mlsanityData to 7 Oct 2026

Limits of Personalizing Differential Privacy Budgets

arXiv 2605.13503 ↗First posted 13 May 2026cs.CR cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

A key technical difficulty in differential privacy is selecting a privacy budget that satisfies privacy requirements while maximizing utility. A natural and well-studied workaround is to use personalized privacy budgets, which may differ across agents. In this paper, we show that personalized budgets come with major limitations and that for mean estimation, the dominant factor is not full personalization, but rather choosing the right effective privacy budget. This can be achieved through a simple thresholding operator that we describe. Compared with this thresholding baseline, the gains obtained by fully personalized mechanisms are limited. In particular, we precisely quantify the constant-factor improvement in settings with mixed private and public datasets and in private datasets with two levels of privacy requirements. We also establish upper bounds and identify regimes of maximal gain for arbitrary privacy requirements.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 21 of its 32 references that are in the map. First posted on arXiv 13 May 2026.

Builds on

21 of 32 references.

References · 21

  1. 2023How to DP-fy ML: A Practical Guide to Machine Learning with Differential Privacycited by 150
  2. 2020Individual Privacy Accounting via a Renyi Filtercited by 57
  3. 2019Average-Case Averages: Private Algorithms for Smooth Sensitivity and Mean Estimationcited by 51
  4. 2019Limits of Private Learning with Access to Public Datacited by 38
  5. 2022Private Estimation with Public Datacited by 31
  6. 2015Heterogeneous Differential Privacycited by 30
  7. 2022Optimal and Differentially Private Data Acquisition: Central and Local Mechanismscited by 22
  8. 2023Have it your way: Individualized Privacy Assignment for DP-SGDcited by 21
  9. 2025Correlated Noise Mechanisms for Differentially Private Learningcited by 13
  10. 2022Optimal Data Acquisition with Privacy-Aware Agentscited by 12
  11. 2022Individualized PATE: Differentially Private Machine Learning with Individual Privacy Guaranteescited by 10
  12. 2023Optimal Differentially Private Model Training with Public Datacited by 9
9 more
  1. 2014Redrawing the Boundaries on Purchasing Data from Privacy-Sensitive Individualscited by 8
  2. 2018Individual Sensitivity Preprocessing for Data Privacycited by 8
  3. 2023Mean Estimation Under Heterogeneous Privacy: Some Privacy Can Be Freecited by 6
  4. 2023Mean Estimation Under Heterogeneous Privacy Demandscited by 6
  5. 2024Oracle-Efficient Differentially Private Learning with Public Datacited by 5
  6. 2024Public-data Assisted Private Stochastic Optimization: Power and Limitationscited by 5
  7. 2025Setting varepsilon is not the Issue in Differential Privacycited by 3
  8. 2025Managing Correlations in Data and Privacy Demandcited by 2
  9. 2026Data Sharing with Endogenous Choices over Differential Privacy Levelscited by 1

Built on by

Not cited yet.