Bounding User Contributions for User-Level Differentially Private Mean Estimation

V. Arvind Rameshwar, Anshoo Tandon · 2025

Read Bounding User Contributions for User-Level Differentially Private Mean Estimation by V. Arvind Rameshwar, Anshoo Tandon (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-04749.txt.

Description

Releasing an average while protecting each person's privacy sounds simple until one heavy user can swamp the sum. Arvind Rameshwar and Anshoo Tandon revisit in 2025 how to cap each user's contribution before averaging, under user-level differential privacy. Older studies assumed tidy data: independent samples, equal counts per user. Real datasets break both.

The authors pin down exactly which capping strategy minimizes the worst-case error over all datasets, allowing users to contribute unevenly. Experiments add a surprise: even on tidy independent data, their strategy beats the field's standard capping method on average error by a wide margin.

Catalog record

CREATORV. Arvind Rameshwar, Anshoo Tandon
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE36 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-04749

Browse the full library · About the Tower of Babel