Privately Estimating a Gaussian: Efficient, Robust, and Optimal
Privately Estimating a Gaussian: Efficient, Robust, and Optimal
复制标题
私下估计高斯:高效、稳健且最优
DOI:
10.1145/3564246.3585194
复制
发表时间:
2023
期刊:
影响因子:
--
通讯作者:
Zhang, Fred
中科院分区:
文献类型:
--
作者:
Alabi, Daniel;Kothari, Pravesh K.;Tankala, Pranay;Venkat, Prayaag;Zhang, Fred
In this work, we give efficient algorithms for privately estimating a Gaussian distribution in both pure and approximate differential privacy (DP) models with optimal dependence on the dimension in the sample complexity.In the pure DP setting, we give an efficient algorithm that estimates an unknownd-dimensional Gaussian distribution up to an arbitrary tiny total variation error usingO(d2logκ) samples while tolerating a constant fraction of adversarial outliers. Here, κ is the condition number of the target covariance matrix. The sample bound matches best non-private estimators in the dependence on the dimension (up to a polylogarithmic factor). We prove a new lower bound on differentially private covariance estimation to show that the dependence on the condition number κ in the above sample bound is also tight. Prior to our work, only identifiability results (yielding inefficient super-polynomial time algorithms) were known for the problem.In the approximate DP setting, we give an efficient algorithm to estimate an unknown Gaussian distribution up to an arbitrarily tiny total variation error usingO(d2) samples while tolerating a constant fraction of adversarial outliers. Prior to our work, all efficient approximate DP algorithms incurred a super-quadratic sample cost or were not outlier-robust. For the special case of mean estimation, our algorithm achieves the optimal sample complexity ofO(d), improving on aO(d1.5) bound from prior work.Our pure DP algorithm relies on a recursive private preconditioning subroutine that utilizes recent work of Hopkins et al. (STOC 2022) on private mean estimation. Our approximate DP algorithms are based on a substantial upgrade of the method of stabilizing convex relaxations introduced by Kothari et al. (COLT 2022). In particular, we improve on their mechanism by using a new unnormalized entropy regularization and a new and surprisingly simple mechanism for privately releasing covariances.
登录
查看更多内容
DOI:
--
发表时间:
2021
期刊:
Symposium on the Theory of Computing
影响因子:
--
作者:
Samuel B. Hopkins;Gautam Kamath;Mahbod Majid
通讯作者:
Mahbod Majid
DOI:
10.1137/1.9781611975031.171
发表时间:
2017-04
期刊:
ArXiv
影响因子:
--
作者:
Ilias Diakonikolas;Gautam Kamath;D. Kane;Jerry Li;Ankur Moitra;Alistair Stewart
通讯作者:
Ilias Diakonikolas;Gautam Kamath;D. Kane;Jerry Li;Ankur Moitra;Alistair Stewart
DOI:
--
发表时间:
2018-05
期刊:
--
影响因子:
--
作者:
Gautam Kamath;Jerry Li;Vikrant Singhal;Jonathan Ullman
通讯作者:
Gautam Kamath;Jerry Li;Vikrant Singhal;Jonathan Ullman
DOI:
--
发表时间:
2021-12
期刊:
--
影响因子:
--
作者:
Pravesh Kothari;Pasin Manurangsi;A. Velingker
通讯作者:
Pravesh Kothari;Pasin Manurangsi;A. Velingker
DOI:
--
发表时间:
2021
期刊:
International Conference on Machine Learning
影响因子:
--
作者:
Eliad Tsfadia;E. Cohen;Haim Kaplan;Y. Mansour;Uri Stemmer
通讯作者:
Uri Stemmer