4 papers
Defending against Backdoor Attacks via Module Switching
Weijun Li, Ansh Arora, Xuanli He +2
Backdoor attacks pose a serious threat to deep neural networks (DNNs), allowing adversaries to implant triggers for hidden behaviors in inference. Defending against such vulnerabil…
Beyond Theoretical Bounds: Empirical Privacy Loss Calibration for Text Rewriting Under Local Differential Privacy
Weijun Li, Arnaud Grivet Sébert, Qiongkai Xu +2
The growing use of large language models has increased interest in sharing textual data in a privacy-preserving manner. One prominent line of work addresses this challenge through…
Cut the Deadwood Out: Backdoor Purification via Guided Module Substitution
Yao Tong, Weijun Li, Xuanli He +2
Model NLP models are commonly trained (or fine-tuned) on datasets from untrusted platforms like HuggingFace, posing significant risks of data poisoning attacks. A practical yet und…
Seeing the Forest through the Trees: Data Leakage from Partial Transformer Gradients
Weijun Li, Qiongkai Xu, Mark Dras
Recent studies have shown that distributed machine learning is vulnerable to gradient inversion attacks, where private training data can be reconstructed by analyzing the gradients…