AI Paper Atlas

Research map / Adversarial robustness and security / Backdoor and poisoning

Reliable Poisoned Sample Detection against Backdoor Attacks Enhanced by Sharpness Aware Minimization

Mingda Zhang, Mingli Zhu, Zihao Zhu, Li Shen, Baoyuan Wu

ICLR 2026 · Adversarial robustness and security › Backdoor and poisoning · approach: Detecting poisoned samples · 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
train detector model with SAMSpectre · CIFAR-10 ResNet18, BadNetsTPR (%): 51.1 -> 88.4Improved
train detector model with SAMBeatrix · CIFAR-10 ResNet18, BlendedTPR (%): 5.0 -> 99.8Improved
train detector model with SAMAC · CIFAR-10 ResNet18, BadNetsFPR / F1: F1 97.1 -> 42.5Hurt
train detector model with SAMSCAn · CIFAR-10 ResNet18, BadNetsTPR (%): 96.0 -> 95.2No clear effect
train detector model with SAMSCAn · Tiny-ImageNet, WaNetTPR (%): 0.0 -> 71.1Improved
add SAM training alone (ablation)SS · CIFAR-10 BlendedTPR (%): 32.9 -> 90.6Improved
add feature scaling alone (ablation)SS · CIFAR-10 BlendedTPR (%): 32.9 -> 60.4Improved
add feature scaling on top of SAM (ablation)SS · CIFAR-10 BlendedTPR (%): 90.6 -> 94.6Improved
add feature scaling on top of SAM (ablation)Beatrix · CIFAR-10 BlendedTPR (%): 79.8 -> 99.8Improved
use SAM at very low poisoning ratio 0.1%SAM-enhanced PSD · CIFAR-10 ResNet18average TPR: TPR below 60%Mixed
vary SAM perturbation budget rho (ablation)SAM-enhanced PSD · CIFAR-10 ResNet18detection performanceNo clear effect
train with SAM instead of vanilla training (ablation)BadNets backdoor model · CIFAR-10 ResNet18silhouette coefficient: 0.19 -> 0.32Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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