Designing thermal radiation metamaterials via a hybrid adversarial autoencoder and Bayesian optimization

Abstract

Designing thermal radiation metamaterials is challenging especially for problems with high degrees of freedom and complex objectives. In this Letter, we develop 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 determined in a compressed two-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.