Metraipolis › archive › dc_ea5ce545c013

Reliable Estimation of Differential Privacy under Selection Bias — arXiv:2610.27021

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.27021
License (as archived): CC0-1.0
Archived: 2026-10-21T07:07:24.518Z
SHA-256 of the archived text: 45b2ad2866191e9e5c11515b81ee373e0b98aed05aff6a651f4d33cd06d2524a
Stable URL of this page: https://metraipolis.com/documento/dc_ea5ce545c013

How to cite

Reliable Estimation of Differential Privacy under Selection Bias — arXiv:2610.27021 — arXiv (en), https://arxiv.org/abs/2610.27021 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ea5ce545c013 (Metraipolis archive dc_ea5ce545c013, sha256 45b2ad2866191e9e5c11515b81ee373e0b98aed05aff6a651f4d33cd06d2524a)

BibTeX

@misc{dcea5ce545c013,
  title = {Reliable Estimation of Differential Privacy under Selection Bias — arXiv:2610.27021},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.27021},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ea5ce545c013; sha256 45b2ad2866191e9e5c11515b81ee373e0b98aed05aff6a651f4d33cd06d2524a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.27021. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Statistical privacy audits are essential for verifying the theoretical privacy guarantees of differentially private (DP) mechanisms. Auditors typically use Membership Inference Attacks (MIAs) to estimate a lower bound on the true privacy leakage of DP mechanisms at a specified confidence level, which should not exceed the theoretical privacy upper bound guaranteed by DP. However, we found that many auditing implementations in prior work introduce a selection bias when running MIAs, which leads to statistically invalid and overestimated lower bounds. Because most audits are not tight, this bias has thus far gone unnoticed: the estimated lower bounds remained well below the theoretical upper bounds. Nevertheless, accurately evaluating this lower bound is critical, as a flawed statistical assessment could lead to false alarms on correct mechanisms or an incorrect evaluation of the guarantee's tightness. This work employs tools from the statistical literature to correct this bias. We present an overview of the different correction methods and assess their respective advantages. Furthermore, we perform a comprehensive re-analysis of results returned by prior audit implementations and show that without statistical correction, these audits violate the theoretical upper bounds up to four times more often than their claimed confidence level implies. With correction applied, the resulting lower bounds are statistically valid, and violations occur at no more than the nominal rate.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Cryptography and Security (cs.CR) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Machine Learning (stat.ML).
Pagina del lavoro: https://arxiv.org/abs/2610.27021 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-21T07:05:43.519Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).