Generalized Model of the Superconducting Sigma Neuron

Abstract

The superconducting sigma neuron is a single-junction interferometer, with a part of the circuit shunted by an additional inductance, which is also used to generate the output signal. It was previously predicted that the transfer function of this device would be close to the sigmoidal one at a certain relation between the values of inductance of its parts. This interferometer can be fabricated as a multilayer thin-film structure over a superconducting screen, which allows measuring the output magnetic flux in a single element. An analysis of the experimental data showed that the use of a superconducting screen did not ensure complete independence of the sigma neuron elements, as was assumed in the theoretical model. This paper presents a generalized model of the stationary state of a sigma neuron that takes into account the interaction between all its parts, including the input and readout elements.