{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,6]],"date-time":"2026-08-06T17:50:48Z","timestamp":1786038648898,"version":"3.56.0"},"reference-count":32,"publisher":"Oxford University Press (OUP)","issue":"9","license":[{"start":{"date-parts":[[2023,8,30]],"date-time":"2023-08-30T00:00:00Z","timestamp":1693353600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100001659","name":"German Research Foundation","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100001659","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001659","name":"Deutsche Forschungsgemeinschaft","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100001659","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2023,9,2]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Motivation<\/jats:title>\n                    <jats:p>Federated Learning (FL) is gaining traction in various fields as it enables integrative data analysis without sharing sensitive data, such as in healthcare. However, the risk of data leakage caused by malicious attacks must be considered. In this study, we introduce a novel attack algorithm that relies on being able to compute sample means, sample covariances, and construct known linearly independent vectors on the data owner side.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>We show that these basic functionalities, which are available in several established FL frameworks, are sufficient to reconstruct privacy-protected data. Additionally, the attack algorithm is robust to defense strategies that involve adding random noise. We demonstrate the limitations of existing frameworks and propose potential defense strategies analyzing the implications of using differential privacy. The novel insights presented in this study will aid in the improvement of FL frameworks.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Availability and implementation<\/jats:title>\n                    <jats:p>The code examples are provided at GitHub (https:\/\/github.com\/manuhuth\/Data-Leakage-From-Covariances.git). The CNSIM1 dataset, which we used in the manuscript, is available within the DSData R package (https:\/\/github.com\/datashield\/DSData\/tree\/main\/data).<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1093\/bioinformatics\/btad531","type":"journal-article","created":{"date-parts":[[2023,8,29]],"date-time":"2023-08-29T08:03:43Z","timestamp":1693296223000},"source":"Crossref","is-referenced-by-count":8,"title":["Accessibility of covariance information creates vulnerability in Federated Learning frameworks"],"prefix":"10.1093","volume":"39","author":[{"ORCID":"https:\/\/orcid.org\/0009-0004-4518-5704","authenticated-orcid":false,"given":"Manuel","family":"Huth","sequence":"first","affiliation":[{"name":"Institute of Computational Biology, Helmholtz Munich , Neuherberg 85764, Germany"},{"name":"Life and Medical Sciences Institute, Faculty of Mathematics and Natural Sciences, University of Bonn , Bonn 53115, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jonas","family":"Arruda","sequence":"additional","affiliation":[{"name":"Life and Medical Sciences Institute, Faculty of Mathematics and Natural Sciences, University of Bonn , Bonn 53115, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Roy","family":"Gusinow","sequence":"additional","affiliation":[{"name":"Institute of Computational Biology, Helmholtz Munich , Neuherberg 85764, Germany"},{"name":"Life and Medical Sciences Institute, Faculty of Mathematics and Natural Sciences, University of Bonn , Bonn 53115, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7901-2172","authenticated-orcid":false,"given":"Lorenzo","family":"Contento","sequence":"additional","affiliation":[{"name":"Life and Medical Sciences Institute, Faculty of Mathematics and Natural Sciences, University of Bonn , Bonn 53115, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2010-4977","authenticated-orcid":false,"given":"Evelina","family":"Tacconelli","sequence":"additional","affiliation":[{"name":"Division of Infectious Diseases, Department of Diagnostics and Public Health, University of Verona , Verona 37124, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4935-3312","authenticated-orcid":false,"given":"Jan","family":"Hasenauer","sequence":"additional","affiliation":[{"name":"Life and