Uniformity Testing under User-Level Local Privacy

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Canonne, Clément L., Gentle, Abigail, Singhal, Vikrant
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911223569711104
author Canonne, Clément L.
Gentle, Abigail
Singhal, Vikrant
author_facet Canonne, Clément L.
Gentle, Abigail
Singhal, Vikrant
contents We initiate the study of distribution testing under \emph{user-level} local differential privacy, where each of $n$ users contributes $m$ samples from the unknown underlying distribution. This setting, albeit very natural, is significantly more challenging that the usual locally private setting, as for the same parameter $\varepsilon$ the privacy guarantee must now apply to a full batch of $m$ data points. While some recent work consider distribution \emph{learning} in this user-level setting, nothing was known for even the most fundamental testing task, uniformity testing (and its generalization, identity testing). We address this gap, by providing (nearly) sample-optimal user-level LDP algorithms for uniformity and identity testing. Motivated by practical considerations, our main focus is on the private-coin, symmetric setting, which does not require users to share a common random seed nor to have been assigned a globally unique identifier.
format Preprint
id arxiv_https___arxiv_org_abs_2510_18379
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Uniformity Testing under User-Level Local Privacy
Canonne, Clément L.
Gentle, Abigail
Singhal, Vikrant
Data Structures and Algorithms
Cryptography and Security
Discrete Mathematics
We initiate the study of distribution testing under \emph{user-level} local differential privacy, where each of $n$ users contributes $m$ samples from the unknown underlying distribution. This setting, albeit very natural, is significantly more challenging that the usual locally private setting, as for the same parameter $\varepsilon$ the privacy guarantee must now apply to a full batch of $m$ data points. While some recent work consider distribution \emph{learning} in this user-level setting, nothing was known for even the most fundamental testing task, uniformity testing (and its generalization, identity testing). We address this gap, by providing (nearly) sample-optimal user-level LDP algorithms for uniformity and identity testing. Motivated by practical considerations, our main focus is on the private-coin, symmetric setting, which does not require users to share a common random seed nor to have been assigned a globally unique identifier.
title Uniformity Testing under User-Level Local Privacy
topic Data Structures and Algorithms
Cryptography and Security
Discrete Mathematics
url https://arxiv.org/abs/2510.18379