Metraipolis › archive › dc_82cf2818c4aa
Adaptive Diffusion Freezing: Privacy-preserving Diffusion Models Against Membership Inference Attacks — arXiv:2609.10608
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.10608
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T05:46:19.963Z
- SHA-256 of the archived text
4e80ee55e7d512dc72300a2280012f378a774c033c03e4da1ec3da937a0f4d96- Stable URL of this page
https://metraipolis.com/documento/dc_82cf2818c4aa
How to cite
Adaptive Diffusion Freezing: Privacy-preserving Diffusion Models Against Membership Inference Attacks — arXiv:2609.10608 — arXiv (en), https://arxiv.org/abs/2609.10608 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_82cf2818c4aa (Metraipolis archive dc_82cf2818c4aa, sha256 4e80ee55e7d512dc72300a2280012f378a774c033c03e4da1ec3da937a0f4d96)
BibTeX
@misc{dc82cf2818c4aa,
title = {Adaptive Diffusion Freezing: Privacy-preserving Diffusion Models Against Membership Inference Attacks — arXiv:2609.10608},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.10608},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_82cf2818c4aa; sha256 4e80ee55e7d512dc72300a2280012f378a774c033c03e4da1ec3da937a0f4d96},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.10608. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Diffusion models have achieved remarkable success in generative tasks across various areas, however their training process raises significant privacy concerns, particularly under membership inference attacks (MIAs). Prior studies on privacy-preserving of diffusion models fail to balance privacy, utility, and efficiency. To address this gap, we propose a novel framework of privacy-preserving diffusion models, Adaptive Diffusion Freezing (ADF), which can defend against MIAs with better trade-off. By leveraging cross-timestep adaptive freezing training, ADF explicitly control the participation of different data subsets across diffusion timesteps via a mask matrix, which reduces the over-memorization and leads to more uniform model behaviors between member and nonmember samples. To construct a freezing mask matrix that effectively reduce membership leakage without unnecessarily harming generation quality, we introduce a pretraining-based risk-aware freezing policy to estimate MIA risk based on memorization tendency, and suppress the contribution of the subset-timestep pairs with higher risk. Evaluations on multiple datasets demonstrate that ADF provides effective defense performance as well as state-of-the-art privacy-utility-efficiency trade-off performance compared to various baselines. Tema degli esploratori: fiducia digitale e crittografia (retro-catalogo arXiv: annunci del 2026-09-11 in cs.CR). Categorie arXiv: Cryptography and Security (cs.CR) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.10608 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-16T05:30:39.364Z dalla pagina https://arxiv.org/catchup/cs.CR/2026-09-11?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).