Classifying Images with CoLaNET Spiking Neural Network -- the MNIST Example

TL;DR

This paper demonstrates how the CoLaNET spiking neural network architecture can be effectively used for supervised image classification, achieving accuracy comparable to advanced machine learning methods on MNIST.

Contribution

It introduces the application of the CoLaNET SNN architecture to image classification and evaluates its performance on MNIST, showing competitive accuracy without convolutional layers.

Findings

CoLaNET achieves near state-of-the-art accuracy on MNIST.

Image pixel brightness is encoded by spike count during presentation.

The architecture is suitable for supervised learning in image classification.

Abstract

In the present paper, it is shown how the columnar/layered CoLaNET spiking neural network (SNN) architecture can be used in supervised learning image classification tasks. Image pixel brightness is coded by the spike count during image presentation period. Image class label is indicated by activity of special SNN input nodes (one node per class). The CoLaNET classification accuracy is evaluated on the MNIST benchmark. It is demonstrated that CoLaNET is almost as accurate as the most advanced machine learning algorithms (not using convolutional approach).