Home
Deutsch
Advanced Search
Browse by ...
Publication overview
Contact persons
Help
DOI
PMC
 |
|
as soon as is submitted to ZB.
Publishing neural networks in drug discovery might compromise training data privacy.
J. Cheminformatics 17:38 (2025)
This study investigates the risks of exposing confidential chemical structures when machine learning models trained on these structures are made publicly available. We use membership inference attacks, a common method to assess privacy that is largely unexplored in the context of drug discovery, to examine neural networks for molecular property prediction in a black-box setting. Our results reveal significant privacy risks across all evaluated datasets and neural network architectures. Combining multiple attacks increases these risks. Molecules from minority classes, often the most valuable in drug discovery, are particularly vulnerable. We also found that representing molecules as graphs and using message-passing neural networks may mitigate these risks. We provide a framework to assess privacy risks of classification models and molecular representations, available at https://github.com/FabianKruger/molprivacy . Our findings highlight the need for careful consideration when sharing neural networks trained on proprietary chemical structures, informing organisations and researchers about the trade-offs between data confidentiality and model openness.
Altmetric
Additional Metrics?
Edit extra informations
Login
Publication type
Article: Journal article
Document type
Scientific Article
Keywords
Cheminformatics ; Drug Discovery ; Machine Learning ; Membership Inference Attack ; Privacy ; Qsar
e-ISSN
1758-2946
Journal
Journal of Cheminformatics
Quellenangaben
Volume: 17,
Issue: 1,
Article Number: 38
Publisher
BioMed Central
Publishing Place
Campus, 4 Crinan St, London N1 9xw, England
Non-patent literature
Publications
Reviewing status
Peer reviewed
Institute(s)
Institute of Structural Biology (STB)
Grants
HORIZON EUROPE European Research Council