AI Paper Atlas

Research map / Adversarial robustness and security / Black-box and transfer attacks

Class-feature Watermark: A Resilient Black-box Watermark Against Model Extraction Attacks

Yaxin Xiao, Qingqing Ye, Zi Liang, Haoyang Li, RongHua Li, Huadi Zheng, Haibo Hu

AAAI 2026 · Adversarial robustness and security › Black-box and transfer attacks · approach: Data-free substitute training

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 WRK (boundary reshaping + feature shifting) to remove watermarksentangled black-box watermarks · CIFAR-10 ResNet18watermark success rate (%): EWE 99.88 -> 8.75; Blend 99.95 -> 9.80Improved
apply existing removal methods against entangled watermarksNC, I-BAU, CLP, FP, NAD, SEAM, FST · CIFAR-10 ResNet18watermark success rate (%)No clear effect
use class-level artifact (OOD synthetic class) instead of sample-level artifactsCFW · CIFAR-10 to Speech Commands, copy model after MEA+WRKWSR_LC (%): WSR_LC >= 70.15 under MEA and WRKImproved
add RepS loss (ablation)CFW · CIFAR-10 copy modelWSR_LC (%): 70.00 -> 94.00 (w/ RepS only)Improved
use SNNL instead of RepS for entanglement (ablation)CFW · CIFAR-10 victimaccuracy (%): 93.70 -> 90.13Hurt
add CD2 loss (ablation)CFW · CIFAR-10 copy modelWSR_LC (%): 84.00 -> 94.00 (CD2 only vs full)Improved
add CD2 loss (ablation)CFW · CIFAR-10 copy modelintra-class variance (x10^2): 10.30 -> 2.74Improved
add RepS and CD2 together vs CD2 only (ablation)CFW · CIFAR-10 copy modelrepresentation entanglement RE: 0.04 -> 0.75Improved
verify via label clustering instead of WSRCFW · CIFAR-10 under WRKWSR vs clustering: WSR 100% -> 19.60% while representations stay clusteredImproved
apply CFW on ImageNetteCFW · ImageNetteaccuracy: about 2% utility dropHurt

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