VariationalInferenceforSequentialDatawithFutureLikelihoodEstimatesGeon-HyeongKim1YoungsooJang1HongseokYang1Kee-EungKim12Abstractdiscretelatentvariables),learningandanalyzingthemodelsforhigh-dimensi...
VariationalAutoencoderswithRiemannianBrownianMotionPriorsDimitrisKalatzis1DavidEklund2GeorgiosArvanitidis3SørenHauberg1AbstractFigure1.ThelatentspacepriorsoftwoVAEstrainedonthedigit1fromMNIST.Left...
VarianceReducedCoordinateDescentwithAcceleration:NewMethodwithaSurprisingApplicationtoFinite-SumProblemsFilipHanzely1DmitryKovalev1PeterRichta´rik1Abstractcontrast,ifψisnotseparable,thecorrespond...
TransparencyPromotionwithModel-AgnosticLinearCompetitorsHassanRafique1TongWang2QihangLin2ArshiaSighani3Abstractderstandandrationalize.Drivenbythepracticalneeds,researchershaveshiftedtheirfocustoacc...
TransformersareRNNs:FastAutoregressiveTransformerswithLinearAttentionAngelosKatharopoulos12ApoorvVyas12NikolaosPappas3Franc¸oisFleuret12Abstractbytheglobalreceptivefieldofself-attention,whichpro-c...
TrainingDeepEnergy-BasedModelswithf-DivergenceMinimizationLantaoYu1YangSong1JiamingSong1StefanoErmon1Abstractbasedmodels(Dinhetal.,2014;2016;Kingma&Dhariwal,2018)andsum-productnetworks(Poon&Domingo...
TrainingBinaryNeuralNetworksthroughLearningwithNoisySupervisionKaiHan12YunheWang2YixingXu2ChunjingXu2EnhuaWu13ChangXu4AbstractFigure1.Frameworkoflearningbinaryneuronswithnoisysuper-vision.Anetworkf...
TheComplexityofFindingStationaryPointswithStochasticGradientDescentYoelDrori1OhadShamir12Abstractisnottominimizef(x)overx,butrather∇f(x).Thisquestionoffindingstationarypointshasgainedmoreatten-Wes...
TeachingwithLimitedInformationontheLearner’sBehaviourFerdinandoCicalese1SergioFilho2EduardoLaber2MarcoMolinaro2Abstractbeenontheinteractivesetting(Liuetal.,2017;Chenetal.,2018;Liuetal.,2018;Dasgup...
StructuredPredictionwithPartialLabellingthroughtheInfimumLossVivienCabannes1AlessandroRudi1FrancisBach1AbstractPartiallabellinghasbeenstudiedinthecontextofclassifi-cation(Couretal.,2011;Nguyen&Caru...
StudentSpecializationinDeepRectifiedNetworkswithFiniteWidthandInputDimensionYuandongTian1Abstract1.IntroductionWeconsideradeepReLU/LeakyReLUstu-WhileDeepLearninghasachievedgreatsuccessindiffer-dent...
StochasticDifferentialEquationswithVariationalWishartDiffusionsMartinJørgensen1MarcPeterDeisenroth2HughSalimbeni3AbstractWhymodeltheprocessnoise?Assumethatintheexampleabove,thetwostatesrepresentme...
Stochasticbanditswitharm-dependentdelaysAnneGaelManegueu1ClaireVernade2AlexandraCarpentier1MichalValko3AbstractAsaresult,westudystochasticdelayedbanditsforwhichthedelaydistributionsarearm-dependent...
StochasticCoordinateMinimizationwithProgressivePrecisionforStochasticConvexOptimizationSudeepSalgia1QingZhao1SattarVakili2Abstractknown,theexpectationofF(x;ξ)overξcannotbeanalyti-callycharacteriz...
SpectralClusteringwithGraphNeuralNetworksforGraphPoolingFilippoMariaBianchi1DanieleGrattarola2CesareAlippi23AbstractMessage-passingMinCutPoolMessage-passingSpectralclustering(SC)isapopularclusterin...
SparseSubspaceClusteringwithEntropy-NormLiangBai1JiyeLiang1Abstracthigh-dimensionaldatawhichisubiquitousinreal-worlddataminingapplications,suchasimageprocessing,textInthispaper,weprovideanexplicitt...
SparseGaussianProcesseswithSphericalHarmonicFeaturesVincentDutordoir1NicolasDurrande1JamesHensman2AbstractyWeintroduceanewclassofinter-domainvari-biasationalGaussianprocesses(GP)wheredataismappedon...
SourceSeparationwithDeepGenerativePriorsVivekJayaramJohnThickstunAbstractwithnofurtherconstraintsorregularization,solvingEqua-tion(1)forxishighlyunderdetermined.Classical“blind”Despitesubstantial...
SIGUA:ForgettingMayMakeLearningwithNoisyLabelsMoreRobustBoHan12GangNiu2XingruiYu3QuanmingYao4MiaoXu25IvorW.Tsang3MasashiSugiyama26Abstractasweightdecay(Krogh&Hertz,1991)anddropout(Sri-vastavaetal.,...
SequentialTransferinReinforcementLearningwithaGenerativeModelAndreaTirinzoni1RiccardoPoiani1MarcelloRestelli1AbstractAkeyquestioniswhatandhowknowledgeshouldbetrans-ferred(Taylor&Stone,2009).Asforth...