Designing thermal radiation metamaterials via hybrid adversarial autoencoder and Bayesian optimization

TL;DR

This paper introduces a hybrid AI approach combining adversarial autoencoders and Bayesian optimization to efficiently design thermal radiation metamaterials with targeted properties, significantly reducing computational costs.

Contribution

It presents a novel hybrid materials informatics framework that enables rapid and cost-effective design of thermal metamaterials with high-dimensional features.

Findings

Design of narrowband thermal emitters at various wavelengths achieved

Reduced computational effort by calculating less than 0.001% of candidate structures

Framework adaptable to other thermal radiation metamaterials

Abstract

Designing thermal radiation metamaterials is challenging especially for problems with high degrees of freedom and complex objective. In this letter, we have developed a hybrid materials informatics approach which combines the adversarial autoencoder and Bayesian optimization to design narrowband thermal emitters at different target wavelengths. With only several hundreds of training data sets, new structures with optimal properties can be quickly figured out in a compressed 2-dimensional latent space. This enables the optimal design by calculating far less than 0.001\% of the total candidate structures, which greatly decreases the design period and cost. The proposed design framework can be easily extended to other thermal radiation metamaterials design with higher dimensional features.