The article describes a novel fluxonic processor that integrates photonic synapse events. The processor consists of a transmitter, neuron, dendrite, and synapse, all connected in an electrical circuit.
The transmitter receives a neuron signal and produces output photons. The neuron receives a dendrite signal, processes it, and communicates the processed signal back to the transmitter. The dendrite receives a synapse signal, processes it, and communicates the processed signal to the neuron. The synapse receives an input photon, processes it, and communicates the processed signal to the dendrite.
The processor operates by converting electrical signals into fluxons (quantized magnetic fluxes) using Josephson isolators. The fluxons are then processed and integrated in the respective loops (receiver, integration) of each component (synaptic, dendritic, neuronal). The processed signals are converted back to electrical signals to be communicated between components.
The process enables fluxonic processing of photonic synapse events, allowing for the integration of photonic signals with traditional electronic processing. This could potentially lead to more efficient and compact neuromorphic computing systems that combine the best features of both photonic and electronic processing.
Source: https://www.nist.gov/patents/fluxonic-processor-and-processing-photonic-synapse-events
Keywords: photonic, Josephson, fluxonic
