Benchmarking deep learning-based models on nanophotonic inverse design problems

Abstract

Photonic inverse design concerns the problem of finding photonic structures with target optical properties. However, traditional methods based on optimization algorithms are time-consuming and computationally expensive. Recently, deep learning-based approaches have been developed to tackle the problem of inverse design efficiently. Although most of these neural network models have demonstrated high accuracy in different inverse design problems, no previous study has examined the potential effects under given constraints in nanomanufacturing. Additionally, the relative strength of different deep learning-based inverse design approaches has not been fully investigated. Here, we benchmark three commonly used deep learning models in inverse design: Tandem networks, Variational Auto-Encoders, and Generative Adversarial Networks. We provide detailed comparisons in terms of their accuracy, diversity, and robustness. We find that tandem networks and Variational Auto-Encoders give the best accuracy, while Generative Adversarial Networks lead to the most diverse predictions. Our findings could serve as a guideline for researchers to select the model that can best suit their design criteria and fabrication considerations. In addition, our code and data are publicly available, which could be used for future inverse design model development and benchmarking.

Toward High‐Efficiency Ultrahigh Numerical Aperture Freeform Metalens: From Vector Diffraction Theory to Topology Optimization

“…Moreover, as topology optimization is a gradient-based optimization method, the presented freeform metalens are locally optimal rather than globally optimal, which means that the focusing efficiency could be further improved. In fact, deep neural network could serve as a kind of robust global optimizers design by combining topology optimization, [49,50] offering a promising direction for large-scale meta-devices design.…”

Multiscale Optical Field Manipulation via Planar Digital Optics

“…In addition, recent years have seen significant advances in guided-mode conversion applications. One direct method is matching the refractive index difference between different modes by the phase gradient induced equivalent wavevector. , Nevertheless, for complex and multifunctional devices, an inverse design strategy is preferred due to the high freedom-of-degree of subwavelength structures. − …”

Classical and generalized geometric phase in electromagnetic metasurfaces

“…For example, by matching the geometric phases of nano-slots on silver to specific superimpositions of the inward and outward surface plasmon profiles for the two spins, arbitrary spin-dependent orbitals can be generated in a slot-free region [58] . In addition to the traditional design approach, heuristic design methodology has been adopted in nanophotonics to improve the design efficiency and device performances [59] .…”