Artificial neural network circuits with Josephson devices

Abstract

A novel approach to Josephson devices for computer applications is described. With an artificial neural network scheme, Josephson devices will be expected to develop a new paradigm for future computer systems. Circuit configurations for a neuron with Josephson devices are described. A combination of a variable bias source and Josephson devices is proposed for a synapse circuit. The bias source signal is steered by the Josephson device input signal and becomes the synapse output signal. These output signals are summed up at the specific resistor or inductor to produce the weighted sum of Josephson devices input signals. According to the error signal, the bias source value is corrected. This corresponds to the learning procedure. Because Josephson devices are threshold logic circuits themselves, they are used as soma circuits. The cell structure of the artificial neural network is discussed.