This work proposes Burst Spiking Neural Networks (BuSNNs), built upon Burst-enhanced Spiking Neurons and a Dynamic Weight Constraint (DWC) mechanism, which mitigates perturbation-induced transitions in activation states and thereby enhances robustness.
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.
This work proposes a noisy group neuron (NGN) model, which incorporates population-level synchronous resetting and neural stochasticity as fundamental computational mechanisms, and develops the NGN method as a framework that combines the NGN model with backpropagation learning based on mean-field dynamics.
Yajie Zhai, Yanmei Kang, Meng Li et al.· 0 citations
Spiking Neural Networks (SNNs) represent a biologically grounded computational paradigm in which information is carried by discrete spike events, conferring structural compatibility with low-power, event-driven neuromorphic hardware. A persistent impediment to realising the full energy benefit of SNNs in practice is the near-universal adoption of fixed firing thresholds, which prevent individual neurons from self-regulating their activity and permit task-irrelevant spike generation to accumulate unchecked. This paper introduces and evaluates a Homeostatically Regulated Adaptive Threshold (HAT) mechanism in which each neuron’s firing threshold is updated at every time step in proportion to the deviation between its exponentially smoothed firing rate and a designer-specified target rate. The rule is derived from the proportional control framework and mirrors the intrinsic excitability regulation observed in biological cortical circuits. A novel metric, the Target Achievement Error (TAE), is defined to quantify how faithfully the population reaches its intended operating point. Computational cost is assessed using a synaptic operations (SynOps) proxy that is hardware-agnostic and scales linearly with spike volume. A two-stage screening procedure selects the best configurations from 27 candidates by first enforcing an accuracy constraint and then ranking them by spike reduction, TAE, energy, and accuracy. Controlled experiments on MNIST show that the fixed-threshold baseline attains a mean validation accuracy of 92.52%±1.13% across three random seeds, with an average of 21.86 × 106 spikes and an estimated SynOps energy of 211.70 mJ. The top-ranked homeostatic configuration (α = 0.1, ftarget = 0.1, γ = 0.2) achieves 92.02% validation accuracy with a 15.54% reduction in total spike activity, incurring only a 0.50% absolute accuracy shortfall. These results indicate that homeostatic threshold regulation offers a structurally non-invasive route to improved inference energy in gradient-trained SNNs.
Gripsy Paul Mannickathan, Yeldo K. Varghese, Sahala Mariyam P. S. et al.· Journal of Network and Infor...· 0 citations
Spiking neural networks offer a promising route toward low-power sequence computation on neuromorphic hardware, but they continue to lag behind attention-based artificial neural networks on long-context tasks. A central open question is whether this gap reflects only implementation and optimization limitations, or whether architectural features of spiking computation impose unfavorable learnability constraints as sequence length increases. Here, we address this question using a covering-number analysis of feedforward non-leaky integrate-and-fire (nLIF) networks in the probably approximately correct framework. Building on causal-piece decompositions and local Lipschitz continuity, we derive a global sensitivity bound for feedforward nLIF networks and extend it from single-token inputs to multi-token spike sequences. For fixed architectures under stated boundedness and margin assumptions, the resulting sufficient worst-case sample requirement has leading quadratic dependence on sequence length. This dependence arises from cumulative causal participation across time and depth, which increases global sensitivity along active spike paths. We then test the mechanistic implications of this theory using finite-sample cue-recall and teacher–student benchmarks across spiking, recurrent, and attention-based model classes. In cue-recall, an early cue must be retained across distractors and reported at a final query token; in teacher–student, labels are generated by a fixed nLIF teacher, separating representability from finite-sample learnability. Unconstrained feedforward spiking models show sequence-length sensitivity, elevated hidden spike-participation density, and increased samples-to-threshold burden. Post-spike refractoriness, leak-mediated forgetting, learned lateral inhibition, and activity-constrained winner-take-all competition reduce hidden participation and improve empirical robustness in task- and regime-dependent ways. Together, these results identify diffuse causal-set growth as a fundamental architectural bottleneck for baseline feedforward spiking sequence models and suggest that scalable neuromorphic sequence architectures will require circuit mechanisms that explicitly constrain temporal accumulation and effective spike participation.
William Fishell, Gordon Fishell, Suraj Honnuraiah· Neuromorphic Computing and E...· 0 citations
Understanding how biological neural networks are shaped via local plasticity mechanisms can lead to energy-efficient and self-adaptive information processing systems, which promises to mitigate some of the current roadblocks in edge computing systems. While biology makes use of spikes to seamless use both spike timing and mean firing rate to modulate synaptic strength, most models focus on one of the two. In this work, we present a Hebbian local learning rule that models synaptic modification as a function of calcium traces tracking neuronal activity. We show how the rule reproduces results from spike time and spike rate protocols from neuroscientific studies. Moreover, we use the model to train spiking neural networks on MNIST digit recognition to show and explain what sort of mechanisms are needed to learn real-world patterns. We show how our model is sensitive to correlated spiking activity and how this enables it to modulate the learning rate of the network without altering the mean firing rate of the neurons nor the hyparameters of the learning rule. To the best of our knowledge, this is the first work that showcases how spike timing and rate can be complementary in their role of shaping the connectivity of spiking neural networks.
Willian Soares Girāo, Nicoletta Risi, Caroline Geisler et al.· Neuromorphic Computing and E...· 0 citations
It is aimed at proving that SNNs have potential in such areas as computer vision, robotics, and speech recognition, and their role in overcoming the barrier between artificial and biological neural systems is proved.
Mesala Sravani, K. Kumari, S. M. Reddy· International Journal of Unc...· 0 citations
Spiking neural network (SNN) has drawn substantial research focus due to its high biological interpretability, low energy consumption and effectiveness in time-dependent data processing. However, existing methods fail to fully integrate the biological interpretability of SNN with the physical interpretability of time-frequency transform methods in machinery fault diagnosis. To this end, we propose a spiking time-frequency patching (STFP) spiking neural network model, in which a STFP module is designed for time-frequency feature extraction inside the network, and a spiking temporal–spatial attention module is designed to focus on important features of temporal–spatial dimensions. In addition, the spiking residual network used in the model is also improved in neurons. Extensive experimental results on two datasets of rotating machinery parts demonstrate that, the proposed model achieves superior diagnostic performance benefiting from the designed modules, furnishing an end-to-end fault diagnosis method excelling in accuracy, stability, and interpretability.
Shilong Zhu, Jun Wang, Weiguo Huang et al.· IEEE Transactions on Reliabi...· 0 citations