4 papers
Cert-SSBD: Certified Backdoor Defense with Sample-Specific Smoothing Noises
Ting Qiao, Yingjia Wang, Xing Liu +3
Deep neural networks (DNNs) are vulnerable to backdoor attacks, where an attacker manipulates a small portion of the training data to implant hidden backdoors into the model. The c…
DSSmoothing: Toward Certified Dataset Ownership Verification for Pre-trained Language Models via Dual-Space Smoothing
Ting Qiao, Xing Liu, Wenke Huang +3
Large web-scale datasets have driven the rapid advancement of pre-trained language models (PLMs), but unauthorized data usage has raised serious copyright concerns. Existing datase…
SSCL-BW: Sample-Specific Clean-Label Backdoor Watermarking for Dataset Ownership Verification
Yingjia Wang, Ting Qiao, Xing Liu +3
The rapid advancement of deep neural networks (DNNs) heavily relies on large-scale, high-quality datasets. However, unauthorized commercial use of these datasets severely violates…
CertDW: Towards Certified Dataset Ownership Verification via Conformal Prediction
Ting Qiao, Yiming Li, Jianbin Li +5
Deep neural networks (DNNs) rely heavily on high-quality open-source datasets (e.g., ImageNet) for their success, making dataset ownership verification (DOV) crucial for protecting…