DP-Sniper: Black-Box Discovery of Differential Privacy Violations using Classifiers
Authors:
- 1 May 2021
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 391-409
Abstract
We present DP-Sniper, a practical black-box method that automatically finds violations of differential privacy.DP-Sniper is based on two key ideas: (i) training a classifier to predict if an observed output was likely generated from one of two possible inputs, and (ii) transforming this classifier into an approximately optimal attack on differential privacy.Our experimental evaluation demonstrates that DP-Sniper obtains up to 12.4 times stronger guarantees than state-of-the-art, while being 15.5 times faster. Further, we show that DP-Sniper is effective in exploiting floating-point vulnerabilities of naively implemented algorithms: it detects that a supposedly 0.1-differentially private implementation of the Laplace mechanism actually does not satisfy even 0.25-differential privacy.Keywords
This publication has 29 references indexed in Scilit:
- The Composition Theorem for Differential PrivacyIEEE Transactions on Information Theory, 2017
- Understanding the sparse vector technique for differential privacyProceedings of the VLDB Endowment, 2017
- LightDP: towards automating differential privacy proofs2017
- Advanced Probabilistic Couplings for Differential Privacy2016
- Proving Differential Privacy via Probabilistic Couplings2016
- Higher-Order Approximate Relational Refinement Types for Mechanism Design and Differential Privacy2015
- Probabilistic Relational Reasoning for Differential PrivacyACM Transactions on Programming Languages and Systems, 2013
- Testing the Lipschitz Property over Product Distributions with Applications to Data Privacy2013
- The Tight Constant in the Dvoretzky-Kiefer-Wolfowitz InequalityThe Annals of Probability, 1990
- THE USE OF CONFIDENCE OR FIDUCIAL LIMITS ILLUSTRATED IN THE CASE OF THE BINOMIALBiometrika, 1934