AffineInvariantAnalysisofFrank-WolfeonStronglyConvexSetsThomasKerdreux1LewisLiu23SimonLacosteJulien2345DamienScieur43AbstractAlgorithm1Frank-WolfeAlgorithmLMOLine-searchItisknownthattheFrank-Wolfe(...
AccuracyontheLine:ontheStrongCorrelationBetweenOut-of-DistributionandIn-DistributionGeneralizationJohnMiller1RohanTaori2AditiRaghunathan2ShioriSagawa2PangWeiKoh2VaishaalShankar1PercyLiang2YairCarmo...
AFunctionalPerspectiveonLearningSymmetricFunctionswithNeuralNetworksAaronZweig1JoanBruna12AbstractSuchsymmetricfunctionsappearnaturallyacrossseveraldomains,includingparticlephysics,computergraphics...
AStatisticalPerspectiveonDistillationAdityaKrishnaMenon1AnkitSinghRawat1SashankJ.Reddi1SeungyeonKim1SanjivKumar1Abstractetal.,2019).onecommonlyacceptedintuitionfromHintonetal.(2015)isthattheteacher...
ARepresentationLearningPerspectiveontheImportanceofTrain-ValidationSplittinginMeta-LearningNikunjSaunshi1ArushiGupta1WeiHu1Abstractfrommany“train”taskstolearnausefulpriorthatcanhelpsolvenew“test...
UnderstandingtheImpactofModelIncoherenceonConvergenceofIncrementalSGDwithRandomReshuffeShaocongMa1YiZhou1Abstractthesamplelossofthei-thdatasample.Suchaproblemformulationcoversavarietyofmachinelearn...
UnderstandingContrastiveRepresentationLearningthroughAlignmentandUniformityontheHypersphereTongzhouWang1PhillipIsola1AbstractContrastiverepresentationlearninghasbeenout-Alignment:Similarsampleshave...
TheEffectofNaturalDistributionShiftonQuestionAnsweringModelsJohnMiller1KarlKrauth1BenjaminRecht1LudwigSchmidt1AbstractmerelytoobtainhighscoresontheSQuADleaderboard,butrathertogeneralizetonewexample...
TensordenoisingandcompletionbasedonordinalobservationsChanwooLee1MiaoyanWang1Abstracttensor,therebyefficientlyreducingtheintrinsicdimensioninbothproblems.Higher-ordertensorsarisefrequentlyinapplica...
TeachingwithLimitedInformationontheLearner’sBehaviourFerdinandoCicalese1SergioFilho2EduardoLaber2MarcoMolinaro2Abstractbeenontheinteractivesetting(Liuetal.,2017;Chenetal.,2018;Liuetal.,2018;Dasgup...
StochasticFlowsandGeometricOptimizationontheOrthogonalGroupKrzysztofChoromanski12DavidCheikhi3JaredDavis1ValeriiLikhosherstov4AchilleNazaret3AchrafBahamou2XingyouSong1MrugankAkarte2JackParker-Holde...
onVariationalLearningofControllableRepresentationsforTextwithoutSupervisionPengXu1JackieChiKitCheung123YanshuaiCao1Abstracttexts.Recently,VAEsandotherunsupervisedgenerativemodelshavefoundsuccessesi...
onUnbalancedOptimalTransport:AnAnalysisofSinkhornAlgorithmKhiemPham1KhangLe1NhatHo2TungPham13HungBui1Abstractogy(Schiebingeretal.,2019),computationalimaging(Leeetal.,2019),deeplearning(Yang&Uhler,2...
onThompsonSamplingwithLangevinAlgorithmsEricMazumdar1AldoPacchiano1Yi-AnMa23PeterL.Bartlett14MichaelI.Jordan14Abstractexploitationtradeoffs(Aueretal.,2002;LattimoreandSzepesva´ri,2020),whereinanal...
onValidationandPlanningofAnOptimalDecisionRulewithApplicationinHealthcareStudiesHengruiCai1WenbinLu1RuiSong1Abstractsionforallindividuals.Anumberofmethodshavebeendevelopedforestimatingoptimaldecisi...
ontheTheoreticalPropertiesoftheNetworkJackknifeQiaohuiLin1RobertLunde1PurnamritaSarkar1Abstractetal.,2017;Gonenetal.,2010;Kallaugheretal.,2019).However,comparativelylittleattentionhasbeenpaidtoWest...
ontheUnreasonableEffectivenessoftheGreedyAlgorithm:GreedyAdaptstoSharpnessSebastianPokutta1MohitSingh2AlfredoTorrico3Abstract(Kempeetal.,2003).Giventheimportanceofsubmodularoptimization,therehasbee...
ontheSampleComplexityofAdversarialMulti-SourcePACLearningNikolaKonstantinov1EliasFrantar12DanAlistarh1ChristophH.Lampert1Abstractetal.,2019).Robustnessattrainingtime,however,isrepre-sentedlesspromi...
ontheNumberofLinearRegionsofConvolutionalNeuralNetworksHuanXiong1LeiHuang2MengyangYu2LiLiu2FanZhu2LingShao12AbstractoneexplanationforthesuperiorityofNNsistheirpow-erfulexpressivity,i.e.,theycanrepr...
onthePowerofCompressedSensingwithGenerativeModelsAkshayKamath1SushrutKarmalkar1EricPrice1Abstractofthecompressedrepresentationofxthantoitsambientdimensionn.Thegoalofcompressedsensingistolearnastruc...