Superconducting implementation of neural networks using fluxon pulses

The authors fabricated neural-based superconducting integrated circuits by using Nb/AlO/sub x//Nb Josephson junctions, and demonstrated the operation of a 2-b neural-based A/D converter. Fluxon pulses were used as neural impulses, and a Josephson junction was used as a threshold element. The conductance values of resistors by which Josephson transmission lines are connected represent fixed synaptic strengths. The preliminary experimental result suggests that variable critical currents of DC-SQUIDs (superconducting quantum interference devices) may provide synapses with variable strength.<
>