Medical Sciences Institute, Faculty of Mathematics and Natural Sciences, University of Bonn , Bonn 53115, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2023,8,30]]},"reference":[{"key":"2023092219034778600_btad531-B1","author":"Abadi","year":"2015"},{"key":"2023092219034778600_btad531-B2","first-page":"308","author":"Abadi","year":"2016"},{"key":"2023092219034778600_btad531-B3","doi-asserted-by":"crossref","first-page":"1329","DOI":"10.1038\/s41591-021-01460-0","article-title":"Remove obstacles to sharing health data with researchers outside of the European union","volume":"27","author":"Bentzen","year":"2021","journal-title":"Nat Med"},{"key":"2023092219034778600_btad531-B4","author":"Chen","year":"2019"},{"key":"2023092219034778600_btad531-B5","doi-asserted-by":"crossref","first-page":"1735","DOI":"10.1038\/s41591-021-01506-3","article-title":"Federated learning for predicting clinical outcomes in patients with COVID-19","volume":"27","author":"Dayan","year":"2021","journal-title":"Nat Med"},{"key":"2023092219034778600_btad531-B6","doi-asserted-by":"crossref","first-page":"211","DOI":"10.1561\/0400000042","article-title":"The algorithmic foundations of differential privacy","volume":"9","author":"Dwork","year":"2014","journal-title":"FNT Theor Comput Sci"},{"key":"2023092219034778600_btad531-B7","doi-asserted-by":"crossref","first-page":"214001","DOI":"10.1088\/1361-6560\/ac97d9","article-title":"OpenFL: the open federated learning library","volume":"67","author":"Foley","year":"2022","journal-title":"Phys Med Biol"},{"key":"2023092219034778600_btad531-B8","first-page":"16937","article-title":"Inverting gradients-how easy is it to break privacy in federated learning?","volume":"33","author":"Geiping","year":"2020","journal-title":"Adv Neural Inf Process Syst"},{"key":"2023092219034778600_btad531-B9","author":"Hansen","year":"2021"},{"key":"2023092219034778600_btad531-B10","doi-asserted-by":"crossref","first-page":"e1003321","DOI":"10.1371\/journal.pmed.1003321","article-title":"Comorbidities associated with mortality in 31,461 adults with COVID-19 in the United States: a federated electronic medical record analysis","volume":"17","author":"Harrison","year":"2020","journal-title":"PLoS Med"},{"key":"2023092219034778600_btad531-B11","first-page":"7232","article-title":"Evaluating gradient inversion attacks and defenses in federated learning","volume":"34","author":"Huang","year":"2021","journal-title":"Adv Neural Inf Process Syst"},{"key":"2023092219034778600_btad531-B12","doi-asserted-by":"crossref","first-page":"3649","DOI":"10.1007\/s00394-022-02909-9","article-title":"Associations between exploratory dietary patterns and incident type 2 diabetes: a federated meta-analysis of individual participant data from 25 cohort studies","volume":"61","author":"Jannasch","year":"2022","journal-title":"Eur J Nutr"},{"key":"2023092219034778600_btad531-B9630962","doi-asserted-by":"crossref","first-page":"20","DOI":"10.1109\/MSEC.2020.3039941","article-title":"A taxonomy of attacks on federated learning","volume":"19","author":"Jere","year":"2021","journal-title":"IEEE Security"},{"key":"2023092219034778600_btad531-B13","author":"Ludwig","year":"2020"},{"key":"2023092219034778600_btad531-B15","doi-asserted-by":"crossref","first-page":"e1008880","DOI":"10.1371\/journal.pcbi.1008880","article-title":"Orchestrating privacy-protected big data analyses of data from different resources with R and DataSHIELD","volume":"17","author":"Marcon","year":"2021","journal-title":"PLoS Comput Biol"},{"key":"2023092219034778600_btad531-B16","first-page":"1273","author":"McMahan"},{"key":"2023092219034778600_btad531-B17","doi-asserted-by":"crossref","first-page":"82","DOI":"10.1016\/j.jaci.2022.01.023","article-title":"Associations of earlylife pet ownership with asthma and allergic sensitization: a meta-analysis of more than 77,000 children from the EU child cohort network","volume":"150","author":"Moira","year":"2022","journal-title":"J Allergy Clin Immunol"},{"key":"2023092219034778600_btad531-B18","doi-asserted-by":"crossref","first-page":"1223","DOI":"10.3390\/nu13041223","article-title":"Heterogeneity of associations between total and types of fish intake and the incidence of type 2 diabetes: federated meta-analysis of 28 prospective studies including 