PuSH - Publication Server of Helmholtz Zentrum München

Export: TXT Text RIS Endnote (RIS) BIB BibTeX
Ziller, A.* ; Mueller, T.T.* ; Stieger, S. ; Feiner, L.F.* ; Brandt, J.* ; Braren, R.* ; Rueckert, D.* ; Kaissis, G.

Reconciling privacy and accuracy in AI for medical imaging.

Nat. Mach. Intell., DOI: 10.1038/s42256-024-00858-y (2024)
Publ. Version/Full Text Research data DOI
Open Access Hybrid
Creative Commons Lizenzvertrag
Artificial intelligence (AI) models are vulnerable to information leakage of their training data, which can be highly sensitive, for example, in medical imaging. Privacy-enhancing technologies, such as differential privacy (DP), aim to circumvent these susceptibilities. DP is the strongest possible protection for training models while bounding the risks of inferring the inclusion of training samples or reconstructing the original data. DP achieves this by setting a quantifiable privacy budget. Although a lower budget decreases the risk of information leakage, it typically also reduces the performance of such models. This imposes a trade-off between robust performance and stringent privacy. Additionally, the interpretation of a privacy budget remains abstract and challenging to contextualize. Here we contrast the performance of artificial intelligence models at various privacy budgets against both theoretical risk bounds and empirical success of reconstruction attacks. We show that using very large privacy budgets can render reconstruction attacks impossible, while drops in performance are negligible. We thus conclude that not using DP at all is negligent when applying artificial intelligence models to sensitive data. We deem our results to lay a foundation for further debates on striking a balance between privacy risks and model performance.
Impact Factor
Scopus SNIP
Altmetric
18.800
0.000
Tags
Annotations
Special Publikation
Hide on homepage

Edit extra information
Edit own tags
Private
Edit own annotation
Private
Hide on publication lists
on hompage
Mark as special
publikation
Publication type Article: Journal article
Document type Scientific Article
Language english
Publication Year 2024
HGF-reported in Year 2024
ISSN (print) / ISBN 2522-5839
e-ISSN 2522-5839
Journal Nature machine intelligence
Publisher Springer
Publishing Place [London]
Reviewing status Peer reviewed
Institute(s) Institute for Machine Learning in Biomed Imaging (IML)
POF-Topic(s) 30205 - Bioengineering and Digital Health
Research field(s) Enabling and Novel Technologies
PSP Element(s) G-507100-001
Grants German Academic Exchange Service (DAAD) under the Kondrad Zuse School of Excellence for Reliable AI (RelAI)
Bavarian Collaborative Research Project PRIPREKI of the Free State of Bavaria
Bavarian State Ministry for Science and the Arts through the Munich Centre for Machine Learning
Bavarian Cancer Research Center (BZKF, Lighthouse AI and Bioinformatics) - German Federal Ministry of Education and Research
Project 'NUM 2.0'
Federal Ministry of Education and Research (BMBF)
ERC Grant
German Ministry of Education and Research (BMBF)
Bundesministerium fr Bildung und Forschung (Federal Ministry of Education and Research)
DOI 10.1038/s42256-024-00858-y
WOS ID 001252364500003
Scopus ID 85196513360
Erfassungsdatum 2024-06-27
[➜Report correction]