What type of study is this?

This is a Quantitative Study study.

October 12, 2025Open Access

Sequentially Auditing Differential Privacy

Puntos clave

The proposed method detects differential privacy violations significantly faster than previous techniques, enhancing efficiency.
It achieves Type I error control while allowing decomposition of the auditing processes, crucial for black-box mechanisms.
Unlike older auditing methods, this test can identify violations within a single training run, minimizing resource use.
Experiments show a dramatic reduction in needed sample sizes, needing only a few hundred examples instead of 50K.

Resumen

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 under one training run, unlike prior methods needing full model training.

Leer artículo completoexternamente

Me gusta

Guardar

Ver artículo completo