EUVIMEDEUVIMED entdecken
Crossref

RPGen: Robust and Differentially Private Synthetic Image Generation

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

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (13) · 10449-10457

DOI: 10.1609/aaai.v40i13.38016 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Zihao Wang und 5 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

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' parameter robustness to mitigate DP noise effects without compromising privacy guarantees. At its core, RPGen employs adversarial model perturbation (AMP) during public pre-training to build resilience against perturbations, but we identify and tackle the critical issue of robustness transferability across domains. RPGen achieves this through a three-step process: (1) A pre-trained classifier infers labels for private images, aggregated into a class distribution noised with Gaussian mechanism for DP, and public samples are selected to match this privatized distribution for domain alignment; (2) The diffusion model is pre-trained on this curated subset with adversarial model perturbation to foster robustness; (3) The model undergoes fine-tuning on private data using DP-SGD. This synergy of robustness augmentation and transferability optimization yields high-fidelity synthesis. Extensive evaluations on ImageNet for pre-training, with CelebA and CIFAR-10 for synthesis, show RPGen outperforming state-of-the-art baselines across epsilon in 1, 5, 10. On average, it achieves 20.18% lower FID and 5.45% higher classification accuracy. Ablations confirm the efficacy of domain curation and modest perturbations, establishing RPGen as a new benchmark for privacy-utility trade-offs in image generation.

Zitieren

APA

Wang, Z., Peng, H., Dong, W., Wei, Y., Sun, L., & Yu, Z. (2026). RPGen: Robust and Differentially Private Synthetic Image Generation. Proceedings of the AAAI Conference on Artificial Intelligence, 40(13), 10449-10457. https://doi.org/10.1609/aaai.v40i13.38016

Vancouver

Wang Z, Peng H, Dong W, Wei Y, Sun L, Yu Z. RPGen: Robust and Differentially Private Synthetic Image Generation. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(13):10449-10457. doi:10.1609/aaai.v40i13.38016

BibTeX
@article{Wang2026,
  author = {Wang, Zihao and Peng, Hao and Dong, Wei and Wei, Yuecen and Sun, Li and Yu, Zhengtao},
  title = {RPGen: Robust and Differentially Private Synthetic Image Generation},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {13},
  pages = {10449--10457},
  doi = {10.1609/aaai.v40i13.38016}
}
Kostenlos starten →