ICENODE · Integrated cryogenic optoelectronic-superconducting nanowire nodes

Lines connect the coordinator with its partners.

Project objective

Global energy demand from artificial intelligence infrastructure is rising rapidly. This is not sustainable. Neuromorphic hardware offers critical solutions to this problem by implementing brain-inspired, energy-efficient information processing. This project will establish a proof-of-concept neuromorphic device based on on-chip photonic interactions with optically sensitive junctions defined in superconductor nanomaterials. Nanoscale, high-aspect-ratio crystals, will be assembled into spiking neurons in a flexible, axon-like geometry suitable for neural networks. Here one nanocrystal generates on-chip light spikes that lowers the threshold of a superconducting weak link in another nanocrystal, switching it to its resistive state producing a voltage spike. By combining unmatched speed-to-power ratios of superconducting quantum devices with the brain-like parallelism of on-chip nanophotonic interactions, this project addresses critical bottlenecks of cryogenic neuromorphics. Core actions of the project are: (i) set up a deterministic nanocrystal deposition platform for nodal engineering, (ii) discern photon-driven from thermal switching across key material combinations to determine the mechanism and extract heat-management design rules and (iii) demonstrate a proof-of-concept spiking node showing key biological functionality crucial for neural networks. This project brings together the host’s strengths in nanophotonics and my background in quantum devices to deliver a competitive neuromorphic building block. This fellowship will extend my expertise from quantum devices into nanophotonics, deepen my network within the neuromorphic community, and lay the groundwork for an independent research group in cryogenic neuromorphics.

Original text from CORDIS.

Participants

• LUNDS UNIVERSITET · LundCoordinatorSweden

Links

• View on CORDIS
• DOI: 10.3030/101269798