UnsupervisedPartRepresentationbyFlowCapsulesSaraSabour12AndreaTagliasacchi12SorooshYazdani1GeoffreyE.Hinton12DavidJ.Fleet12AbstractFigure1:Self-supervisedtrainingforlearningprimarycapsules:Animagee...
StabilizingEquilibriumModelsbyJacobianRegularizationShaojieBai1VladlenKoltun2J.ZicoKolter1Abstractequilibriumzbytheimplicitfunctiontheorem(Krantz&Parks,2012),irrespectiveofthemethodusedtosolveforDe...
SketchEmbedNet:LearningNovelConceptsbyImitatingDrawingsAlexanderWang12MengyeRen12RichardS.Zemel123Abstractofsketchgenerationmodels,rangingfromgenerativeadver-sarialnetworks(GANs)(Isolaetal.,2017;Li...
SequentialDomainAdaptationbySynthesizingDistributionallyRobustExpertsBaharTaskesen1Man-ChungYue2Jose´Blanchet3DanielKuhn1VietAnhNguyen34Abstractabsenceofmeaningfuldatatotuneanyparameters.How-ever,...
OvercomingCatastrophicForgettingbyBayesianGenerativeRegularizationPatrickH.Chen1WeiWei2Cho-jui,Hsieh1BoDai3Abstractelsfailtotacklecontinuallearningwiththephenomenonofcatastrophicforgetting,wheredee...
Meta-Cal:Well-controlledPost-hocCalibrationbyRankingXingchenMa1MatthewB.Blaschko1AbstractToachievethisobjective,onecandesignmodelswithin-trinsicallylowcalibrationerrors.Thesemodels(WilsonInmanyappl...
MakingtransportmorerobustandinterpretablebymovingdatathroughasmallnumberofanchorpointsChi-HengLin1MehdiAzabou12EvaL.Dyer123Abstracttivemodeling(MartinArjovsky&Bottou,2017;Tolstikhinetal.,2017),docu...
LearningRepresentationsbyHumans,forHumansSophieHilgard1NirRosenfeld2MahzarinBanaji3JackCao3DavidC.Parkes1Abstractetal.,2017)andwhereusershaveexternalinformationorwanttoexerciseagencyandusetheirownj...
LearningBinaryDecisionTreesbyArgminDifferentiationValentinaZantedeschi12MattJ.Kusner2VladNiculae3Abstractquadraticprogram(relaxedfromMIP)Weaddresstheproblemoflearningbinarydeci-siontreesthatpartiti...
LearningbyTurning:NeuralArchitectureAwareOptimisationYangLiu1JeremyBernstein2MarkusMeister2YisongYue2AbstractThisobservationmotivatesthecombinedstudyofarchi-tectureandoptimisation,andthispaperexplo...
ImitationbyPredictingObservationsAndrewJaegle1YurySulsky1ArunAhuja1JakeBruce1RobFergus1GregWayne1Abstract2009;Huberetal.,2009).Whilemostalgorithmsforimita-tionlearningassumethatdemonstrationscontai...
GraphNeuralNetworksInspiredbyClassicalIterativeAlgorithmsYongyiYang1†TangLiu1†YangkunWang2†JinjingZhou3QuanGan3ZheweiWei4ZhengZhang3ZengfengHuang1DavidWipf3Abstract1.IntroductionDespitetherecent...
FILTRA:RethinkingSteerableCNNbyFilterTransformBoLi1QiliWang1GimHeeLee2Abstractputishard-bakedtotransformaccordinglywhentheinputreflectsorrotates.AplentyofworksdevelopthisideaSteerableCNNimposesthep...
Few-shotLanguageCoordinationbyModelingTheoryofMindHaoZhu1GrahamNeubig1YonatanBisk1Abstractprocess,withawidevarietyofworksexaminingcommuni-cationbetweenagentsviaeithercompletelyartificialemer-Nomani...
CombOptNet:FittheRightNP-HardProblembyLearningIntegerProgrammingConstraintsAnselmPaulus1MichalRol´ınek1V´ıtMusil2BrandonAmos3GeorgMartius1SoILlvPerFigure1:CombOptNetasamoduleinadeeparchitecture...
AutoencoderImageInterpolationbyShapingtheLatentSpaceAlonOring1ZoharYakhini1YacovHel-Or1Abstractexistingobjects,cameraparametersandadditionalscenecomponents,suchaslightingandthereflectivecharacteris...
BORE:BayesianOptimizationbyDensity-RatioEstimationLouisC.Tiao12AaronKlein3MatthiasSeeger3EdwinV.Bonilla21Ce´dricArchambeau3FabioRamos14Abstractdensity1(x)Bayesianoptimization(BO)isamongthemostg(x)...
TrainingNeuralNetworksforandbyInterpolationLeonardBerrada1AndrewZisserman2M.PawanKumar2Abstractdimensionalnon-convexfunctions.Inpractice,themainalgorithmsofchoiceareStochasticGradientDescent(SGD)In...
ReinforcementLearningforMolecularDesignGuidedbyQuantumMechanicsGregorN.C.Simm1RobertPinsler1Jose´MiguelHerna´ndez-Lobato1AbstractFigure1.Visualizationofthemoleculardesignprocesspresentedinthiswor...
PrivateQueryReleaseAssistedbyPublicDataRaefBassily1AlbertCheu2ShayMoran3AleksandarNikolov4JonathanUllman2ZhiweiStevenWu5Abstracterrorα.Acentralquestioninthisproblemisconcernedwithcharacterizingthe...