Skip to main content
QUICK REVIEW

[论文解读] Spiking Layer-Adaptive Magnitude-based Pruning

Junqiao Wang, Zhehang Ye|arXiv (Cornell University)|Mar 16, 2026
Advanced Memory and Neural Computing被引用 0
一句话总结

SLAMP 将面向时间的输出失真建模引入层自适应幅值剪枝,能够在保持准确性与时间保真度的前提下进行更积极的剪枝。

ABSTRACT

Spiking Neural Networks (SNNs) provide energy-efficient computation but their deployment is constrained by dense connectivity and high spiking operation costs. Existing magnitude-based pruning strategies, when naively applied to SNNs, fail to account for temporal accumulation, non-uniform timestep contributions, and membrane stability, often leading to severe performance degradation. This paper proposes Spiking Layer-Adaptive Magnitude-based Pruning (SLAMP), a theory-guided pruning framework that generalizes layer-adaptive magnitude pruning to temporal SNNs by explicitly controlling worst-case output distortion across layers and timesteps. SLAMP formulates sparsity allocation as a temporal distortion-constrained optimization problem, yielding time-aware layer importance scores that reduce to conventional layer-adaptive pruning in single-timestep limit. An efficient two-stage procedure is derived, combining temporal score estimation, global sparsity allocation, and magnitude pruning with retraining for stability recovery. Experiments on CIFAR10, CIFAR100, and the event-based CIFAR10-DVS datasets demonstrate that SLAMP achieves substantial connectivity and spiking operation reductions while preserving accuracy, enabling efficient and deployable SNN inference.

研究动机与目标

  • 通过在不牺牲时间保真度的前提下 reducing connectivity 与 spike operations 实现 SNN 的能效部署
  • 将 Layer-Adaptive Magnitude-based Pruning (LAMP) 扩展到时序 SNN,考虑时间累积效应
  • 提出一个理论框架,在剪枝下对最坏情况的时序输出失真进行界定
  • 开发一个高效的两阶段剪枝与再训练流程,并具备稳定性恢复能力

提出的方法

  • 提出一个在 T 个时间步内聚合贡献的时序失真约束剪枝目标函数(Eq. 3)
  • 定义一个低失真剪枝预算约束以界定累积的输出扰动(Eq. 4)
  • 计算时间感知的层重要性分数 R^i,反映累积的前突触活动(Eq. 3)
  • 在失真约束下推导剪枝掩码 M^i 与被剪枝权重 f6W^i(Eq. 5 与 6)
  • 提供一个迭代的剪枝–再训练循环:训练收敛后按时间分数剪枝,再细调以恢复膜稳定性(Algorithm 1)
  • 证明在单时间步极限下,R^i 收敛为标准 LAMP(Eq. 8)

实验结果

研究问题

  • RQ1如何将剪枝公式化以显式控制 SNNs 的时序失真?
  • RQ2时间感知的层重要性分数是否能够实现全局稀疏、内存高效的 SNN 而不损失时序保真?
  • RQ3时序剪枝对静态与事件驱动数据集的准确性、连通性和尖峰操作次数有何影响?
  • RQ4两阶段剪枝-微调工作流是否能够在强剪枝下产出稳定、可部署的 SNN?

主要发现

数据集剪枝方法体系结构TAcc. Diff(%)Top-1 Acc.(%)Conn.(%)Param. (M)SOPs (M)
CIFAR10ADMM7 Conv, 2 FC8-0.1390.1925.0315.54-
CIFAR10Grad R6 Conv, 2 FC8-0.3092.5436.7210.43-
CIFAR10ESLSNNResNet192-1.7091.0950.006.30180.56
CIFAR10STDS6 Conv, 2 FC8-0.3592.4911.331.71147.22
CIFAR10UPR6 Conv, 2 FC8-0.7992.051.169.5616.47
CIFAR10SLAMP (Ours)ResNet192+1.2392.4240.004.47112.35
CIFAR100ESLSNNResNet192-0.9973.4850.006.32186.25
CIFAR100UPRSEW ResNet184-4.7569.412.48-6.79
CIFAR100SLAMP (Ours)ResNet192+0.7369.8525.004.47115.72
CIFAR10-DVSESLSNNVGGSNN10-0.2878.3010.000.92129.64
CIFAR10-DVSSTDSVGGSNN10-2.6079.804.670.2438.85
CIFAR10-DVSUPRVGGSNN10-0.5078.300.771.816.75
CIFAR10-DVSSLAMP (Ours)VGGSNN5+0.0978.2140.003.60105.45
  • SLAMP 在 CIFAR10、CIFAR100 与 CIFAR10-DVS 上实现了显著的连通性和尖峰操作减少。
  • 在 CIFAR10 上,剪枝到 40% 连通性时,SLAMP 的 Top-1 准确率提升了 +1.23%。
  • 在 CIFAR100 上,在 25% 连通性下达到 69.85% 的 Top-1 时保持精度并实现 0.73% 的增益(在 69.85% Top-1)。
  • 在 CIFAR10-DVS 上,SLAMP 以 40% 连通性保持事件驱动的准确性(+0.09% Top-1)。
  • 剪枝后进行微调显著提升准确性并降低膜方差,尤其是在低连通性时。
  • 该方法在最终轮次将连通性降低到 10% 或更低,同时保持时序保真度。

更好的研究,从现在开始

从论文设计到论文写作,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。