Panda, Soumyashree S.Soumyashree S.PandaTandan, HarshulHarshulTandanHegde, Ravi S.Ravi S.Hegde2025-08-312025-08-312021-01-01[9781665416016]10.1109/IPC48725.2021.95930402-s2.0-85123201620http://repository.iitgn.ac.in/handle/IITG2025/26369Although metasurfaces offer unprecedented design flexibility, the discovery of optimal structures remains a challenging problem where techniques like Deep learning are becoming significant. We propose novel neural architectures beyond the traditionally employed feedforward architectures and explore their utility in the design of multiple-element metagratings.falseDeep learning | Metasurface design | Photonics inverse designMulti-element Metagrating Design With Densely-connected Neural NetworksConference Paper20211cpConference Proceeding1