arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2203.00263cs.DScs.CRcs.LGmath.OCmath.PR
收藏

基于指数机制的私有凸优化

Private Convex Optimization via Exponential Mechanism

Microsoft Research(微软研究院);University of Washington(华盛顿大学)

Sivakanth Gopi, Yin Tat Lee, Daogao Liu

摘要:

本文研究了非光滑凸函数$F(x)=\mathbb{E}_i f_i(x)$在$\mathbb{R}^d$上的私人优化问题。我们表明,通过在$F(x)$中添加$\ell_2^2$正则化器并从$π(x)\propto \exp(-k(F(x)+μ\|x\|_2^2/2))$中采样来修改指数机制,可以恢复$(ε,δ)$ -DP下的已知最优经验风险和人口损失。此外,我们展示了如何实现这种机制,使用$\widetilde{O}(n \min(d, n))$查询$f_i(x)$的DP-SCO,其中$n$是样本/用户的数量和$d$是环境尺寸。我们还给出了一个(几乎)匹配的下界$\widetildeΩ(n \min(d, n))$的评估查询的数量。我们的结果利用以下工具是独立的兴趣:(1)我们证明了高斯差分隐私(GDP)的指数机制,如果损失函数是强凸的,扰动是Lipschitz。我们的隐私界是\emph{optimal},因为它包括作为特殊情况的高斯机制的隐私,并证明了使用强对数凹测度的等周不等式。(2)我们展示了如何从$\exp(-F(x)-μ\|x\|^2_2/2)$中采样$G$ -Lipschitz $F$,$η$误差在总变差(TV)距离中使用$\widetilde{O}((G^2/μ) \log^2(d/η))$无偏查询$F(x)$。这是第一个其查询复杂度在维度$d$和准确度$η$上都具有\emph{polylogarithmic dependence}的采样器。

英文摘要:

In this paper, we study private optimization problems for non-smooth convex functions $F(x)=\mathbb{E}_i f_i(x)$ on $\mathbb{R}^d$. We show that modifying the exponential mechanism by adding an $\ell_2^2$ regularizer to $F(x)$ and sampling from $π(x)\propto \exp(-k(F(x)+μ\|x\|_2^2/2))$ recovers both the known optimal empirical risk and population loss under $(ε,δ)$-DP. Furthermore, we show how to implement this mechanism using $\widetilde{O}(n \min(d, n))$ queries to $f_i(x)$ for the DP-SCO where $n$ is the number of samples/users and $d$ is the ambient dimension. We also give a (nearly) matching lower bound $\widetildeΩ(n \min(d, n))$ on the number of evaluation queries. Our results utilize the following tools that are of independent interest: (1) We prove Gaussian Differential Privacy (GDP) of the exponential mechanism if the loss function is strongly convex and the perturbation is Lipschitz. Our privacy bound is \emph{optimal} as it includes the privacy of Gaussian mechanism as a special case and is proved using the isoperimetric inequality for strongly log-concave measures. (2) We show how to sample from $\exp(-F(x)-μ\|x\|^2_2/2)$ for $G$-Lipschitz $F$ with $η$ error in total variation (TV) distance using $\widetilde{O}((G^2/μ) \log^2(d/η))$ unbiased queries to $F(x)$. This is the first sampler whose query complexity has \emph{polylogarithmic dependence} on both dimension $d$ and accuracy $η$.

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。

↑