BayesianDifferentialPrivacyforMachineLearningAlekseiTriastcyn1BoiFaltings1AbstractDespitenotableadvances,differentiallyprivateMLstillsuffersfromtwomajorproblems:(a)utilitylossduetoTraditionaldiffer...
AlignedCrossEntropyforNon-AutoregressiveMachineTranslationMarjanGhazvininejad1VladimirKarpukhin1LukeZettlemoyer1OmerLevy1AbstractTargetYittastesprettygoodthoughNon-autoregressiveMachinetranslationm...
MixtureModelsforDiverseMachineTranslation:TricksoftheTradeTianxiaoShen1MyleOtt2MichaelAuli2Marc’AurelioRanzato2Abstractstylessuchasformal/informal,literal/notliteral,etc.Thisraisesthequestionofhow...
HOList:AnEnvironmentforMachineLearningofHigher-OrderTheoremProvingKshitijBansal1SarahLoos1MarkusRabe1ChristianSzegedy1StewartWilcox1AbstractriSpeech(Panayotovetal.,2015)forspeechrecognition,theNet...
FaultToleranceinIterative-ConvergentMachineLearningAurickQiao12BryonAragam3BingjingZhang1EricP.Xing123AbstractMachinelearning(ML)trainingalgorithmsoftenpossessaninherentself-correctingbehaviordueto...
DataShapley:EquitableValuationofDataforMachineLearningAmirataGhorbani1JamesZou2Abstractnentofthemarketplace,similartolaborandcapital(Posner&Weyl,2018).Ithasbeensuggestedthatcertaindatacon-Asdatabec...
AUCµ:APerformanceMetricforMulti-ClassMachineLearningModelsRossS.Kleiman1DavidPage12Abstractmulti-classAUCanalog.However,inourworkweshowthatMcanfailtoreturnascoreof1(perfectperformance),Theareaunde...
TowardsBlack-boxIterativeMachineTeachingWeiyangLiu1BoDai1XingguoLi2ZhenLiu1JamesM.Rehg1LeSong13AbstractCommonKnowledgeRepresentationSpaceInthispaper,wemakeanimportantstepto-wardstheblack-boxMachine...
ResidualUnfairnessinFairMachineLearningfromPrejudicedDataNathanKallus1AngelaZhou2Abstractnewquestionsaboutthepossibleharmsoflearningfromdatawhichissubjecttohistoricalbias.Unlikeclean-cutpre-Recentw...
OrthogonalMachineLearning:PowerandLimitationsLesterMackey1VasilisSyrgkanis1IliasZadik12Abstractinaccurateinferencesconcerningtheparametersofinterest(thetreatmenteffects).√DoubleMachinelearningprov...
GeneralizedRobustBayesianCommitteeMachineforLarge-scaleGaussianProcessRegressionHaitaoLiu1JianfeiCai2YiWang3Yew-SoonOng24Abstractpredictivedistributionsattestpoints.InordertoscalestandardGaussianpr...
DelayedImpactofFairMachineLearningLydiaT.Liu1SarahDean1EstherRolf1MaxSimchowitz1MoritzHardt1Abstractvantagedgroupsinthepopulation(ExecutiveOfficeofthePresident,2016;Barocas&Selbst,2016).Consequentl...
CompetitiveCachingwithMachineLearnedAdviceThodorisLykouris1SergeiVassilvitskii2Abstractdoingso,thesystemmayinvestitseffortsonreducingtheerroronthemajorityofinputs,attheexpenseofincreasedWedevelopaf...
AsynchronousByzantineMachineLearning(thecaseofSGD)GeorgiosDamaskinos1ElMahdiElMhamdi1RachidGuerraoui1RhicheekPatra1MahsaTaziki1Abstract1.IntroductionAsynchronousdistributedMachinelearningso-Tokeepu...
AnalyzingUncertaintyinNeuralMachineTranslationMyleOtt1MichaelAuli1DavidGrangier1Marc’AurelioRanzato1Abstractpublicbenchmarks.However,severalopenquestionsre-main(Koehn&Knowles,2017).Inthiswork,wean...
SARAH:ANovelMethodforMachineLearningProblemsUsingStochasticRecursiveGradientLamM.Nguyen1JieLiu1KatyaScheinberg12MartinTaka´cˇ1AbstractProblemsofthistypearisefrequentlyinsupervisedlearn-Inthispape...
Resource-efficientMachineLearningin2KBRAMfortheInternetofThingsAshishKumar1SaurabhGoyal2ManikVarma1Abstractory.TheBBCMicro:Bithasa32bitARMCortexM0mi-crocontrolleroperatingat16MHzwith16KBSRAMandThis...
Multi-ClassOptimalMarginDistributionMachineTengZhang1Zhi-HuaZhou1AbstractRecently,themargintheoryforBoostinghasfinallybeendefended(Gao&Zhou,2013),andhasdisclosedthattheRecentstudiesdisclosethatmaxi...
IterativeMachineTeachingWeiyangLiu1BoDai1AhmadHumayun1CharleneTay2ChenYu2LindaB.Smith2JamesM.Rehg1LeSong1AbstractPassiveDatasetProvidetrainingsetLearningSamplequeryInthispaper,weconsidertheproblemo...