Burst Spiking Neural Networks

2026-07-05

AI Summary

This work addresses the vulnerability of spiking neural networks (SNNs) to input perturbations—stemming from binary activation—and the lack of effective weight constraints by proposing Bursting Spiking Neural Networks (BuSNN). The approach introduces Burst-enhanced Spiking Neurons (BSNs) that mitigate abrupt activation shifts caused by minor input disturbances through graded spiking responses. Additionally, a dynamic weight constraint (DWC) mechanism, conditioned on neuronal activation states, is devised to enhance robustness while preserving the inherent energy efficiency of SNNs. Theoretical analysis and experiments on CIFAR-10 and ImageNet demonstrate that BuSNN outperforms existing SNN and ANN baselines on CIFAR-10. On ImageNet, using MS ResNet-34 as the backbone, BuSNN achieves a 3.18% improvement in Top-1 accuracy and a 2.66% gain in corruption robustness, approaching the performance of 8-bit ANNs.

Abstract

A central goal of current Spiking Neural Network (SNN) research is to improve their accuracy toward becoming low-power alternatives to Artificial Neural Networks (ANNs). This work further argues that realizing this ambition requires improving not only accuracy but also robustness, defined as the ability to maintain correct predictions under input perturbations. We identify two key issues in existing SNN methods that undermine robustness. First, binary spiking activations can produce large activation-state changes under small perturbations. Second, the lack of effective weight constraints makes network outputs more sensitive to input variations. To this end, we propose Burst Spiking Neural Networks (BuSNNs), built upon Burst-enhanced Spiking Neurons (BSNs) and a Dynamic Weight Constraint (DWC) mechanism. BSNs incorporate burst firing to provide a graded spiking pattern. This spiking mechanism mitigates perturbation-induced transitions in activation states and thereby enhances robustness. DWC penalizes connection weights based on activation states, effectively reducing weight magnitudes and improving robustness while preserving accuracy. We provide theoretical analyses to support these robustness effects. Experimental results further show that, on smaller-scale benchmarks such as CIFAR-10, BuSNNs outperform both SNN and ANN counterparts in accuracy and robustness. On large-scale ImageNet, BuSNN with the MS ResNet-34 backbone further improves top-1 accuracy and corruption robustness over the corresponding SNN baseline by 3.18% and 2.66%, respectively. Despite using spike-based activations, BuSNNs surpass 4-bit activation-quantized ANN baselines and approach 8-bit ANN baselines on ImageNet. They also preserve SNNs' low-power advantage. This work studies the accuracy-robustness problem in SNNs, advancing their practical viability in robust and energy-efficient applications.

Problem

• Spiking Neural Networks
• robustness
• accuracy
• input perturbations
• burst spiking

Innovation

• Burst Spiking Neural Networks
• Robustness
• Burst Firing
• Dynamic Weight Constraint
• Spiking Neural Networks