GuaranteesforGreedyMaximizationofNon-submodularFunctionswithApplicationsAndrewAnBian1JoachimM.Buhmann1AndreasKrause1SebastianTschiatschek1AbstractwhereV={v1,...,vn}isthegroundset.Specifically,inexp...
GlobaloptimizationofLipschitzfunctionsCe´dricMalherbe1NicolasVayatis1Abstractglobaloptimization(Pinte´r,1991),black-boxoptimization(Jonesetal.,1998)orderivative-freeoptimization(Rios&Thegoalofthe...
GeometryofNeuralNetworkLossSurfacesviaRandomMatrixTheoryJeffreyPennington1YasamanBahri1Abstract2014;Choromanskaetal.,2015;Neyshaburetal.,2015),architecturedesign,andgeneralization(Keskaretal.).Unde...
FailuresofGradient-BasedDeepLearningShaiShalev-Shwartz1OhadShamir2ShakedShammah1Abstractbothparties:fromapractitioner’sperspective,emphasizingthedifficultiesprovidespracticalinsightstothetheoreti-...
EvaluatingtheVarianceofLikelihood-RatioGradientEstimatorsSeiyaTokui12IsseiSato32Abstractforevariancereductioniscrucialforpracticallearning.However,fewthingsareknownaboutitstheoreticalas-Thelikeliho...
EfficientOrthogonalParametrisationofRecurrentNeuralNetworksUsingHouseholderReflectionsZakariaMhammedi12AndrewHellicar2AshfaqurRahman2JamesBailey1Abstractinanerrorsurface,associatedwithsomeobjective...
DifferentiallyPrivateLearningofUndirectedGraphicalModelsUsingCollectiveGraphicalModelsGarrettBernstein1RyanMcKenna1TaoSun1DanielSheldon12MichaelHay3GeromeMiklau1AbstractDifferentialprivacyisawidely...
Cost-OptimalLearningofCausalGraphsMuratKocaoglu1AlexDimakis1SriramVishwanath1AbstractSupposeshedecidestoperformaninterventiononthedietvariable.ThisentailsforcingthedesireddietaryrestrictionsWeconsi...
ConvergenceAnalysisofProximalGradientwithMomentumforNonconvexOptimizationQunweiLi1YiZhou1YingbinLiang1PramodK.Varshney1AbstractAlgorithm1APGInthiswork,weinvestigatetheacceleratedprox-Input:y1=x1=x0...
ClusteringbySumofNorms:StochasticIncrementalAlgorithm,ConvergenceandClusterRecoveryAshkanPanahi1DevdattDubhashi2FredrikD.Johansson3ChiranjibBhattacharyya4Abstractrequiresrandomlyinitializingtheclus...
BayesianModelsofDataStreamswithHierarchicalPowerPriorsAndre´sMasegosa12ThomasD.Nielsen3HelgeLangseth2Dar´ıoRamos-Lo´pez1AntonioSalmero´n1AndersL.Madsen34Abstractgettingapproaches(Honkela&Valpo...
AutomaticDiscoveryoftheStatisticalTypesofVariablesinaDatasetIsabelValera1ZoubinGhahramani12Abstractplore,findpatternsormakepredictionsonthedata.Asanexample,apredictiontaskissolveddifferentlydependi...
AnalyticalGuaranteesonNumericalPrecisionofDeepNeuralNetworksCharbelSakrYongjuneKimNareshShanbhagAbstractdeepresidualnetwork.ThishighcomplexityofdeepneuralnetworkspreventsitsdeploymentonenergyandThe...
AnalysisandOptimizationofGraphDecompositionsbyLiftedMulticutsAndreaHornˇa´kova´1Jan-HendrikLange1BjoernAndres1AbstractFigure1.Adecompositionofagraphisapartitionofthenodesetintoconnectedsubsets.A...
AnAnalyticalFormulaofPopulationGradientfortwo-layeredReLUnetworkanditsApplicationsinConvergenceandCriticalPointAnalysisYuandongTian1Abstractandwhysimplemethodslikegradientdescentcansolvethecomplica...
AnAdaptiveTestofIndependencewithAnalyticKernelEmbeddingsWittawatJitkrittum1ZoltánSzabó2ArthurGretton1AbstractAbasicnonlineardependencemeasureistheHilbert-SchmidtIndependenceCriterion(HSIC),whichi...
AdaNet:AdaptiveStructuralLearningofArtificialNeuralNetworksCorinnaCortes1XavierGonzalvo1VitalyKuznetsov1MehryarMohri21ScottYang2Abstractoflayersandunitsspecifiedsincethereneedstobeatleastonepaththr...