Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
Q: Fachverband Quantenoptik und Photonik
Q 7: Precision Measurements
Q 7.3: Poster
Dienstag, 21. September 2021, 16:30–18:30, P
Tandem Neural Network for Design of High-Reflectivity Metamirrors — •Liam Shelling Neto1,3, Anastasiia Sorokina1,3, Johannes Dickmann1,3, Jan Meyer1,3, Tim Käseberg2, and Stefanie Kroker2,3 — 1Laboratory for Emerging Nanometrology ( LENA), Technical University of Braunschweig, Langer Kamp 6a/b, 38106 Braunschweig — 2Physikalisch-Technische Bundesanstalt, Bundesallee 100, 38116 Braunschweig — 3Cluster of Excellence QuantumFrontiers
In recent years, sub-wavelength structures that interact with light gained increasing attention thanks their ability to manipulate different aspects of the impinging electromagnetic wave, e.g. phase, amplitude or even polarization. The composition of such artificial structures pave the way for a multitude of applications such as ultrathin metalenses or hologram generation. In order to control the vast design space that unfolds with the desired flexibility of those nanostructures, many approaches have been reported in the past. Deep Learning efficiently tackles the problem of large parameter spaces since that is part of its intrinsic nature. In this Poster, we utilize a Tandem Neural Network to design focusing metamirrors with excellent phase agreement while maximizing reflectivity within a given design space.