AI Paper Atlas

Research map / Adversarial robustness and security / Backdoor and poisoning

Stealthy Backdoor Attack in Federated Learning via Adaptive Layer-wise Gradient Alignment

Qingqian Yang, Peishen Yan, Xiaoyu Wu, Jiaru Zhang, Tao Song, Yang Hua, Hao Wang, Liangliang Wang, Haibing Guan

ICCV 2025 · Adversarial robustness and security › Backdoor and poisoning · approach: Backdoors in federated learning · 4 citations (OpenAlex)

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
add layer-wise gradient alignment instead of no alignment (ablation)BadNets-style attack · CIFAR-10 VGG19, FLAMEbackdoor accuracy (%): 7.05 -> 97.46Improved
add layer-wise gradient alignment instead of no alignment (ablation)BadNets-style attack · CIFAR-100 ResNet18, Multi-Metricsbackdoor accuracy (%): 0.37 -> 97.05Improved
use layer-wise instead of model-wise alignment (ablation)LGA · CIFAR-10 VGG19, FLAMEbackdoor accuracy (%): 45.79 -> 97.46Improved
use layer-wise instead of model-wise alignment (ablation)LGA · CIFAR-100 ResNet18, no defensebackdoor accuracy (%): 73.29 -> 96.70Improved
use layer-wise instead of model-wise alignment (ablation)LGA · CIFAR-10 ResNet18, FLAMEbackdoor accuracy (%): 85.32 -> 98.16Improved
add layer-wise alignment to alignment-free attack (no defense) (ablation)BadNets · CIFAR-10 VGG19, no defensebackdoor accuracy (%): 97.70 -> 96.16Hurt
add alignment to trigger-optimization attack IBAIBA · FLAMEbackdoor accuracy (%): 2.71 -> 98.04Improved
add alignment to trigger-optimization attack CerpCerp · FLAMEbackdoor accuracy (%): 29.84 -> 99.64Improved
add alignment to A3FLA3FL · Multi-Metricsbackdoor accuracy (%): 86.29 -> 99.89Improved
use LGA instead of LP attackLP attack · CIFAR100 ResNet18 no defensebest backdoor accuracy (%): 72.24 -> 98.51Improved
use LGA instead of LP attack on malicious acceptanceLP attack · CIFAR10 ResNet18, MultiKrumMAR: 0.72 -> 0.68Hurt
increase attack interval F to 5LGA vs LP attack · CIFAR-100backdoor accuracyImproved

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