Newest papers

Diffusion Sensitivity Geometry and Strong Duality for Optimal Privacy-Utility Tradeoffs

Tue Do, Daniel Alabi · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.28991 · cited by 0 works (OpenAlex) · open-access PDF

French

Les approches existantes à la confidentialité différentielle dans les modèles de diffusion reposent sur des limites de sensibilité lâches et du bruit isotrope de Gauss, conduisant à une erreur d'estimation inutile. Nous développons un cadre pour le trade-off de confidentialité et d'utilité optimal des dénonciateurs de diffusion idéaux, en remplaçant les limites conservatrices par des caractéristiques de sensibilité exactes et une calibration du bruit probable optimale. Premièrement, nous montrons que la sensibilité du pire sur les ensembles de données de taille arbitraire peut être caractérisée avec précision en utilisant seulement deux enregistrements. Pour un domaine de données de taille $m$, cela réduit l'analyse à des contraintes maximales de $m(

Machine translation (Meta M2M100); the English original follows.

Abstract

Existing approaches to differential privacy in diffusion models rely on loose sensitivity bounds and isotropic Gaussian noise, leading to unnecessary estimation error. We develop a framework for the optimal privacy-utility tradeoff of ideal diffusion denoisers, replacing conservative bounds with exact sensitivity characterizations and provably optimal noise calibration. First, we show that worst-case sensitivity over datasets of arbitrary size can be characterized exactly using only two records. For a data domain of size $m$, this reduces the analysis to at most $m(m-1)$ constraints, regardless of dataset size. Our characterization can be arbitrarily tighter than conventional diameter-based sensitivity bounds, revealing potentially unbounded improvements in noise calibration. Second, we obtain exact privacy guarantees for Rényi, approximate, and Gaussian differential privacy, as well as mutual-information privacy for binary records, avoiding the looseness of previous generic privacy conversions. Finally, we show that the minimum achievable weighted mean-squared error can be computed through a semidefinite program and prove strong duality, providing certificates of global optimality.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Sensitivity (control systems), Differential privacy, Duality (order theory), Gaussian, Mathematics, Noise (video)

Is this paper real? This page was built from OpenAlex and arXiv records, not generated. Software can verify any reference in one call: GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.28991, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txt

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange