JoiningdatasetsviadataaugmentationinthelabelspaceforneuralnetworksJakeZhao(Junbo)1MingfengOu23LinjiXue2YunkaiCui2SaiWu1GangChen1AbstractInstead,itsversatilitycanbesubstantiallyenhancedbyanovelframe...
ImprovedConfidenceBoundsfortheLinearLogisticModelandApplicationstoBanditsKwang-SungJun1LalitJain2BlakeMason3HoussamNassif4Abstractafeaturevectorforeacharm.Acommonassumptionisthattherewardisanoisyli...
HyperHyperNetworksfortheDesignofAntennaArraysShaharLutati1LiorWolf12AbstractDesigningantennasisachallenginginverseproblem:whilemappingthestructureoftheantennatoitsradiationprop-Wepresentdeeplearnin...
HowImportantistheTrain-ValidationSplitinMeta-Learning?YuBai1MinshuoChen2PanZhou1TuoZhao2JasonD.Lee3ShamKakade4HuanWang1CaimingXiong1Abstract1.IntroductionMeta-learningaimstoperformfastadaptationMet...
HeterogeneityfortheWin:One-ShotFederatedClusteringDonKurianDennis1TianLi1VirginiaSmith1AbstractInthecaseoffederatedlearning,clusteringhasfoundap-plicationsinclient-selection(Choetal.,2020),personal...
GuaranteesforTuningtheStepSizeusingaLearning-to-LearnApproachXiangWang1ShuaiYuan1ChenweiWu1RongGe1Abstractrinetal.(2015)consideredtheideaoftuningtheseparame-tersbyoptimization—thatis,considerameta...
GeometryoftheLossLandscapeinOverparameterizedNeuralNetworks:SymmetriesandInvariancesBerfinS¸ims¸ek12Franc¸oisGed1ArthurJacot1FrancescoSpadaro1Cle´mentHongler1WulframGerstner2JohanniBrea2Abstrac...
GenerativeVideoTransformer:CanObjectsbetheWords?Yi-FuWu1JaesikYoon12SungjinAhn13Abstractinteresttodevelopananalogousgenerativepre-trainingpro-cedureforvideos,thecomputationaloverheadindealingTransf...
GeneralizationBoundsinthePresenceofOutliers:aMedian-of-MeansStudyPierreLaforgue1GuillaumeStaerman2StephanClémençon2Abstractcorruptionoftheobservationsacquired.Fromastatisticalperspective,twoframe...
FindingtheStochasticShortestPathwithLowRegret:theAdversarialCostandUnknownTransitionCaseLiyuChen1HaipengLuo1Abstractendwithinafixednumberofstepsisextensivelystudiedinrecentyears(oftenknownasepisodi...
EvaluatingtheImplicitMidpointIntegratorforRiemannianManifoldHamiltonianMonteCarloJamesA.Brofos1RoyR.Lederman1AbstractFigure1.Trajectoriescomputedbytheleapfrogintegratorandtheimplicitmidpointintegra...
EquivariantMessagePassingforthePredictionofTensorialPropertiesandMolecularSpectraKristofT.Schu¨tt12OliverT.Unke12MichaelGastegger13Abstractoryquicklybecomeprohibitiveforlargesystemsandthepredictio...
DisentanglingSyntaxandSemanticsintheBrainwithDeepNetworksCharlotteCaucheteux12AlexandreGramfort1Jean-RemiKing23AbstractA.B.theactivationsoflanguagetransformerslikeLexicalCompositionalGPT-2havebeens...
DimensionalityReductionforSum-of-DistancesMetricZhiliFeng1PraneethKacham1DavidP.Woodruff1Abstractsubspaceandperformthetaskonthelow-dimensionalpoints.GivenasetofnpointsA={a1,...,an}inRd,Wegiveadimen...
DifferentiallyPrivateAggregationintheShuffleModel:AlmostCentralAccuracyinAlmostaSingleMessageBadihGhazi1RaviKumar1PasinManurangsi1RasmusPagh21AmerSinha3Abstractwithextensionsincluding(private)distr...
DFACFramework:FactorizingtheValueFunctionviaQuantileMixtureforMulti-AgentDistributionalQ-LearningWei-FangSun123Cheng-KuangLee2Chun-YiLee1Abstractoptimizetheoverallrewardsineachepisode.Nevertheless,...
DetectionofSignalintheSpikedRectangularModelsJiHyungJung1HyeWonChung2JiOonLee13Abstractv∈RNwithu=v=1.theparameterλcorrespondstothesignal-to-noiseratio(SNR).Weconsidertheproblemofdetectingsignalsi...
DenseforthePriceofSparse:ImprovedPerformanceofSparselyInitializedNetworksviaaSubspaceOffsetIlanPrice12JaredTanner12Abstract1.1.CompetingPrioritiesforSparseNetworksThatneuralnetworksmaybeprunedtohig...
DecentralizedRiemannianGradientDescentontheStiefelManifoldShixiangChen1AlfredoGarcia1MingyiHong2ShahinShahrampour1AbstractagentoptimizationproblemWeconsiderdistributednon-convexoptimization1nwherea...
CrystallizationLearningwiththeDelaunayTriangulationJiaqiGu1GuoshengYin1AbstractandE(2i)<∞.Nonparametricregressionisacollectionofmethodsforestimatingtheconditionalexpectationfunc-BasedontheDelaunay...