A federated learning architecture for secure and private neuroimaging analysis
Patterns · 2024-08 · 5 (8) · 101031
DOI: 10.1016/j.patter.2024.101031 ↗Kurz gefasst
Fachartikel in der Zeitschrift Patterns, 2024-08. Verfasst von Dimitris Stripelis und 11 weiteren Autor:innen. Beteiligte Einrichtung: University of Southern California, Information Sciences Institute, Marina del Rey, CA 90292, USA. Die Arbeit wurde bisher 36-mal zitiert (Crossref) und nennt 106 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: NIA NIH HHS, Defense Advanced Research Projects Agency, National Institutes of Health.
Abstract (Originalsprache)
The amount of biomedical data continues to grow rapidly. However, collecting data from multiple sites for joint analysis remains challenging due to security, privacy, and regulatory concerns. To overcome this challenge, we use federated learning, which enables distributed training of neural network models over multiple data sources without sharing data. Each site trains the neural network over its private data for some time and then shares the neural network parameters (i.e., weights and/or gradients) with a federation controller, which in turn aggregates the local models and sends the resulting community model back to each site, and the process repeats. Our federated learning architecture, MetisFL, provides strong security and privacy. First, sample data never leave a site. Second, neural network parameters are encrypted before transmission and the global neural model is computed under fully homomorphic encryption. Finally, we use information-theoretic methods to limit information leakage from the neural model to prevent a "curious" site from performing model inversion or membership attacks. We present a thorough evaluation of the performance of secure, private federated learning in neuroimaging tasks, including for predicting Alzheimer's disease and for brain age gap estimation (BrainAGE) from magnetic resonance imaging (MRI) studies in challenging, heterogeneous federated environments where sites have different amounts of data and statistical distributions.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- University of Southern California, Information Sciences Institute, Marina del Rey, CA 90292, USA.
- University of Southern California, Computer Science Department, Los Angeles, CA 90089, USA.
- University of Southern California, Imaging Genetics Center, Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, Marina del Rey, CA 90292, USA.
- University of California, Riverside, Riverside, CA 92521, USA.
Förderung
- NIA NIH HHS
- Defense Advanced Research Projects Agency
- National Institutes of Health
Zitieren
APA
Stripelis, D., Gupta, U., Saleem, H., Dhinagar, N., Ghai, T., Anastasiou, C., Sánchez, R., Steeg, G. V., Ravi, S., Naveed, M., Thompson, P. M., & Ambite, J. L. (2024). A federated learning architecture for secure and private neuroimaging analysis. Patterns, 5(8), 101031. https://doi.org/10.1016/j.patter.2024.101031
Vancouver
Stripelis D, Gupta U, Saleem H, Dhinagar N, Ghai T, Anastasiou C, et al. A federated learning architecture for secure and private neuroimaging analysis. Patterns. 2024;5(8):101031. doi:10.1016/j.patter.2024.101031
BibTeX
@article{Stripelis2024,
author = {Stripelis, Dimitris and Gupta, Umang and Saleem, Hamza and Dhinagar, Nikhil and Ghai, Tanmay and Anastasiou, Chrysovalantis and Sánchez, Rafael and Steeg, Greg Ver and Ravi, Srivatsan and Naveed, Muhammad and Thompson, Paul M and Ambite, José Luis},
title = {A federated learning architecture for secure and private neuroimaging analysis},
journal = {Patterns},
year = {2024},
volume = {5},
number = {8},
pages = {101031},
doi = {10.1016/j.patter.2024.101031}
}
EUVIMED