Design of learning circuit for single-electron neural networks

In this paper, we proposed a new single-electron (SE) circuit that can learn after being fabricated. We designed it whose result changes with the number of signals arriving at the output. To design it, we designed an SE flip flop circuit and an SE switch circuit to control signal propagation. The circuit can be used as a learning circuit. We performed character recognition as operation test of the proposed circuit. The results show that our SE neural network using a learning circuit can simultaneously learn multiple types of input patterns.