Mathematical model of a neuromorphic network based on memristive elements

The article discusses the modeling of interconnected memory elements within a neuromorphic network. A mathematical model is proposed that describes a hardware analog implementation of a spiking neural network with memristive elements as synaptic weights and a learning mechanism based on the STDP (Spike-Timing Dependent Plasticity) rule. As a memristor model, a model describing a hafnium oxide (HfO2) element is used. A numerical simulation of the operation of one and two interconnected neurons with 64 and 128 synapses has been carried out. The process of training the network to recognize certain templates is demonstrated. (c) 2020 Elsevier Ltd. All rights reserved.

Automated summary

The article discusses the modeling of interconnected memory elements within a neuromorphic network using a mathematical model and numerical simulation to demonstrate the training process of the network.