RiemannianConvexPotentialMapsSamuelCohen1BrandonAmos2YaronLipman23AbstractFigure1.Illustrationofadiscretec-concavefunction(blue)overabasemanifoldM(boldline).Theseconsistofdiscretecompo-Modelingdist...
DANCE:EnhancingsaliencyMapsusingdecoysYangYoungLu1WenboGuo2XinyuXing2WilliamStaffordNoble3Abstract1.IntroductionSaliencymethodscanmakedeepneuralnetworkDeepneuralnetworks(DNNs)deliverremarkableperfo...
Multi-FrequencyVectorDiffusionMapsYifengFan1ZhizhenZhao1Abstractkernelforvectorfieldsonthedatamanifold.Thecorre-spondingadjacencymatrixisbasedonedgeweightsandWeintroducemulti-frequencyvectordiffusi...
OntheSpectrumofRandomFeaturesMapsofHighDimensionalDataZhenyuLiao1RomainCouillet12Abstractfeaturesrelevanttosomegiventask.ThenonlinearitiesmaketheserepresentationsmoremightybutmeanwhileRandomfeature...
SphericalStructuredFeatureMapsforKernelApproximationYuemingLyu1Abstractdenotesthenonlinearfunction,W∈Rd×NisconstructedbyNi.i.dsamplesdrawnfromadistributiondefinedbyWeproposeSphericalStructuredFea...