AccelerateCNNsfromThreeDimensions:AComprehensivePruningFrameworkWenxiaoWang12MinghaoChen1ShuaiZhao1LongChen34JinmingHu1HaifengLiu1DengCai1XiaofeiHe1WeiLiu2Abstract57Mostneuralnetworkpruningmethods,...
AFreeLunchfromANN:TowardsEfficient,AccurateSpikingNeuralNetworksCalibrationYuhangLi12ShikuangDeng1XinDong3RuihaoGong4ShiGu1AbstractFeaturesTrainingConversionCalibrationSpikingNeuralNetwork(SNN)hasb...
TransformationofReLU-basedrecurrentneuralnetworksfromdiscrete-timetocontinuous-timeZahraMonfared1DanielDurstewitz12AbstractMorerecently,inthenaturalsciences,biologyandphysicsinparticular,RNNswereal...
TaskUnderstandingfromConfusingMulti-taskDataXinSu12YizhouJiang1ShangqiGuo13FengChen132AbstractTaskAnnotationLabelAnnotationBeyondmachinelearning’ssuccessinthespecificTaskTAasnkno1tationSampleAnnot...
StrengthfromWeakness:FastLearningUsingWeakSupervisionJoshuaRobinson1StefanieJegelka1SuvritSra1Abstractplexmodelonarelated,largedatatask,andtothenusesthelearnedfeaturesforfinetuningonthesmall-datata...
StrategyproofMeanEstimationfromMultiple-ChoiceQuestionsAnsonKahng1GregoryKehne2ArielD.Procaccia3Abstractadmittoprovideanestimateoftheprobabilitypithattheadmitmatriculates,butthisisproblematicbecaus...
SampleFactory:Egocentric3DControlfromPixelsat100000FPSwithAsynchronousReinforcementLearningAlekseiPetrenko12ZhehuiHuang2TusharKumar2GauravSukhatme2VladlenKoltun1AbstractHwangboetal.,2019;Molchanove...
RecoveryofSparseSignalsfromaMixtureofLinearSamplesAryaMazumdar1SoumyabrataPal1AbstractWewillrefertothevaluesreturnedbytheoraclegiventhesequeriesassamples.Mixtureoflinearregressionsisapopularlearnin...
RankAggregationfromPairwiseComparisonsinthePresenceofAdversarialCorruptionsArpitAgarwal1ShivaniAgarwal1SanjeevKhanna1PrathameshPatil1Abstract1.IntroductionRankaggregationfrompairwisepreferenceshasT...
PrivateCountingfromAnonymousMessages:Near-OptimalAccuracywithVanishingCommunicationOverheadBadihGhazi1RaviKumar1PasinManurangsi1RasmusPagh21Abstractprotectingpotentiallysensitive,user-specificinfor...
PeerLossFunctions:LearningfromNoisyLabelswithoutKnowingNoiseRatesYangLiu1HongyiGuo2Abstracttensivepriorworksintheliteraturethataimtodevelopal-gorithmstolearnmodelsthatarerobusttolabelnoise(By-Learn...
OptimizationfromStructuredSamplesforCoverageFunctionsWeiChen1XiaomingSun23JialinZhang23ZhijieZhang23Abstracteters.Thecomputationaltaskistooptimizeanobjectivefunctiongivenparametersofthemodel.Onesuc...
OptimalRobustLearningofDiscreteDistributionsfromBatchesAyushJain1AlonOrlitsky1AbstractIfanadversarycancorruptaβfractionofthedata,aloss≥βisunavoidable.Fortunately,thatisnotnecessarilyso.Manyappli...
OnlineContinualLearningfromImbalancedDataAristotelisChrysakis1Marie-FrancineMoens1Abstractisavastlydifferentstory.Overourlivesweperceiveastreamoftemporallycorrelated,unlabeledobservations,andAwell-...
Neuro-SymbolicVisualReasoning:Disentangling“Visual”from“Reasoning”SaeedAmizadeh1HamidPalangi2OleksandrPolozov2YichenHuang2KazuhitoKoishida1AbstractRecentadvancesincomputervision,representationl...
NetGANwithoutGAN:fromRandomWalkstoLow-RankApproximationsLucaRendsburg1HolgerHeidrich1UlrikevonLuxburg12Abstractstructure(suchasthecommunitystructure,thecharacteris-ticpathlengths,etc).Beingabletocr...
MinimaxRateforLearningfromPairwiseComparisonsintheBTLModelJulienHendrickx1AlexOlshevsky2VenkateshSaligrama2AbstractAnalternativelineofresearchhasexploreddatafusionbasedonbettercalibratedmeasures,su...
MetaVarianceTransfer:LearningtoAugmentfromTheOthersSeong-JinPark1SeungjuHan1JiwonBaek1InsooKim1JuhwanSong1HaeBeomLee2Jae-JoonHan1SungJuHwang2Abstractcepts.However,collectingacompletedatasetwithallp...
MedianMatrixCompletion:fromEmbarrassmenttoOptimalityWeidongLiu1XiaojunMao2RaymondK.W.Wong3Abstractflixdatahaslessthan1%ofobservedentriesofamatrixwitharound5×105rowsand2×104customers.Withtech-Inth...
LearningStructuredLatentFactorsfromDependentData:AGenerativeModelFrameworkfromInformation-TheoreticPerspectiveRuixiangZhang12MasanoriKoyama3KatsuhikoIshiguro3Abstractbuilt-inratherheuristicwaystoen...