AI Paper Atlas

Research map / Adversarial robustness and security / Backdoor and poisoning

Effective Backdoor Defense by Exploiting Sensitivity of Poisoned Samples

Weixin Chen, Baoyuan Wu, Haoqian Wang

NeurIPS 2022 · Adversarial robustness and security › Backdoor and poisoning · approach: Detecting poisoned samples · 11 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 D-ST (FCT-based distinguishment + secure training) instead of DBDsecure training from scratch · CIFAR-10 avgASR (%): 96.40 -> 1.21Improved
use D-BR instead of ANP for backdoor removalbackdoor removal · CIFAR-10 avgACC (%): 84.91 -> 92.42Improved
semi-supervised contrastive learning (SS-CTL) instead of CTL for the feature extractor (ablation)D-ST · CIFAR-10 BadNets-all2oneACC (%): 85.63 -> 92.77Improved
SS-CTL instead of S-CTL for the feature extractor (ablation)D-ST · CIFAR-10 SIGASR (%): 99.77 -> 0.00Improved
SS-CTL instead of S-CTL for the feature extractor (ablation)D-ST · CIFAR-10 ACC across attacksACC (%): 94.51 -> 90.07 (SIG)Hurt
replace FCT metric with Spectral signatures or DBD metric in SD module (ablation)D-BR · CIFAR-10 BadNets, Blend, CLASRHurt
replace FCT metric with ABL loss metric (ablation)D-BR · CIFAR-10 BlendASR (%): 24.72 -> <0.5Hurt
use unlearning+relearning instead of pure unlearning (ablation)BR module · CIFAR-10ACCImproved
use unlearning+relearning instead of pure relearning (ablation)BR module · CIFAR-10ASRImproved
add negative CE term on distinguished poisoned samples (lambda_p > 0) (ablation)mixed cross-entropy loss · CIFAR-10ASRImproved
increase lambda_p from 0.001 to 1 (ablation)mixed cross-entropy loss · CIFAR-10ACCHurt

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