Physics Letters.

[44]  Christian Ledig,et al.  Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).