A Polynomial Time, Pure Differentially Private Estimator for Binary Product Distributions

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteur principal: Singhal, Vikrant
Format: Preprint
Publié: 2023
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866917574960218112
author Singhal, Vikrant
author_facet Singhal, Vikrant
contents We present the first $\varepsilon$-differentially private, computationally efficient algorithm that estimates the means of product distributions over $\{0,1\}^d$ accurately in total-variation distance, whilst attaining the optimal sample complexity to within polylogarithmic factors. The prior work had either solved this problem efficiently and optimally under weaker notions of privacy, or had solved it optimally while having exponential running times.
format Preprint
id arxiv_https___arxiv_org_abs_2304_06787
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle A Polynomial Time, Pure Differentially Private Estimator for Binary Product Distributions
Singhal, Vikrant
Data Structures and Algorithms
Cryptography and Security
Machine Learning
We present the first $\varepsilon$-differentially private, computationally efficient algorithm that estimates the means of product distributions over $\{0,1\}^d$ accurately in total-variation distance, whilst attaining the optimal sample complexity to within polylogarithmic factors. The prior work had either solved this problem efficiently and optimally under weaker notions of privacy, or had solved it optimally while having exponential running times.
title A Polynomial Time, Pure Differentially Private Estimator for Binary Product Distributions
topic Data Structures and Algorithms
Cryptography and Security
Machine Learning
url https://arxiv.org/abs/2304.06787