Show simple item record

dc.contributor.authorMohapatra, Shubhankar
dc.date.accessioned2020-05-28 16:51:48 (GMT)
dc.date.available2020-05-28 16:51:48 (GMT)
dc.date.issued2020-05-28
dc.date.submitted2020-05-21
dc.identifier.urihttp://hdl.handle.net/10012/15939
dc.description.abstractSupervised machine learning tasks require large labelled datasets. However, obtaining such datasets is a difficult task and often leads to noisy labels due to human errors or adversarial perturbation. Recent studies have shown multiple methods to tackle this problem in the non-private scenario, yet this remains an unsolved problem when the dataset is private. In this work, we aim to train a model on a sensitive dataset that contains noisy labels such that (i) the model has high test accuracy and (ii) the training process satisfies (ε,δ)-differential privacy. Noisy labels, as studied in our work, are generated by flipping labels in the training set, from the true source label(s) to other targets (s). Our approach, Diffindo, constructs a differentially private stochastic gradient descent algorithm which removes suspicious points based on their noisy gradients. We show experiments on datasets across multiple domains with different class balance properties. Our results show that the proposed algorithm can remove up to 100% of the points with noisy labels in the private scenario while restoring the precision of the targeted label and testing accuracy to its no-noise counterparts.en
dc.language.isoenen
dc.publisherUniversity of Waterlooen
dc.subjectNoisy Labelsen
dc.subjectPrivacyen
dc.subjectDifferential Privacyen
dc.subjectPrivate Machine Leaningen
dc.subjectDiffindoen
dc.subjectPrivate health informaticsen
dc.subjectPrivate robust learningen
dc.titleDifferentially Private Learning with Noisy Labelsen
dc.typeMaster Thesisen
dc.pendingfalse
uws-etd.degree.departmentDavid R. Cheriton School of Computer Scienceen
uws-etd.degree.disciplineComputer Scienceen
uws-etd.degree.grantorUniversity of Waterlooen
uws-etd.degreeMaster of Mathematicsen
uws.contributor.advisorHe, Xi
uws.contributor.advisorChen, Helen
uws.contributor.affiliation1Faculty of Mathematicsen
uws.published.cityWaterlooen
uws.published.countryCanadaen
uws.published.provinceOntarioen
uws.typeOfResourceTexten
uws.peerReviewStatusUnrevieweden
uws.scholarLevelGraduateen


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record


UWSpace

University of Waterloo Library
200 University Avenue West
Waterloo, Ontario, Canada N2L 3G1
519 888 4883

All items in UWSpace are protected by copyright, with all rights reserved.

DSpace software

Service outages