Optical Neural Networks

A learned scattering matrix guides the inverse design of optical neural networks. The holographic projection symbolizes the transfer of a trained optical operator into a physical nanostructure, while connected waveguides carry light through the device toward classified class. This separation of task learning from electromagnetic realization enables efficient design of nanophotonic neural networks. More details can be found in the Research Article by Uğur Teğin and co-workers (DOI: 10.1002/nap2.70302).