LearningandSamplingofAtomicInterventionsfromObservationsArnabBhattacharyya1SutanuGayen1SaravananKandasamy2AshwinMaran3N.V.Vinodchandran4Abstracttime.WestudytheproblemofefficientlyestimatingWeextend...
LearningandEvaluatingContextualEmbeddingofSourceCodeAdityaKanade12PetrosManiatis2GogulBalakrishnan2KensenShi2Abstractstandtheauthor’sintentsothattheycanmaintainandextendthecode.Developersusemeanin...
LandscapeConnectivityandDropoutStabilityofSGDSolutionsforOver-parameterizedNeuralNetworksAlexanderShevchenko1MarcoMondelli1Abstractanover-parameterizednetworkviaSGDtypicallyleadstoasolutionthathass...
KernelizedSteinDiscrepancyTestsofGoodness-of-fitforTime-to-EventDataTamaraFerna´ndez1ArthurGretton1Nicola´sRivera2WenkaiXu1Abstracttimeofapatientinaclinicaltrial(Collett,2015;Biswasetal.,2007);or...
ImprovingtheGatingMechanismofRecurrentNeuralNetworksAlbertGu1CaglarGulcehre2TomPaine2MattHoffman2RazvanPascanu2Abstract(Chungetal.,2014,GRUs).ThesegatedRNNshavebeenverysuccessfulinseveraldifferenta...
ImprovingRobustnessofDeep-Learning-BasedImageReconstructionAnkitRaj1YoramBresler1BoLi2Abstract1.IntroductionDeep-learning-basedmethodsforvariousappli-Adversarialexamplesfordeeplearningbasedmethodsh...
ImplicitRegularizationofRandomFeatureModelsArthurJacot1BerfinS¸ims¸ek12FrancescoSpadaro1Cle´mentHongler1FranckGabriel1Abstractparametersexhibitsaso-called‘double-descent’curveformanymodelsincl...
ImplicitdifferentiationofLasso-typemodelsforhyperparameteroptimizationQuentinBertrand1QuentinKlopfenstein2MathieuBlondel3SamuelVaiter4AlexandreGramfort1JosephSalmon5Abstract1.IntroductionSettingreg...
HierarchicalGenerationofMolecularGraphsusingStructuralMotifsWengongJin1ReginaBarzilay1TommiJaakkola1Abstractgradientsacrosstheiterativesteps.Graphgenerationtechniquesareincreasinglybe-Largemolecule...
HarmonicDecompositionsofConvolutionalNetworksMeyerScetbon12ZaidHarchaoui2Abstractlyzeamulti-layerperceptron(MLP)insteadactingonvec-torizedimages,whichhasthedownsidethesetasidetheWepresentadescripti...
GuidedLearningofNonconvexModelsthroughSuccessiveFunctionalGradientOptimizationRieJohnson1TongZhang2AbstractOuroriginalmotivationwasfunctionalgradientlearningofadditivemodelsingradientboosting(Fried...
GraphOpt:LearningOptimizationModelsofGraphFormationRakshitTrivedi1JiachenYang1HongyuanZha12AbstractFigure1.⌦issetoflatentobjectivefunctions{Fi},anyofwhichcouldleadtoobservedgraphGwhenoptimised.Our...
GraphStructureofNeuralNetworksJiaxuanYou1JureLeskovec1KaimingHe2SainingXie2Abstractinformationfromitsinputneurons,throughhiddenlayersallthewaytotheoutputneurons(McClellandetal.,1986).Neuralnetworks...
GradientDICE:RethinkingGeneralizedofflineEstimationofStationaryValuesShangtongZhang1BoLiu2ShimonWhiteson1Abstractevaluationismoreflexible.Wecanevaluateanewpolicywithexistingdatainareplaybuffer(Lin,...
GoWide,ThenNarrow:EfficientTrainingofDeepThinNetworksDennyZhou1MaoYe2ChenChen1TianjianMeng1MingxingTan1XiaodanSong1QuocLe1QiangLiu2DaleSchuurmans1Abstractfeletal.,2019;Brownetal.,2020).Toenlargeamo...
GlobalConcavityandOptimizationinaClassofDynamicDiscreteChoiceModelsYidingFeng1EkaterinaKhmelnitskaya2DenisNekipelov2Abstract1.IntroductionDiscretechoicemodelswithunobservedhetero-Dynamicdiscretecho...
GeneralizationErrorofGeneralizedLinearModelsinHighDimensionsMelikasadatEmami1MojtabaSahraee-Ardakan12ParthePandit12SundeepRangan3AlysonK.Fletcher12Abstractaclassofgeneralizedlinearmodels(GLMs)ofthe...
GeneralizationandRepresentationalLimitsofGraphNeuralNetworksVikasK.Garg1StefanieJegelka1TommiJaakkola1Abstractblycombiningthiswithv’sembedding.Theexactformoftheaggregateandcombinestepsvariesacross...
GamificationofPureExplorationforLinearBanditsRe´myDegenne1PierreMe´nard2XuedongShang3MichalValko4Abstracthighconfidencetoagivenqueryusingasfewsamplesaspossible.Weinvestigateanactivepure-explorati...
FullLawIdentificationinGraphicalModelsofMissingData:CompletenessResultsRaziehNabi1RohitBhattacharya1IlyaShpitser1Abstractetal.,2008;Marstonetal.,2010).MNARmechanismsareexpectedtooccurquiteofteninpr...