AI Paper Atlas

Research map / Adversarial robustness and security / Certified and empirical robustness

Characterizing the Evasion Attackability of Multi-label Classifiers

Zhuo Yang, Yufei Han, Xiangliang Zhang

AAAI 2021 · Adversarial robustness and security › Certified and empirical robustness · approach: Attack evaluation and defense critique · 12 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 GASE label expansion instead of random, oblivious, or loss-guided label searchMulti-label classifiers evaluated with PGD · Creepware, Genbase, VOC2012, Planetaverage number of flipped labelsImproved
Use GASE instead of primitive greedy search (PGS)Multi-label classifiers evaluated with PGD · Creepware, Genbase, VOC2012, Planetaverage number of flipped labelsNo clear effect
Add a low-rank nuclear-norm constraint (ablation)Linear SVM and DNN multi-label classifiers · Creepware, Genbase, VOC2012, Planetattackability indicatorImproved
Add adversarial training without a low-rank constraint (ablation)Linear SVM and DNN multi-label classifiers · Creepware, Genbase, VOC2012, Planetattackability indicatorNo clear effect
Combine adversarial training with a low-rank nuclear-norm constraint (ablation)Linear SVM and DNN multi-label classifiers · Creepware, Genbase, VOC2012, Planetattackability indicatorImproved
Increase the strength of the nuclear-norm constraint (ablation)Linear SVM and DNN multi-label classifiers · Unperturbed test data across four datasetsclassification accuracyHurt

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