PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.
Deep learning has shown notable results in electromagnetic metamaterials research. However, one of the major outstanding challenges is the required large dataset needed for the training stage of the neural network, which may take several months to complete. In order to mitigate this data bottleneck issue, we demonstrate a transfer learning approach to deep learning, which takes advantage of the similar underlying physics between related problems. We demonstrate transfer learning on metasurfaces with a reduced dataset and produce similar accurate results. We overview current efforts and give an outlook for the future of deep learning metamaterials.
Willie J. Padilla
"Deep transfer learning for metamaterials research", Proc. SPIE PC12420, Terahertz, RF, Millimeter, and Submillimeter-Wave Technology and Applications XVI, PC1242005 (13 March 2023); https://doi.org/10.1117/12.2664582
ACCESS THE FULL ARTICLE
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.
The alert did not successfully save. Please try again later.
Willie J. Padilla, "Deep transfer learning for metamaterials research," Proc. SPIE PC12420, Terahertz, RF, Millimeter, and Submillimeter-Wave Technology and Applications XVI, PC1242005 (13 March 2023); https://doi.org/10.1117/12.2664582