Supervised Training of Siamese Spiking Neural Networks with Earth Mover's Distance

TL;DR

This paper presents a supervised training method for Siamese Spiking Neural Networks using Earth Mover's Distance, demonstrating efficient, low-energy, and low-latency image classification on the MNIST dataset.

Contribution

Introduces a novel supervised training framework for SNNs with EMD, achieving high accuracy with fewer neurons and improved efficiency.

Findings

Achieved F1-score up to 0.9386 on MNIST

Used only 15% of neurons compared to traditional models

Sparse coding reduced model speed by 45%

Abstract

This study adapts the highly-versatile siamese neural network model to the event data domain. We introduce a supervised training framework for optimizing Earth Mover's Distance (EMD) between spike trains with spiking neural networks (SNN). We train this model on images of the MNIST dataset converted into spiking domain with novel conversion schemes. The quality of the siamese embeddings of input images was evaluated by measuring the classifier performance for different dataset coding types. The models achieved performance similar to existing SNN-based approaches (F1-score of up to 0.9386) while using only about 15% of hidden layer neurons to classify each example. Furthermore, models which did not employ a sparse neural code were about 45% slower than their sparse counterparts. These properties make the model suitable for low energy consumption and low prediction latency applications.