Optimal Guarantees for Auditing Rényi Differentially Private Machine Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kim, Benjamin D., Varshney, Lav R., Alabi, Daniel
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914585630474240
author Kim, Benjamin D.
Varshney, Lav R.
Alabi, Daniel
author_facet Kim, Benjamin D.
Varshney, Lav R.
Alabi, Daniel
contents We study black-box auditing for machine learning algorithms that claim R \ 'enyi differential privacy (RDP) guarantees. We introduce an auditing framework, based on hypothesis testing, that directly estimates Rényi divergence between neighboring executions using the Donsker-Varadhan (DV) variational estimator. Our analysis yields explicit and non-asymptotic confidence intervals for RDP auditing via class-restricted DV estimators, separating statistical estimation error from algorithmic privacy leakage. We prove matching minimax lower bounds showing that, up to logarithmic factors, our sample-complexity guarantees are information-theoretically optimal, thereby establishing the first optimal guarantees for auditing RDP via DV estimators. Empirically, we instantiate our framework for auditing DP-SGD in a fully black-box setting. Across MNIST and CIFAR-10, and over a wide range of privacy regimes, our auditors produce a strong overall improvement on empirical RDP lower bounds compared to prior state-of-the-art black-box methods especially at small and moderate Rényi orders where accurate auditing is most challenging.
format Preprint
id arxiv_https___arxiv_org_abs_2605_21938
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Optimal Guarantees for Auditing Rényi Differentially Private Machine Learning
Kim, Benjamin D.
Varshney, Lav R.
Alabi, Daniel
Machine Learning
Cryptography and Security
Information Theory
We study black-box auditing for machine learning algorithms that claim R \ 'enyi differential privacy (RDP) guarantees. We introduce an auditing framework, based on hypothesis testing, that directly estimates Rényi divergence between neighboring executions using the Donsker-Varadhan (DV) variational estimator. Our analysis yields explicit and non-asymptotic confidence intervals for RDP auditing via class-restricted DV estimators, separating statistical estimation error from algorithmic privacy leakage. We prove matching minimax lower bounds showing that, up to logarithmic factors, our sample-complexity guarantees are information-theoretically optimal, thereby establishing the first optimal guarantees for auditing RDP via DV estimators. Empirically, we instantiate our framework for auditing DP-SGD in a fully black-box setting. Across MNIST and CIFAR-10, and over a wide range of privacy regimes, our auditors produce a strong overall improvement on empirical RDP lower bounds compared to prior state-of-the-art black-box methods especially at small and moderate Rényi orders where accurate auditing is most challenging.
title Optimal Guarantees for Auditing Rényi Differentially Private Machine Learning
topic Machine Learning
Cryptography and Security
Information Theory
url https://arxiv.org/abs/2605.21938