Designing thermal radiation metamaterials via hybrid adversarial autoencoder and Bayesian optimization
This work addresses the problem of efficient metamaterial design for researchers in materials science and engineering, though it appears incremental as it builds on existing methods like adversarial autoencoders and Bayesian optimization.
The authors tackled the challenge of designing thermal radiation metamaterials with high degrees of freedom by developing a hybrid approach combining adversarial autoencoder and Bayesian optimization, resulting in the ability to design narrowband thermal emitters at target wavelengths using only hundreds of training datasets and calculating far less than 0.001% of total candidate structures.
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.