RevisitingPointCloudShapeClassificationwithaSimpleandEffectiveBaselineAnkitGoyal1HeiLaw1BoweiLiu1AlejandroNewell1JiaDeng1AbstractFigure1.PerformanceofdifferentmodelsonModelNet40.Thoseusing>1024poin...
OntheImplicitBiasofInitializationShape:BeyondInfinitesimalMirrorDescentShaharAzulay1EdwardMoroshko2MorShpigelNacson2BlakeWoodworth3NathanSrebro3AmirGloberson1DanielSoudry2Abstractparameterizedmodel...
MultidimensionalShapeConstraintsMayaR.Gupta1ErezLouidor1OlexanderMangylov1NobuyukiMorioka1TamanNarayan1SenZhao1AbstractShapeconstraintshistoricallyhavebeendefinedintermsofhowthefunctionrespondstoch...
ShapeConstraintsforSetFunctionsAndrewCotter1MayaR.Gupta1HeinrichJiang1ErezLouidor1JamesMuller1TamanNarayan1SerenaWang1TaoZhu1AbstractForexample,supposeonewantstopredicthowgoodarestaurantisbasedonit...
GeodesicConvolutionalShapeOptimizationPierreBaque1EdoardoRemelli1Franc¸oisFleuret21PascalFua1AbstractXtrainFωy(Xtrain)Fωz(Xtrain)AerodynamicShapeoptimizationhasmanyindus-(a)GCNNFωy(Xtest)Fωz(X...
CognitivePsychologyforDeepNeuralNetworks:AShapeBiasCaseStudySamuelRitter1DavidG.T.Barrett1AdamSantoro1MattM.Botvinick1Abstracttothinkofthesemodelsasblackboxes,andtoquestionwhethertheycanbeunderstoo...