3 papers
cs.LG2026
DP-SelFT: Differentially Private Selective Fine-Tuning for Large Language Models
Haichao Sha, Zihao Wang, Yuncheng Wu +2
Large language models (LLMs) are commonly adapted to downstream tasks through fine-tuning, but fine-tuning data often contains sensitive information that may be leaked by the resul…
cs.LG2024
Clip Body and Tail Separately: High Probability Guarantees for DPSGD with Heavy Tails
Haichao Sha, Yang Cao, Yong Liu +3
Differentially Private Stochastic Gradient Descent (DPSGD) is widely utilized to preserve training data privacy in deep learning, which first clips the gradients to a predefined no…
cs.CR2023
PCDP-SGD: Improving the Convergence of Differentially Private SGD via Projection in Advance
Haichao Sha, Ruixuan Liu, Yixuan Liu +1
The paradigm of Differentially Private SGD~(DP-SGD) can provide a theoretical guarantee for training data in both centralized and federated settings. However, the utility degradati…