Presentation
5 March 2021 AI-enabled intelligent architectures for designing versatile broadband metamaterial absorbers
Author Affiliations +
Abstract
We report a framework comprising of a combination of sequence models (e.g. RNN, LSTM, Bi LSTM etc) and deep neural networks (DNNs) to tackle the forward problem of predicting the optical response for a given geometry of a broadband, terahertz metamaterial absorber based on Au split-ring resonators. We obtained a training and validation losses of 0.0062 and 0.0042 respectively. The test dataset for this model yielded a loss of 0.0026. Using our model, we were able to predict the spectral response of similar metamaterial absorber geometries in less than 0.5 seconds.
Conference Presentation
© (2021) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Parama Pal, Prajith Pillai, Deepak Jain, and Beena Rai "AI-enabled intelligent architectures for designing versatile broadband metamaterial absorbers", Proc. SPIE 11695, High Contrast Metastructures X, 1169512 (5 March 2021); https://doi.org/10.1117/12.2578201
Advertisement
Advertisement
KEYWORDS
Metamaterials

Inverse optics

Inverse problems

Nanophotonics

Neural networks

Optical design

Optical metamaterials

Back to Top