SuperpolynomialLowerboundsforLearningOne-LayerNeuralNetworksusingGradientDescentSurbhiGoel1AravindGollakota1ZhihanJin2SushrutKarmalkar1AdamKlivans1AbstractOurResults.Inthispaperwegivethefirstsuperp...
SharpCompositionboundsforGaussianDifferentialPrivacyviaEdgeworthExpansionQinqingZheng1JinshuoDong1QiLong1WeijieSu1Abstracttheprioralgorithms’output,anadversarycanadaptivelyanalyzethedatasettocause...
SampleComplexityboundsfor1-bitCompressiveSensingandBinaryStableEmbeddingswithGenerativePriorsZhaoqiangLiu1SelwynGomes1AvtanshTiwari2JonathanScarlett1Abstractreticalguaranteesincludingsharpperforman...
RefinedboundsforAlgorithmConfiguration:TheKnife-edgeofDualClassApproximabilityMaria-FlorinaBalcan1TuomasSandholm1234EllenVitercik1AbstractHutteretal.,2009;Kadiogluetal.,2010;Sandholm,2013):givenatr...
OptimisticboundsforMulti-outputPredictionHenryW.J.Reeve1AtaKaba´n1Abstractoftheoutputspace.Whilstmodernapplicationsofmulti-outputpredictiondealwithincreasinglylargedatasets,theyWeinvestigatethecha...
Optimalboundsbetweenf-DivergencesandIntegralProbabilityMetricsRohitAgrawal1ThibautHorel2Abstractdivergencebetweentwodistributionsµandνasanav-eragecostofthelikelihoodratio,thatis,Iφ(µν):=Thefam...
Near-TightMargin-BasedGeneralizationboundsforSupportVectorMachinesAllanGrønlund1LiorKamma1KasperGreenLarsen1AbstractmarginSVM.SupportVectorMachines(SVMs)areamongtheminw2mostfundamentaltoolsforbina...
Near-optimalsamplecomplexityboundsforlearningLatentk−polytopesandapplicationstoAd-MixturesChiranjibBhattacharyya1RavindranKannan2Abstractmixtrues,aproblemnottackledbefore.Ad-mixturesaregeneralizat...
Near-optimalRegretboundsforStochasticShortestPathAlonCohen1HaimKaplan12YishayMansour12AvivRosenberg2AbstractThefocusofthisworkisonregretminimizationinSSP.Itbuildsonextensiveliteratureontheoreticala...
LowerComplexityboundsforFinite-SumConvex-ConcaveMinimaxOptimizationProblemsGuangzengXie1LuoLuo2YijiangLian3ZhihuaZhang45AbstractconvexsuchthatX⊆RdxandY⊆Rdy.Thisformula-tioncontainsseveralpopularm...
LinearLowerboundsandConditioningofDifferentiableGamesAdamIbrahim1WaïssAzizian2GauthierGidel1IoannisMitliagkas1Abstractmind(Meschederetal.,2017),andtomakemattersworse,havebeenoftentunedsuboptimally...
ImprovedboundsonMinimaxRegretunderLogarithmicLossviaSelf-ConcordanceBlairBilodeau123DylanJ.Foster4DanielM.Roy123AbstractTheloglosspenalizestheplayerbasedonhowmuchprob-abilitymasstheyplaceontheactua...
EstimationofboundsonPotentialOutcomesForDecisionMakingMaggieMakar1FredrikJohansson2JohnGuttag1DavidSontag1Abstractneedstoknowifthepotentialoutcomesundertreatmentandnon-treatmentfallwithin2–3.Learn...
DissectingNon-VacuousGeneralizationboundsbasedontheMean-FieldApproximationKonstantinosPitas1Abstract(a)Explaininghowoverparametrizedneuralnet-Figure1.Risk-ComplexityplotforMNIST10:Theareabelowworks...
Continuous-timeLowerboundsforGradient-basedAlgorithmsMichaelMuehlebach1MichaelI.Jordan1Abstractindependentlowerboundsthenresultfromanunboundedincreaseintheproblemdimension.Thisestablishes,forThisar...
ConcentrationboundsforCVaRestimation:Thecasesoflight-tailedandheavy-taileddistributionsPrashanthL.A.1KrishnaJagannathan2RaviKumarKolla3Abstract(VaR)andConditional-Value-at-Risk(CVaR)aretworisk-awar...
TighterProblem-DependentRegretboundsinReinforcementLearningwithoutDomainKnowledgeusingValueFunctionboundsAndreaZanette1EmmaBrunskill2AbstractFortunatelyinpracticereinforcementlearningalgorithmsof-t...
POLITEX:RegretboundsforPolicyIterationUsingExpertPredictionYasinAbbasi-Yadkori1PeterL.Bartlett2KushBhatia2NevenaLazic´3CsabaSzepesvári4GellértWeisz4Abstractmodel-basedalgorithms,andtheoreticalev...
OnVariationalboundsofMutualInformationBenPoole1SherjilOzair12Aa¨ronvandenOord3AlexanderA.Alemi1GeorgeTucker1AbstractFigure1.Schematicofvariationalboundsofmutualinformationpresentedinthispaper.Node...
OnCertifyingNon-uniformboundsagainstAdversarialAttacksChenLiu1RyotaTomioka2VolkanCevher1AbstractCWattack(Carlini&Wagner,2017).MostattackmethodssearchfortheadversarialexamplebyutilizingthegradientTh...