Abstract:Employing deep learning models to design high-performance metasurfaces has garnered significant attention due to its potential benefits in terms of accuracy and efficiency. A deep learning-based metasurface design framework typically comprises a forward prediction path for predicting optical responses and a backward retrieval path for generating geometrical configurations. In the forward design path, a specific geometrical configuration corresponds to a unique optical response. However, in the inverse design path, a single performance metric can correspond to multiple potential designs. This one-to-many mapping poses a significant challenge for deep learning models and can potentially impede their performance. Although representing the inverse path as a probabilistic distribution is a widely adopted method for tackling this problem, accurately capturing the posterior distribution to encompass all potential solutions remains an ongoing challenge. Furthermore, in most pioneering works, the forward and backward paths are captured using separate models. However, the knowledge acquired from the forward path does not contribute to the training of the backward model. This separation of models adds complexity to the system and can hinder the overall efficiency and effectiveness of the design framework. Here, we utilized an invertible neural network (INN) to simultaneously model both the forward and inverse process. Unlike other frameworks, INN focuses on the forward process and implicitly captures a probabilistic model for the inverse process. Given a specific optical response, the INN enables the recovery of the complete posterior over the parameter space. This capability allows for the generation of novel designs that are not present in the training data. Through the integration of the INN with the angular spectrum method, we have developed an efficient and automated end-to-end metasurface design and evaluation framework. This novel approach eliminates the need for human intervention and significantly speeds up the design process. Utilizing this advanced framework, we have effectively designed high-efficiency metalenses and dual-polarization metasurface holograms. This approach extends beyond dielectric metasurface design, serving as a general method for modeling optical inverse design problems in diverse optical fields.

High-efficiency Transmissive Tunable Metasurfaces for Binary Cascaded Diffractive Layers

An Enhanced Diffractive Neural Network for Metasurface Holograms with High Resolution, Low Noise, and Uniform Intensity

Neural Network-Based Surrogate Model for Inverse Design of Metasurfaces

Metasurface-enabled on-chip multiplexed diffractive neural networks in the visible

All dielectric metasurface based diffractive neural networks for 1-bit adder

Metasurface Enabled Multi‐Target and Multi‐Wavelength Diffraction Neural Networks

Terahertz Metasurface Design with Arbitrary Hexagon Patterns Based on Deep Learning

A programmable diffractive deep neural network based on a digital-coding metasurface array

Transfer-Learning-Assisted Inverse Metasurface Design for 30% Data Savings

Heterogeneous Transfer‐Learning‐Enabled Diverse Metasurface Design

Direct Field-to-pattern Monolithic Design of Holographic Metasurface Via Residual Encoder-Decoder Convolutional Neural Network

Metasurface-Based Image Classification Using Diffractive Deep Neural Network

Spectral Transfer-Learning-based Metasurface Design Assisted by Complex-Valued Deep Neural Network

End-to-End Diverse Metasurface Design and Evaluation Using an Invertible Neural Network

Electromagnetic Manipulation Evolution from Stacked Meta‐Atoms to Spatially Cascaded Metasurfaces

Integrated Photonic Metasystem for Image Classifications at Telecommunication Wavelength.

Direct electromagnetic information processing with planar diffractive neural network

All‐Dielectric Metasurface Empowered Optical‐Electronic Hybrid Neural Networks

Multiplexed supercell metasurface design and optimization with tandem residual networks

Reconfigurable Holography Based on Diffractive Deep Neural Networks

High-Capacity Metasurface at Limits of Polarization and Wavelength Multiplexing