Privacy-preserving data sharing via probabilistic modeling

Loading...
Thumbnail Image
Access rights
openAccess
Journal Title
Journal ISSN
Volume Title
A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
Date
2021-07-09
Major/Subject
Mcode
Degree programme
Language
en
Pages
10
Series
Patterns, Volume 2, issue 7
Abstract
Differential privacy allows quantifying privacy loss resulting from accession of sensitive personal data. Repeated accesses to underlying data incur increasing loss. Releasing data as privacy-preserving synthetic data would avoid this limitation but would leave open the problem of designing what kind of synthetic data. We propose formulating the problem of private data release through probabilistic modeling. This approach transforms the problem of designing the synthetic data into choosing a model for the data, allowing also the inclusion of prior knowledge, which improves the quality of the synthetic data. We demonstrate empirically, in an epidemiological study, that statistical discoveries can be reliably reproduced from the synthetic data. We expect the method to have broad use in creating high-quality anonymized data twins of key datasets for research.
Description
Funding Information: This work was supported by the Academy of Finland (grants 325573 , 325572 , 319264 , 313124 , 303816 , 303815 , 297741 , and 292334 and the Flagship program Finnish Center for Artificial Intelligence [FCAI]). We thank the Carat group for access to the Carat data ( http://carat.cs.helsinki.fi/ ) and the CARING study group ( https://www.caring-diabetes.eu/ ) for access to the ARD data. Publisher Copyright: © 2021 The Authors
Keywords
differential privacy, DSML 2: Proof-of-Concept: Data science output has been formulated, implemented, and tested for one domain/problem, machine learning, open data, probabilistic modeling, synthetic data
Other note
Citation
Jälkö, J, Lagerspetz, E, Haukka, J, Tarkoma, S, Honkela, A & Kaski, S 2021, ' Privacy-preserving data sharing via probabilistic modeling ', Patterns, vol. 2, no. 7, 100271 . https://doi.org/10.1016/j.patter.2021.100271