Inter Spike Interval and Stochasticity Engineering of Floating Gate Technology-based Neurons for Spiking Neural Network Hardware

An analytical model has been developed for a stochastic spiking neuron with the floating gate (FG) technology. The electron injection statistics have been applied to the leaky-integrate-and-fire (LIF) neuron model for the first time and inter spike interval (ISI) stochasticity has been quantified. The model has demonstrated low voltage operations below 2V and the ability to fine-tune ISI mean values by 6 orders of magnitude and the degree of ISI stochasticity ranging from 0% to 60%. With the strong interests in the stochastic neurons in spiking neural network (SNN), the ISI stochasticity and tunability of the FG-based neuron have potential to realize efficient and robust spiking neural network hardware.