AI Paper Atlas

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

Poisoning with a Pill: Circumventing Detection in Federated Learning

Hanxi Guo, Hao Wang, Tao Song, Tianhang Zheng, Yang Hua, Haibing Guan, Xiangyu Zhang

AAAI 2026 · Federated learning and privacy › Privacy attacks and secure learning · approach: Defending federated learning against poisoning · 1 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 poison pill augmentation to existing poisoning attackssign-flip, Trim, Krum, Min-Max attacks · Fashion-MNIST IID, 20% malicious, FedAvg and 5 defenseserror rate: avg +0.274, max +0.658Improved
add poison pill augmentation to existing poisoning attackssign-flip, Trim, Krum, Min-Max attacks · Fashion-MNIST non-IID, 20% maliciouserror rate: avg +0.281, max +0.637Improved
add poison pill to sign-flipping attack against FedAvgsign-flipping attack · Fashion-MNIST IID, FedAvgerror rate: 0.943 -> 0.667Hurt
add poison pill to Trim attack against FLTrustTrim attack · Fashion-MNIST IID, FLTrusterror rate: 0.109 -> 0.576Improved
add poison pill to Krum attack against Multi-KrumKrum attack · Fashion-MNIST IID, MKrumerror rate: 0.189 -> 0.715Improved
add poison pill to Krum attack against FLDetectorKrum attack · Fashion-MNIST IID, FLDetectorerror rate: 0.786 -> 0.637Hurt
reduce malicious client proportion from 20% to 10% (ablation)poison pill augmentation · Fashion-MNIST cross-siloerror rate increase: avg +0.144, max +0.403 (vs avg +0.274 at 20%)Hurt
switch to cross-device FL (40% clients per round) (ablation)poison pill augmentation · Fashion-MNIST cross-deviceerror rate increase: avg +0.279, max +0.639No clear effect
apply poison pill on CIFAR-10 with AlexNetall four baseline attacks · CIFAR-10, 20% malicious, 30 clientserror rate: avg +0.288Improved
use approximate max pill search instead of approximate min pill search (ablation)pill construction · pill search ablation (extended version)attack error rate: better in 41 out of 56 casesImproved
add extra local training rounds for malicious reference model (ablation)pill poisoning · FLTrust server update, IID and non-IIDcosine similarity to FLTrust updateImproved
add pill adjustment (similarity and distance) (ablation)poison pill injection · Multi-Krum distance scoredistance scoreImproved

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