Sequentially Auditing Differential Privacy

Tomás González, Mateo Dulce Rubio, Aaditya Ramdas, Mónica Ribero

Keywords: Differential privacy, testing by betting

TL;DR: We propose a sequential test for auditing differential privacy that detects violations with sample sizes that are orders of magnitude smaller than existing methods.

Abstract: We propose a practical sequential test for auditing differential privacy guarantees of black-box mechanisms. The test processes streams of mechanisms' outputs providing anytime-valid inference while controlling Type I error, overcoming the fixed sample size limitation of previous batch auditing methods. Experiments show this test detects violations with sample sizes that are orders of magnitude smaller than existing methods, reducing this number from 50K to a few hundred examples, across diverse realistic mechanisms. Notably, it identifies DP-SGD privacy violations in \textit{under} one training run, unlike prior methods needing full model training.

Primary Area: Social and economic aspects of machine learning (e.g., fairness, interpretability, human-AI interaction, privacy, safety, strategic behavior)

Submission Number: 23490