FeatureQuantizationImprovesGANTrainingYangZhao1ChunyuanLi2PingYu1JianfengGao2ChangyouChen1AbstractTrainingGANsisanotoriouslychallengingtask,asitin-volvesoptimizinganon-convexproblemforitsNashequi-T...
FeatureSelectionusingStochasticGatesYutaroYamada1OfirLindenbaum2SahandNegahban1YuvalKluger23Abstract(e.g.,genesorproteins)oftenexceedsthenumberofsam-ples.Inthissetting,Featureselectioncanleadtoimpr...
FeatureNoiseInducesLossDiscrepancyAcrossGroupsFereshteKhani1PercyLiang1Abstractdiscrepancy(Liptonetal.,2018)andadverselong-termimpacts(Liuetal.,2018).Therefore,itisimportanttoTheperformanceofstanda...
EfficientnonparametricstatisticalinferenceonpopulationFeatureimportanceusingShapleyvaluesBrianD.Williamson1JeanFeng2AbstractLikewise,vaccineresearchersmaywanttoknowthemostimportantmolecularphenotyp...
DivideandConquer:LeveragingIntermediateFeatureRepresentationsforQuantizedTrainingofNeuralNetworksAhmedT.Elthakeb1PrannoyPilligundla2FatemehsadatMireshghallah2AlexanderCloninger3HadiEsmaeilzadeh2Abs...
BeyondSignalPropagation:IsFeatureDiversityNecessaryinDeepNeuralNetworkInitialization?YanivBlumenfeld1DarGilboa2DanielSoudry1Abstractadetailedanalysisoftheirdynamicsandgeneralizationproperties.Unlik...
FeatureGroupingasaStochasticRegularizerforHigh-DimensionalStructuredDataSergu¨lAydo¨re1BertrandThirion2Gae¨lVaroquaux2Abstract1.LargeFeaturedimension:Neuroimagingdataareveryhigh-dimensional,duet...
ConcreteAutoencoders:DifferentiableFeatureSelectionandReconstructionAbubakarAbid1MuhammedFatihBalin2JamesZou134AbstracttoreducedimensionalityofdatabyidentifyingthesubsetofrelevantFeaturesinadataset...
CategoricalFeatureCompressionviaSubmodularOptimizationMohammadHosseinBateni1LinChen12HosseinEsfandiari1ThomasFu1VahabS.Mirrokni1AfshinRostamizadeh1Abstractisastandardoccurrence.Forexample,intaskssu...
MSplitLBI:RealizingFeatureSelectionandDenseEstimationSimultaneouslyinFew-shotandZero-shotLearningBoZhao†1XinweiSun†2YanweiFu‡34YuanYao‡5YizhouWang1Abstracting(FSL)andcapturingrelationalstructur...
MISSION:UltraLarge-ScaleFeatureSelectionusingCount-SketchesAmiraliAghazadeh1RyanSpring2DanielLeJeune3GautamDasarathy3AnshumaliShrivastava2RichardG.Baraniuk3Abstractderedpairs(Xi,yi)i∈[n],whereXi∈...
InterpretabilityBeyondFeatureAttribution:QuantitativeTestingwithConceptActivationVectors(TCAV)BeenKimMartinWattenbergJustinGilmerCarrieCaiJamesWexlerFernandaViegasRorySayresAbstractcoefficientweigh...
DeepAsymmetricMulti-taskFeatureLearningHaeBeomLee12EunhoYang32SungJuHwang32Abstracttivetransferproblembyallowingasymmetricknowledgetransferbetweentasksthroughinter-taskparameterregular-WeproposeDee...
ABirth-DeathProcessforFeatureAllocationKonstantinaPalla1DavidKnowles2ZoubinGhahramani34Abstract2.RelatedWorkWeproposeaBayesiannonparametricprioroverWetargetproblemswherethedatadependsonacovariate,f...
SphericalStructuredFeatureMapsforKernelApproximationYuemingLyu1Abstractdenotesthenonlinearfunction,W∈Rd×NisconstructedbyNi.i.dsamplesdrawnfromadistributiondefinedbyWeproposeSphericalStructuredFea...
RegularisingNon-linearModelsUsingFeatureSide-informationAminaMollaysa1PabloStrasser1AlexandrosKalousis1Abstractformationforthelearninginstances;thisinformationisnotavailableintesting.Inmetriclearni...
RandomFeatureExpansionsforDeepGaussianProcessesKurtCutajar1EdwinV.Bonilla2PietroMichiardi1MaurizioFilippone1Abstract2006)suchthattheoutputofeachlayerofGPsformstheinputtotheGPsatthenextlayer,effecti...
McGan:MeanandCovarianceFeatureMatchingGANYoussefMroueh12TomSercu12VaibhavaGoel2Abstractetal.,2015;Dziugaiteetal.,2015).Asshownempiricallyin(Salimansetal.,2016),onecantraintheGANdiscrimi-Weintroduce...
LatentFeatureLassoIanE.H.Yen1Wei-ChengLee2Sung-EnChang2ArunS.Suggala1Shou-DeLin2PradeepRavikumar1Abstractnationoflatentclasses.Forinstance,adocumentobjectcouldbemodeledasaconvexcombinationoftopicob...
BeyondFilters:CompactFeatureMapforPortableDeepModelYunheWang1ChangXu2ChaoXu1DachengTao2Abstractsignificantredundancyinmostofexistingconvolutionalneuralnetworks.Forinstance,theResNet-50(Heetal.,Conv...