AI Paper Atlas

Research map / Federated learning and privacy / Privacy attacks and secure learning

SkyMask: Attack-agnostic Robust Federated Learning with Fine-grained Learnable Masks

Peishen Yan, Hao Wang, Tao Song, Yang Hua, Ruhui Ma, Ningxin Hu, Mohammad Reza Haghighat, Haibing Guan

ECCV 2024 · Federated learning and privacy › Privacy attacks and secure learning · approach: Defending federated learning against poisoning · 6 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
use SkyMask learnable-mask defense instead of FLTrust under fine-grained attacksFL aggregation · CIFAR-100 (ResNet20), Min-Maxtesting accuracy: 0.30 -> 0.44Improved
use SkyMask under no attack instead of FLTrustFL aggregation · CIFAR-100 (ResNet20), no attacktesting accuracy: 0.39 -> 0.44Improved
drop the trusted root model (SkyMask-NR) when under 50% malicious (ablation)SkyMask · CIFAR-10 (ResNet20), 20% malicioustesting accuracy: 0.77 -> 0.77 (Min-Max)No clear effect
use learnable-mask clustering instead of FLDetector for malicious client detectionmalicious client detection · CIFAR-10 (ResNet20), Min-Maxtesting accuracy: 0.11 -> 0.77Improved
use learnable-mask clustering instead of Tolpegin defensemalicious client detection · CIFAR-10 (ResNet20), Fang-Krumtesting accuracy: 0.17 -> 0.77Improved
use SkyMask at high malicious fractions (above 50%) versus FLTrustFL aggregation · Fashion-MNIST (CNN), 40-80% maliciousconvergence and accuracyImproved
use Krum as defense when there is no attackFL aggregation · CIFAR-10/100, no attacktesting accuracy: 0.77 -> 0.54 (CIFAR-10)Hurt
scale to 200 and 500 clients with SkyMaskSkyMask · CIFAR-10 (ResNet20)FPR and FNR: FPR and FNR remain 0% for fine-grained and targeted attacksImproved

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