PuSH - Publikationsserver des Helmholtz Zentrums München

Kaissis, G. ; Ziller, A.* ; Kolek, S.* ; Riess, A. ; Rueckert, D.*

Optimal privacy guarantees for a relaxed threat model: Addressing sub-optimal adversaries in differentially private machine learning.

In: (37th Conference on Neural Information Processing Systems (NeurIPS), 10-16 December 2023, New Orleans, LA). 10010 North Torrey Pines Rd, La Jolla, California 92037 Usa: Neural Information Processing Systems (nips), 2023. 24
Differentially private mechanisms restrict the membership inference capabilities of powerful (optimal) adversaries against machine learning models. Such adversaries are rarely encountered in practice. In this work, we examine a more realistic threat model relaxation, where (sub-optimal) adversaries lack access to the exact model training database, but may possess related or partial data. We then formally characterise and experimentally validate adversarial membership inference capabilities in this setting in terms of hypothesis testing errors. Our work helps users to interpret the privacy properties of sensitive data processing systems under realistic threat model relaxations and choose appropriate noise levels for their use-case.
Weitere Metriken?
[➜Einloggen]
Zusatzinfos bearbeiten [➜Einloggen]
Publikationstyp Artikel: Konferenzbeitrag
Korrespondenzautor
ISSN (print) / ISBN 1049-5258
Konferenztitel 37th Conference on Neural Information Processing Systems (NeurIPS)
Konferzenzdatum 10-16 December 2023
Konferenzort New Orleans, LA
Quellenangaben Band: , Heft: , Seiten: 24 Artikelnummer: , Supplement: ,
Verlag Neural Information Processing Systems (nips)
Verlagsort 10010 North Torrey Pines Rd, La Jolla, California 92037 Usa
Nichtpatentliteratur Publikationen
Institut(e) Institute for Machine Learning in Biomed Imaging (IML)
Förderungen German Federal Ministry of Education and Research
Konrad Zuse School of Excellence in Reliable AI (RelAI)
Bavarian State Ministry for Science and the Arts through the Munich Centre for Machine Learning (MCML)
Helmholtz Junior Research Group grant