Order-AgnosticCrossEntropyforNon-AutoregressiveMachineTranslationCunxiaoDu1ZhaopengTu2JingJiang1AbstractAnumberofrecenteffortshaveexploredwaystoimprovetheNATmodels’abilitytohandlemultimodality.One...
OnlineUnrelatedMachineLoadBalancingwithPredictionsRevisitedShiLi1JiayiXian1Abstractlearnedeasily.Aspredictionsareoftenerror-prone,ide-allytheperformanceofthealgorithmshoulddeteriorateWestudytheonli...
MatrixSketchingforSecureCollaborativeMachineLearningMengjiaoZhang1ShusenWang1Abstractclients,eachclientusesabatchoflocaldatatocomputeastochasticgradient,andtheserveraggregatesthestochas-Collaborati...
MassivelyParallelandAsynchronousTsetlinMachineArchitectureSupportingAlmostConstant-TimeScalingKurugeDarshanaAbeyrathna1BimalBhattarai1MortenGoodwin1SaeedRahimiGorji1Ole-ChristofferGranmo1LeiJiao1Ru...
Markpainting:AdversarialMachineLearningmeetsInpaintingDavidKhachaturov1IliaShumailov12YirenZhao1NicolasPapernot2RossAnderson1Abstractareainwithartificiallygeneratedcontent.Inpaintingisalearnedinter...
MachineUnlearningforRandomForestsJonathanBrophy1DanielLowd1Abstract1.IntroductionRespondingtouserdatadeletionrequests,remov-Recentlegislation(EU,2016;California,2018;Canada,ingnoisyexamples,ordelet...
MixedCrossEntropyLossforNeuralMachineTranslationHaoranLi1WeiLu1AbstractSourceDieGeburtenrategehtweiterzurück.one-hotmodelgroundpredictiontruthInneuralMachinetranslation,crossentropy(CE)isthestanda...
LabelDistributionLearningMachineJingWang12XinGeng12AbstractHAP:1.39descriptiondegree0.30SAD:2.030.25AlthoughLabelDistributionLearning(LDL)hasSUR:1.57witnessedextensiveclassificationapplications,itA...
KnowledgeEnhancedMachineLearningPipelineagainstDiverseAdversarialAttacksNeziheMerveGu¨rel1XiangyuQi2LukaRimanic1CeZhang1BoLi3Abstract1.IntroductionDespitethegreatsuccessesachievedbydeepneu-Recents...
Cross-modelBack-translatedDistillationforUnsupervisedMachineTranslationXuan-PhiNguyen12ShafiqJoty13Thanh-TungNguyen12WuKui2AiTiAw2Abstractmonolingualdatahasbeenactive.WhileRavi&Knight(2011)andKleme...
TransferLearningwithoutKnowing:ReprogrammingBlack-boxMachineLearningModelswithScarceDataandLimitedResourcesYun-YunTsai1Pin-YuChen2Tsung-YiHo1Abstractatargetdomainbasedontheknowledgetransferredfroma...
TheNon-IIDDataQuagmireofDecentralizedMachineLearningKevinHsieh12AmarPhanishayee1OnurMutlu32PhillipB.Gibbons2AbstractTheseconstraintsmotivatetheneedforMLtrainingoverwidelydistributeddata(decentraliz...
OnLearningLanguage-InvariantRepresentationsforUniversalMachineTranslationHaoZhao1JunjieHu1AndrejRisteski1Abstractlectionofparallelcorporaatalargescale,ontheorderofmillionsorevenbillionsofalignedsen...
MeasuringNon-ExpertComprehensionofMachineLearningFairnessMetricsDebjaniSaha1CandiceSchumann1DuncanC.McElfresh1JohnP.Dickerson1MichelleL.Mazurek1MichaelCarlTschantz2Abstractoffairnesscanhavesignific...
AutoML-Zero:EvolvingMachineLearningAlgorithmsFromScratchEstebanReal1ChenLiang1DavidR.So1QuocV.Le1Abstractfield,rangingfromlearningstrategiestonewarchitectures[Rumelhartetal.,1986;LeCunetal.,1995;Ho...
EvaluatingMachineAccuracyonImageNetVaishaalShankar1RebeccaRoelofs2HoriaMania1AlexFang1BenjaminRecht1LudwigSchmidt1AbstractInthispaper,wecontextualizeprogressonImageNetbycomparingawiderangeofImageNe...
DeltaGrad:RapidretrainingofMachinelearningmodelsYinjunWu1EdgarDobriban2SusanB.Davidson1AbstractFigure1.RunningtimeofourDeltaGradalgorithmforretrainingalogisticregressionmodelonRCV1asafunctionofthef...
CoresetsforData-efficientTrainingofMachineLearningModelsBaharanMirzasoleiman1JeffBilmes2JureLeskovec1AbstractTrainingMachinelearningmodelsoftenreducestooptimiz-Incrementalgradient(IG)methods,suchas...
CollaborativeMachineLearningwithIncentive-AwareModelRewardsRachaelHweeLingSim1YehongZhang1MunChoonChan1BryanKianHsiangLow1Abstractfromotherhospitalsandfirmstoimprovethepredictionofsomediseaseprogre...
CertifiedDataRemovalfromMachineLearningModelsChuanGuo1TomGoldstein2AwniHannun2LaurensvanderMaaten2Abstractinferenceattacks(Yeometal.,2018;Carlinietal.,2019)areunsuccessfulondatathatwasremovedfromth...