Model of Spiking Neuron-Glial Network with Memristive Plasticity for Image Classification Problem

In this paper, we propose a new algorithm for classifying images from the MNIST database based on a spiking neural network with memristive plasticity and glial regulation of excitatory synapses. This algorithm encodes images through the dynamics of the spiking neural network and generates a new feature space for image classification using classical machine learning methods. The model's analysis revealed that neuron-glial regulation influences synaptic connections, which ultimately alters the performance of the spiking neural network's image classification.