PrivateAlternatingLeastSquares:PracticalPrivateMatrixCompletionwithTighterRatesSteveChien1PrateekJain1WalidKrichene1SteffenRendle1ShuangSong1AbhradeepThakurta1LiZhang1Abstractknownuser-itemratingsm...
PrivateStochasticConvexOptimization:OptimalRatesin1GeometryHilalAsi1VitalyFeldman2TomerKoren3KunalTalwar2AbstractInthisproblem(DP-SCO),givenni.i.d.samplesz1,...,znfromadistributionP,wewishtorelease...
PrivateAdaptiveGradientMethodsforConvexOptimizationHilalAsi12JohnDuchi23AlirezaFallah41OmidJavidbakht5KunalTalwar5AbstractopingPrivatevariantsofstochasticgradientdescent(SGD),wherealgorithmsguarant...
PracticalandPrivate(Deep)LearningWithoutSamplingorShufflingPeterKairouz1BrendanMcMahan1ShuangSong1OmThakkar1AbhradeepThakurta1ZhengXu1Abstractinthecontextofdistributedsettingslikefederatedlearn-ing...
PAPRIKA:PrivateOnlineFalseDiscoveryRateControlWanrongZhang1GautamKamath2RachelCummings3Abstractoflarge-scaledatasetsandeaseofdataanalysis,whilebene-ficialtosociety,hascreatedaseverecrisisofreproduc...
OneshotDifferentiallyPrivateTop-kSelectionGangQiao1WeijieJ.Su2LiZhang3Abstractkitemswith(approximately)themaximumvaluesamongmgivenvalues.TheproblemofPrivatelyreportingtheBeingabletoefficientlyandac...
LosslessCompressionofEfficientPrivateLocalRandomizersVitalyFeldman1KunalTalwar1AbstractTheconceptofalocalrandomizerdatesbacktotheworkofWarner(1965)whereitwasusedtoencouragetruthful-LocallyDifferent...
LeveragingPublicDataforPracticalPrivateQueryReleaseTerranceLiu1GiuseppeVietri2ThomasSteinke3JonathanUllman4ZhiweiStevenWu1AbstractInpractice,generatingaccuratedifferentiallyPrivatesyn-theticdataset...
LargeScalePrivateLearningviaLow-rankReparametrizationDaYu12HuishuaiZhang2WeiChen2JianYin1Tie-YanLiu2AbstractFigure1.Theproposedreparametrizationscheme.Theresidualweightmakesthereparametrizedoutputt...
DifferentiallyPrivateSlicedWassersteinDistanceAlainRakotomamonjy12LivaRalaivola1AbstracttinousadvancesmadeinMachineLearning(ML).How-ever,asethicalandregulatoryconcernsbecomeprominentDevelopingmachi...
DifferentiallyPrivateQuantilesJenniferGillenwaterMatthewJosephAlexKuleszaAbstracttothisproblem.Informally,thedistributionoveradifferen-tiallyPrivatealgorithm’soutputsmustberelativelyinsen-Quantile...
DifferentiallyPrivateQueryReleaseThroughAdaptiveProjectionSergulAydore1WilliamBrown12MichaelKearns13KrishnaramKenthapadi1LucaMelis1AaronRoth13AnkitSiva1AbstractGaussiannoise.Foradatasetofsizen,this...
DifferentiallyPrivateDensestSubgraphDetectionDungNguyen12AnilVullikanti12Abstract2018).However,networkproblemshavebeenproventobemuchharder,andPrivatealgorithmsformanyproblemsDensestsubgraphdetectio...
DifferentiallyPrivateCorrelationClusteringMarkBun1MarekEliáš2JanardhanKulkarni3Abstractinformation,e.g.,fromtransactionrecordsandpersonalcommunications.Moreover,theinformationwehaveonCorrelationc...
DifferentiallyPrivateBayesianInferenceforGeneralizedLinearModelsTejasKulkarni1JoonasJa¨lko¨1AnttiKoskela2SamuelKaski13AnttiHonkela2Abstractontheregressioncoefficients,suchasdeterminingwhichcoeffi...
DifferentiallyPrivateAggregationintheShuffleModel:AlmostCentralAccuracyinAlmostaSingleMessageBadihGhazi1RaviKumar1PasinManurangsi1RasmusPagh21AmerSinha3Abstractwithextensionsincluding(Private)distr...
DeepReDuce:ReLUReductionforFastPrivateInferenceNandanKumarJha1ZahraGhodsi1SiddharthGarg1BrandonReagen1AbstractTest(top-1)accuracy(%)80DeepReDuceTherecentriseofprivacyconcernshasledre-75SAFENetsearc...
AFrameworkforPrivateMatrixAnalysisinSlidingWindowModelJalajUpadhyay1SarvagyaUpadhyay2Abstractanalysis,includingmatrixanalysis,accuratelycanleakpri-vateinformation(Narayanan&Shmatikov,2006).Asare-We...
PrivateQueryReleaseAssistedbyPublicDataRaefBassily1AlbertCheu2ShayMoran3AleksandarNikolov4JonathanUllman2ZhiweiStevenWu5Abstracterrorα.Acentralquestioninthisproblemisconcernedwithcharacterizingthe...