Privacy Auditing with One (1) Training Run

Thomas Steinke, Milad Nasr, Matthew Jagielski
5/15/2023
cs.LGcs.CRcs.DS

No official code found. Best implementation: zj-jayzhang/one_round_auditing (5 stars, last commit 2024-12-08).

Abstract

We propose a scheme for auditing differentially private machine learning systems with a single training run. This exploits the parallelism of being able to add or remove multiple training examples independently. We analyze this using the connection between differential privacy and statistical generalization, which avoids the cost of group privacy. Our auditing scheme requires minimal assumptions about the algorithm and can be applied in the black-box or white-box setting.

View on arXivView PDF

Code Implementations(1)

zj-jayzhang/one_round_auditingCommunity56%

This is an unofficial code for "Privacy Auditing with One (1) Training Run"

51Shell, PythonDec 8, 20241 years ago

Similar Papers

Tight Auditing of Differentially Private Machine Learning

91% match2023

One-shot Empirical Privacy Estimation for Federated Learning

86% match1 repo7552023

Auditing Differential Privacy Guarantees Using Density Estimation

86% match2024

Auditing f-Differential Privacy in One Run

85% match2024

Adversarial Sample-Based Approach for Tighter Privacy Auditing in Final Model-Only Scenarios

85% match2024

One-shot Empirical Privacy Estimation for Federated Learning

85% match1 repo7552026

Cite this paper

@article{steinke2023privacy,
  title  = {Privacy Auditing with One (1) Training Run},
  author = {Thomas Steinke and Milad Nasr and Matthew Jagielski},
  year   = {2023},
  eprint = {2305.08846},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2305.08846v1},
  journal = {NEURIPS 2023 2023}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2305.08846.svg)](https://codeofpaper.com/paper/2305.08846)

Discussion