arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.04562cs.CR
收藏

从Rényi差分隐私到$f$-差分隐私的最优转换

Optimal conversion from Rényi Differential Privacy to $f$-Differential Privacy

Helmholtz Munich(赫尔霍兹慕尼黑);Technical University of Munich(慕尼黑工业大学);Harvard University(哈佛大学);King’s College London(伦敦国王学院);Hasso-Plattner-Institut(哈索·普拉特纳研究所)

Anneliese Riess, Juan Felipe Gomez, Flavio du Pin Calmon, Julia Anne Schnabel, Georgios Kaissis

AI总结:

本文证明了从Rényi差分隐私(RDP)配置文件到假设检验权衡函数的最优转换规则是基于单阶RDP隐私区域交集的,该规则在所有有效RDP配置文件和所有第一类错误水平下同时达到最优。

摘要:

我们证明了\citet{zhu2022optimalaccountingdifferentialprivacy}附录F.3中提出的猜想:在所有将Rényi差分隐私(RDP)配置文件$τ\mapsto ρ(τ)$映射到有效假设检验权衡$f$的转换规则中,基于单阶RDP隐私区域交集的规则是最优的。该最优性对所有有效RDP配置文件和所有第一类错误水平$α$同时成立。具体地,我们证明在权衡函数空间中,最紧的界是$f_{ρ(\cdot)}(α) = \sup_{τ\geq 0.5} f_{τ,ρ(τ)}(α)$:即每个RDP隐私区域的单阶界的逐点最大值。我们的证明统一并强化了\citet{balle2019hypothesistestinginterpretationsrenyi}、\citet{asoodeh2021variantsdifferentialprivacylossless}和\citet{zhu2022optimalaccountingdifferentialprivacy}的见解。我们的分析依赖于对RDP隐私区域的精确几何刻画,利用其凸性以及其边界完全由伯努利机制决定的事实。我们的结果确立了“RDP隐私区域交集”规则不仅是有效的,而且是最优的:没有其他黑盒转换能在布莱克威尔意义上一致地优于它,这标志着仅从机制的RDP保证所能推断出的隐私信息的根本极限。

英文摘要:

We prove the conjecture stated in Appendix F.3 of \citet{zhu2022optimalaccountingdifferentialprivacy}: among all conversion rules that map a Rényi Differential Privacy (RDP) profile $τ\mapsto ρ(τ)$ to a valid hypothesis-testing trade-off $f$, the rule based on the intersection of single-order RDP privacy regions is optimal. This optimality holds simultaneously for all valid RDP profiles and for all Type I error levels $α$. Concretely, we show that in the space of trade-off functions, the tightest possible bound is $f_{ρ(\cdot)}(α) = \sup_{τ\geq 0.5} f_{τ,ρ(τ)}(α)$: the pointwise maximum of the single-order bounds for each RDP privacy region. Our proof unifies and sharpens the insights of \citet{balle2019hypothesistestinginterpretationsrenyi}, \citet{asoodeh2021variantsdifferentialprivacylossless}, and \citet{zhu2022optimalaccountingdifferentialprivacy}. Our analysis relies on a precise geometric characterization of the RDP privacy region, leveraging its convexity and the fact that its boundary is determined exclusively by Bernoulli mechanisms. Our results establish that the \enquote{intersection-of-RDP-privacy-regions} rule is not only valid, but optimal: no other black-box conversion can uniformly dominate it in the Blackwell sense, marking the fundamental limit of what can be inferred about a mechanism's privacy solely from its RDP guarantees.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