Direct Training via Backpropagation for Ultra-low Latency Spiking Neural Networks with Multi-threshold

TL;DR

This paper introduces a novel backpropagation-based training method for ultra-low latency spiking neural networks with multi-thresholds, achieving high accuracy with only 1-2 time steps, significantly reducing latency and energy consumption.

Contribution

The paper proposes a new training approach for ultra-low latency SNNs using multi-threshold LIF models and approximated derivatives, enabling direct training with minimal time steps.

Findings

Achieves over 99% accuracy on MNIST with 2 time steps.

Outperforms previous SNNs on CIFAR10 with fewer time steps.

Reduces latency and energy consumption in SNN training.

Abstract

Spiking neural networks (SNNs) can utilize spatio-temporal information and have a nature of energy efficiency which is a good alternative to deep neural networks(DNNs). The event-driven information processing makes SNNs can reduce the expensive computation of DNNs and save a lot of energy consumption. However, high training and inference latency is a limitation of the development of deeper SNNs. SNNs usually need tens or even hundreds of time steps during the training and inference process which causes not only the increase of latency but also the waste of energy consumption. To overcome this problem, we proposed a novel training method based on backpropagation (BP) for ultra-low latency(1-2 time steps) SNN with multi-threshold. In order to increase the information capacity of each spike, we introduce the multi-threshold Leaky Integrate and Fired (LIF) model. In our proposed training method, we proposed three approximated derivative for spike activity to solve the problem of the non-differentiable issue which cause difficulties for direct training SNNs based on BP. The experimental results show that our proposed method achieves an average accuracy of 99.56%, 93.08%, and 87.90% on MNIST, FashionMNIST, and CIFAR10, respectively with only 2 time steps. For the CIFAR10 dataset, our proposed method achieve 1.12% accuracy improvement over the previously reported direct trained SNNs with fewer time steps.