AAAI 2026technical0 citations

RPGen: Robust and Differentially Private Synthetic Image Generation

Zihao Wang, Hao Peng, Wei Dong, Yuecen Wei, Li Sun, Zhengtao Yu

Abstract

Differentially private (DP) image synthesis enables the generation of realistic images while bounding privacy leakage, facilitating secure data sharing across organizations. However, the Gaussian noise injected during DP training, such as via DP-SGD, often severely degrades synthesis quality by disrupting model convergence. To address this, we introduce RPGen, a novel framework that enhances diffusion models

BibTeX
@inproceedings{aaai2026_rpgenrobustanddi,
  title = {RPGen: Robust and Differentially Private Synthetic Image Generation},
  author = {Zihao Wang and Hao Peng and Wei Dong and Yuecen Wei and Li Sun and Zhengtao Yu},
  booktitle = {AAAI 2026},
  year = {2026}
}
RPGen: Robust and Differentially Private Synthetic Image Generation · AAAI 2026