Neuromorphic Single-Electron Circuit and its Application to Temporal-Domain Neural Competition

We propose neuromorphic single-electron circuits for fundamental neural components in modern spiking neural networks, aiming at implementing artificial neural networks on a single or multi-layer nano-dot array. A unit circuit consists of a pair of single-electron oscillators. Using these unit circuits with coupling capacitors, we designed a single-electron neuron circuit that consists of excitable axons and dendrites, excitatory and inhibitory synapses, and a soma. We present an application of the neuron circuit in an inhibitory neural network, where the neurons compete with each other in the temporal domain.