Show simple item record

dc.contributor.authorBalog, Matej
dc.contributor.authorTolstikhin, Ilya
dc.contributor.authorSchölkopf, Bernhard
dc.date.accessioned2019-01-11T00:30:25Z
dc.date.available2019-01-11T00:30:25Z
dc.date.issued2017-10-04
dc.identifier.urihttps://www.repository.cam.ac.uk/handle/1810/287784
dc.description.abstractWe lay theoretical foundations for new database release mechanisms that allow third-parties to construct consistent estimators of population statistics, while ensuring that the privacy of each individual contributing to the database protected. The proposed framework rests on two main ideas. First, releasing (an estimate of) the kernel mean embedding of the data generating random variable instead of the database itself still allows third-parties to construct consistent estimators of a wide class of population statistics. Second, the algorithm can satisfy the definition of differential privacy by basing the released kernel mean embedding on entirely synthetic data points, while controlling accuracy through the metric available in a Reproducing Kernel Hilbert Space. We describe two instantiations of the proposed framework, suitable under different scenarios, and prove theoretical results guaranteeing differential privacy of the resulting algorithms and the consistency of estimators constructed from their outputs.
dc.publisherMIR Press
dc.subjectstat.ML
dc.subjectstat.ML
dc.titleDifferentially Private Database Release via Kernel Mean Embeddings
dc.typeConference Object
prism.endingPage431
prism.publicationNameProceedings of Machine Learning Research Volume 80:
prism.startingPage423
prism.volume80
dc.identifier.doi10.17863/CAM.35099
rioxxterms.versionofrecord10.17863/CAM.35099
rioxxterms.versionAM
rioxxterms.licenseref.urihttp://www.rioxx.net/licenses/all-rights-reserved
dc.contributor.orcidBalog, Matej [0000-0002-5552-9855]
rioxxterms.typeConference Paper/Proceeding/Abstract
pubs.funder-project-idEPSRC (1626332)
cam.issuedOnline2018-05-31
pubs.conference-nameInternational Conference on Machine Learning
pubs.conference-start-date2018-07-10
cam.orpheus.successThu Nov 05 11:53:17 GMT 2020 - Embargo updated
cam.orpheus.counter12
pubs.conference-finish-date2018-07-15
rioxxterms.freetoread.startdate2019-05-31


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record