AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

On the Role of Temporal Granularity in the Robustness of Spiking Neural Networks

Mengting Xu, Shi Gu, Peng Lin, De Ma, Huajin Tang, Qian Zheng, Gang Pan

CVPR 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: Graph neural network robustness

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
attack with per-time-step gradients (TG-FGSM) instead of time-averaged gradientvanilla VGG11 SNN · CIFAR10, T=4accuracy under FGSM: 12.71 -> 9.72Improved
add TG-Reg to standard trainingVGG11 SNN · CIFAR10, T=4, PGDaccuracy under attack: 0.38 -> 4.25Improved
add TG-Reg to standard trainingVGG11 SNN · CIFAR10, T=4, FGSMaccuracy under attack: 12.71 -> 28.79Improved
add TG-Reg to standard trainingResNet34 SNN · CIFAR10, T=4, cleanclean accuracy: 93.17 -> 87.32Hurt
add TG-Reg to standard trainingVGG11 SNN · CIFAR100, T=8, cleanclean accuracy: 72.76 -> 70.98Hurt
add TG-Reg on top of PGD adversarial trainingVGG11 SNN · CIFAR10, T=4, PGDaccuracy under attack: 22.35 -> 31.12 (vs RAT)Improved
use TG-Reg instead of FEEL-SNNVGG11 SNN · CIFAR100, T=8, PGDaccuracy under attack: 2.04 -> 4.59Improved
add TG-RegVGG11 SNN · CIFAR100, T=8, black-box PGDaccuracy under attack: 5.77 -> 35.91Improved
add TG-Reg on DVS-CIFAR10VGG11 SNN · DVS-CIFAR10, T=10, PGDaccuracy under attack: 11.20 -> 31.70Improved
add TG-Reg to WideResNet16WideResNet16 SNN · CIFAR100, T=8, PGDaccuracy under attack: 0.02 -> 0.33No clear effect
add TG-Reg to Tiny-ImageNet ResNet19ResNet19 SNN · Tiny-ImageNet, T=4, PGDaccuracy under attack: 2.91 -> 3.87Improved
add TG-Reg regularization (training cost)VGG11 SNN · CIFAR100, T=8training time per epoch: 108s -> 236sHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy