In this presentation, I will describe how a relatively simple electronic device---a 100-nm-width superconducting nanowire patterned in a thin film---can be used for deep learning and potentially other applications in neuromorphic computing. The devices do not rely on Josephson junctions to operate, rather they resemble a cryotron, hearkening back to the early days of superconductivity research. These devices also find use in applications ranging from single-photon detection to memory for superconducting electronics
Nanostructures Laboratory, Research Laboratory of Electronics
Massachusetts Institute of Technology