Abstract
Optical neural networks are often trained “in-silico” on digital simulators, but physical imperfections that cannot be modelled may lead to a “reality gap” between the simulator and the physical system. In this work we present hybrid training, where the weight matrix is trained by computing neuron values optically using the actual physical network.
© 2022 The Author(s)
PDF Article | Presentation VideoMore Like This
Sri Krishna Vadlamani, Ryan Hamerly, and Dirk Englund
FTh1B.3 Frontiers in Optics (FiO) 2022
Zibo Hu, Russell L.T. Schwartz, Maria Solyanik-Gorgone, and Volker J Sorger
JW2A.18 3D Image Acquisition and Display: Technology, Perception and Applications (3D) 2022
John X. Li and Francis T. S. Yu
MN4 OSA Annual Meeting (FIO) 1990