AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

Domain Watermark: Effective and Harmless Dataset Copyright Protection is Closed at Hand

Junfeng Guo, Yiming Li, Lixu Wang, Shu-Tao Xia, Heng Huang, Cong Liu, Bo Li

NeurIPS 2023 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Augmentation and mixture-of-experts generalization · 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 domain watermark instead of backdoor watermarksdataset ownership verification · CIFAR-10harmful degree H: 1.00 -> 0.10 (vs BadNets)Improved
use domain watermark instead of clean-label backdoor watermarksdataset ownership verification · Tiny-ImageNetbenign accuracy (%): 51.79 -> 59.10 (vs UBW-C)Improved
use domain watermark instead of backdoor watermarksdataset ownership verification · Tiny-ImageNetverification success rate (%): 58.06 vs 100 (BadNets)Hurt
increase perturbation budget epsilon (ablation)domain watermark · CIFAR-10 and Tiny-ImageNetverification success rateImproved
increase watermarking rate gamma (ablation)domain watermark · CIFAR-10 and Tiny-ImageNetverification success rateImproved
increase perturbation budget epsilon (ablation)domain watermark · CIFAR-10 and Tiny-ImageNetbenign accuracyNo clear effect
fine-tune the watermarked modeldomain watermark · CIFAR-10VSR and BANo clear effect

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