SUTD-PRCM Dataset and Neural Architecture Search Approach for Complex Metasurface Design

TL;DR

This paper introduces the SUTD-PRCM dataset of approximately 260,000 complex metasurface samples for benchmarking deep learning models and explores neural architecture search to improve metasurface electromagnetic response prediction.

Contribution

It provides a large, complex metasurface dataset for benchmarking and investigates neural architecture modifications to enhance deep learning model performance.

Findings

Convolution stacking is not the dominant element in neural architectures for this task.

Low-level features are preferred over high-level features in metasurface response prediction.

Deep CNN models perform poorly compared to other architectures on this dataset.

Abstract

Metasurfaces have received a lot of attentions recently due to their versatile capability in manipulating electromagnetic wave. Advanced designs to satisfy multiple objectives with non-linear constraints have motivated researchers in using machine learning (ML) techniques like deep learning (DL) for accelerated design of metasurfaces. For metasurfaces, it is difficult to make quantitative comparisons between different ML models without having a common and yet complex dataset used in many disciplines like image classification. Many studies were directed to a relatively constrained datasets that are limited to specified patterns or shapes in metasurfaces. In this paper, we present our SUTD polarized reflection of complex metasurfaces (SUTD-PRCM) dataset, which contains approximately 260,000 samples of complex metasurfaces created from electromagnetic simulation, and it has been used to benchmark our DL models. The metasurface patterns are divided into different classes to facilitate different degree of complexity, which involves identifying and exploiting the relationship between the patterns and the electromagnetic responses that can be compared in using different DL models. With the release of this SUTD-PRCM dataset, we hope that it will be useful for benchmarking existing or future DL models developed in the ML community. We also propose a classification problem that is less encountered and apply neural architecture search to have a preliminary understanding of potential modification to the neural architecture that will improve the prediction by DL models. Our finding shows that convolution stacking is not the dominant element of the neural architecture anymore, which implies that low-level features are preferred over the traditional deep hierarchical high-level features thus explains why deep convolutional neural network based models are not performing well in our dataset.