Limits of Personalizing Differential Privacy Budgets
Edwige Cyffers, Juba Ziani · 2026
Read Limits of Personalizing Differential Privacy Budgets by Edwige Cyffers, Juba Ziani (2026) 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-2605-13503.txt.
Description
Edwige Cyffers and Juba Ziani show that personalised privacy budgets have major limits, at least for mean estimation. Choosing a budget that meets privacy needs while maximising utility is hard, and budgets that differ across agents seem natural. The dominant factor is not full personalisation but the right effective budget, which a simple thresholding operator achieves. Compared with that baseline, gains from full personalisation are limited. They quantify constant-factor improvement with mixed datasets and private datasets at two privacy levels. They establish upper bounds and regimes of maximal gain.
Catalog record
| CREATOR | Edwige Cyffers, Juba Ziani |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50351 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.13503