WassersteinDistributionalNormalizationForRobustDistributionalCertificationofNoisyLabeledDataSungWooPark1JunseokKwon1AbstractWhilethereareseveralmethodsthatcandealwithnoisy-labeleddata,recentmethods...
ScalableComputationsofWassersteinBarycenterviaInputConvexNeuralNetworksJiaojiaoFan1AmirhosseinTaghvaei2YongxinChen1Abstractthepastfewyears,ithasfoundapplicationsinseveralma-chinelearningproblems.Fo...
RethinkingRotatedObjectDetectionwithGaussianWassersteinDistanceLossXueYang123JunchiYan12QiMing4WentaoWang1XiaopengZhang3QiTian3AbstractFigure1.ComparisonofthedetectionresultsbetweenSmoothL1loss-bas...
ProjectionRobustWassersteinBarycentersMinhuiHuang1ShiqianMa2LifengLai1AbstractHowever,computingtheWBforasetofprobabilitydis-tributionsisnotoriouslyhard.ThehardnesscomesfromCollectingandaggregatingi...
First-OrderMethodsforWassersteinDistributionallyRobustMDPsJulienGrand-Cle´ment1ChristianKroer1Abstractpolicies,astheyoptimizeonlyfortheworst-casekernelre-alization,withoutincorporatingdistribution...
DifferentiallyPrivateSlicedWassersteinDistanceAlainRakotomamonjy12LivaRalaivola1AbstracttinousadvancesmadeinMachineLearning(ML).How-ever,asethicalandregulatoryconcernsbecomeprominentDevelopingmachi...
AWassersteinMinimaxFrameworkforMixedLinearRegressionTheoDiamandis1YoninaC.Eldar2AlirezaFallah1FarzanFarnia1AsumanOzdaglar1Abstractexample,speechdataorgeneticdatamayexhibitaclustereddistributionbase...
StrongerandFasterWassersteinAdversarialAttacksKaiwenWu12AllenHouzeWang12YaoliangYu12Abstract✏=0.05✏=0.10✏=0.20✏=0.40Deepmodels,whilebeingextremelyflexibleand`1accurate,aresurprisinglyvulnerable...
StochasticOptimizationforRegularizedWassersteinEstimatorsMarinBallu1QuentinBerthet2FrancisBach3Abstractcantransportmasswithdistributionmeasureµtoanotherdistributionmeasureν,withminimalglobaltrans...
PrincipledLearningMethodforWassersteinDistributionallyRobustOptimizationwithLocalPerturbationsYongchanKwon1WonyoungKim2Joong-HoWon2MyungheeChoPaik2Abstractsothecomputationoftheriskin(1)isimpossible...
BridgingtheGapBetweenf-GANsandWassersteinGANsJiamingSong1StefanoErmon1AbstractVariousGANlearningprocedureshavebeenproposedfordifferentdiscrepancymeasures.f-GANs(NowozinGenerativeadversarialnetworks...
TheWassersteinTransformFacundoMe´moli12ZaneSmith3ZhengchaoWan1Abstracttanceinformationinordertobothenhancefeatures,suchasclusters,presentinthedata,andtodenoisethedata.AsaWeintroducetheWassersteint...
SubspaceRobustWassersteinDistancesFranc¸ois-PierrePaty1MarcoCuturi21Abstractdistancescangrowexponentiallyindimension(Dudley,1969;Fournier&Guillin,2015),whichmeansthatanirre-MakingsenseofWasserstei...
OntheComplexityofApproximatingWassersteinBarycentersAlexeyKroshnin123DarinaDvinskikh41PavelDvurechensky41AlexanderGasnikov512NazariiTupitsa15Ce´sarA.Uribe6Abstractlearningandoptimizationcommunitie...
WassersteinAdversarialExamplesviaProjectedSinkhornIterationsEricWong1FrankR.Schmidt2J.ZicoKolter34Abstract+∆W=Arapidlygrowingareaofworkhasstudiedtheex-+∆∞=istenceofadversarialexamples,datapoints...
WassersteinofWassersteinLossforLearningGenerativeModelsYonatanDukler1WuchenLi1AlexTongLin1GuidoMontu´far123AbstractGenerativeAdversarialNetworks(GANs).TheapplicationoftheWassersteinmetrictodefinet...
StochasticWassersteinBarycentersSebastianClaici1EdwardChien1JustinSolomon1AbstractsumofsquaredWassersteindistancestotheinputdistribu-Wepresentastochasticalgorithmtocomputethetions(Agueh&Carlier,201...
PolicyOptimizationasWassersteinGradientFlowsRuiyiZhang1ChangyouChen2ChunyuanLi1LawrenceCarin1Abstractwiththeenvironment.Policyoptimizationisacorecomponentofrein-Astandardtechniqueforpolicylearningi...
ASimulatedAnnealingBasedInexactOracleforWassersteinLossMinimizationJianboYe1JamesZ.Wang1JiaLi2Abstractryl(x,·)iscomputedinO(m)time,mbeingthecomplex-ityofoutcomevariablesxory.Thispartofcalculationi...
WassersteinGenerativeAdversarialNetworksMartinArjovsky1SoumithChintala2Le´onBottou12AbstractThetypicalremedyistoaddanoisetermtothemodeldis-tribution.Thisiswhyvirtuallyallgenerativemodelsde-Weintro...