956,122 participants","volume":"13","author":"Pastorino","year":"2021","journal-title":"Nutrients"},{"key":"2023092219034778600_btad531-B19","first-page":"8024","author":"Paszke"},{"key":"2023092219034778600_btad531-B20","doi-asserted-by":"crossref","first-page":"1231","DOI":"10.1093\/jn\/nxaa447","article-title":"Associations of total legume, pulse, and soy consumption with incident type 2 diabetes: federated meta-analysis of 27 studies from diverse world regions","volume":"151","author":"Pearce","year":"2021","journal-title":"J Nutr"},{"key":"2023092219034778600_btad531-B21","doi-asserted-by":"crossref","first-page":"e055630","DOI":"10.1136\/bmjopen-2021-055630","article-title":"Unravelling data for rapid evidence-based response to COVID-19: a summary of the unCoVer protocol","volume":"11","author":"Pe\u00f1alvo","year":"2021","journal-title":"BMJ Open"},{"key":"2023092219034778600_btad531-B22","doi-asserted-by":"crossref","first-page":"4734","DOI":"10.1109\/TCOMM.2020.2990686","article-title":"Federated learning with blockchain for autonomous vehicles: analysis and design challenges","volume":"68","author":"Pokhrel","year":"2020","journal-title":"IEEE Trans Commun"},{"key":"2023092219034778600_btad531-B23","doi-asserted-by":"crossref","first-page":"152","DOI":"10.1109\/MNET.011.2000430","article-title":"Federated learning in vehicular networks: opportunities and solutions","volume":"35","author":"Posner","year":"2021","journal-title":"IEEE Netw"},{"key":"2023092219034778600_btad531-B24","doi-asserted-by":"crossref","first-page":"37","DOI":"10.1038\/s41591-018-0272-7","article-title":"Privacy in the age of medical big data","volume":"25","author":"Price","year":"2019","journal-title":"Nat Med"},{"key":"2023092219034778600_btad531-B25","author":"Roth","year":"2022"},{"key":"2023092219034778600_btad531-B26","doi-asserted-by":"crossref","DOI":"10.1007\/978-981-19-8532-4","volume-title":"Krylov Subspace Methods for Linear Systems","author":"Sogabe","year":"2022","edition":"1st edn"},{"key":"2023092219034778600_btad531-B27","doi-asserted-by":"crossref","first-page":"100467","DOI":"10.1016\/j.lanepe.2022.100467","article-title":"Challenges of data sharing in European COVID-19 projects: a learning opportunity for advancing pandemic preparedness and response","volume":"21","author":"Tacconelli","year":"2022","journal-title":"Lancet Reg Health Eur"},{"key":"2023092219034778600_btad531-B28","doi-asserted-by":"crossref","first-page":"265","DOI":"10.1038\/s41586-021-03583-3","article-title":"Swarm learning for decentralized and confidential clinical machine learning","volume":"594","author":"Warnat-Herresthal","year":"2021","journal-title":"Nature"},{"key":"2023092219034778600_btad531-B29","author":"Westerberg","year":"2022."},{"key":"2023092219034778600_btad531-B31","first-page":"16337","author":"Yin","year":"2021"},{"key":"2023092219034778600_btad531-B33","author":"Zhu","year":"8--14  2019. , , :  , ."},{"key":"2023092219034778600_btad531-B34","doi-asserted-by":"crossref","first-page":"111","DOI":"10.1007\/978-3-030-70604-3_5","volume-title":"Federated Learning Systems","author":"Ziller","year":"2021"}],"container-title":["Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bioinformatics\/advance-article-pdf\/doi\/10.1093\/bioinformatics\/btad531\/51308763\/btad531.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/39\/9\/btad531\/51718344\/btad531.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/39\/9\/btad531\/51718344\/btad531.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,9,22]],"date-time":"2023-09-22T15:05:16Z","timestamp":1695395116000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article\/doi\/10.1093\/bioinformatics\/btad531\/7255908"}},"subtitle":[],"editor":[{"given":"Alfonso","family":"Valencia","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2023,8,30]]},"references-count":32,"journal-issue":{"issue":"9","published-print":{"date-parts":[[2023,9,2]]}},"URL":"https:\/\/doi.org\/10.1093\/bioinformatics\/btad531","relation":{"has-preprint":[{"id-type":"doi","id":"10.1101\/2022.10.09.511497","asserted-by":"object"}]},"ISSN":["1367-4811"],"issn-type":[{"value":"1367-4811","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2023,9,1]]},"published":{"date-parts":[[2023,8,30]]},"article-number":"btad531"}}