ontheUniversalityofInvariantNetworksHaggaiMaron1EthanFetaya23NimrodSegol1YaronLipman1AbstractwithrespecttosomegroupGwillbereferredhereasG-invariantnetworks.Constraininglinearlayersinneuralnetworkst...
onthestatisticalrateofnonlinearrecoveryingenerativemodelswithheavy-taileddataXiaohanWei1ZhuoranYang2ZhaoranWang3Abstract1.IntroductionWeconsiderestimatingahigh-dimensionalvec-Considerarandompair(x,...
ontheLong-termImpactofAlgorithmicDecisionPolicies:EffortUnfairnessandFeatureSegregationthroughSocialLearningHodaHeidari1VedantNanda2KrishnaP.Gummadi2AbstractThisrealizationhasrecentlyspawnedanactiv...
onThePowerofCurriculumLearninginTrainingDeepNetworksGuyHacohen12DaphnaWeinshall1Abstracttypicallyreflectstheircomplexity.Thestudentisthengrad-uallyintroducedtotheseconceptsbyincreasingcomplexity,Tr...
ontheLinearSpeedupAnalysisofCommunicationEfficientMomentumSGDforDistributedNon-ConvexOptimizationHaoYu1RongJin1SenYang1Abstractwherefi(x)=E⇠i⇠Di[Fi(x;⇠i)]aresmoothnon-convexfunctionswithpossibly...
ontheLimitationsofRepresentingFunctionsonSetsEdwardWagstaff1FabianB.Fuchs1MartinEngelcke1IngmarPosner1MichaelOsborne1AbstractInputOutputRecentworkontherepresentationoffunctionsϕ+ρf(x1,…,xM)onset...
ontheLimitationsofRepresentingFunctionsonSetsEdwardWagstaff1FabianB.Fuchs1MartinEngelcke1IngmarPosner1MichaelOsborne1AbstractInputOutputRecentworkontherepresentationoffunctionsϕ+ρf(x1,…,xM)onset...
ontheGeneralizationGapinReparameterizableReinforcementLearningHuanWang1StephanZheng1CaimingXiong1RichardSocher1Abstract2018a).Amodelthatperformswellinthetrainingenvi-ronment,mayormaynotperformwellw...
ontheImpactoftheActivationFunctiononDeepNeuralNetworksTrainingSoufianeHayou1ArnaudDoucet1JudithRousseau1Abstractobtainedsimilarresultsusingatopologicalmeasureofex-pressiveness.Theweightinitializati...
ontheImpactoftheActivationFunctiononDeepNeuralNetworksTrainingSoufianeHayou1ArnaudDoucet1JudithRousseau1Abstractobtainedsimilarresultsusingatopologicalmeasureofex-pressiveness.Theweightinitializati...
ontheConvergenceandRobustnessofAdversarialTrainingYisenWang1XingjunMa2JamesBailey2JinfengYi1BowenZhou1QuanquanGu3Abstract2018),theyarevulnerabletoadversarialexamplescraftedbyaddingsmall,humanimperc...
ontheFeasibilityofLearning,RatherthanAssuming,HumanBiasesforRewardInferenceRohinShah1NoahGundotra1PieterAbbeel1AncaD.Dragan1Abstractp(as)µebQ(s,a;r)w!Ourgoalisforagentstooptimizetherightre-wwardfu...
ontheDesignofEstimatorsforBanditOff-PolicyEvaluationNikosVlassis1AurelienBibaut2MariaDimakopoulou1TonyJebara1Abstractofgreatinterest:Givenabanditmodel,whatisalow-riskestimatorofthecounterfactualtar...
ontheConnectionBetweenAdversarialRobustnessandSaliencyMapInterpretabilityChristianEtmann12SebastianLunz3PeterMaass1Carola-BibianeScho¨nlieb3AbstractFigure1.Animageofadog(left),thesaliencymapsofahi...
ontheComputationandCommunicationComplexityofParallelSGDwithDynamicBatchSizesforStochasticNon-ConvexOptimizationHaoYu1RongJin1Abstractvariables⇣⇠Drefertoinstance-labelpairsandF(x;⇣)referstothecor...
ontheComplexityofApproximatingWassersteinBarycentersAlexeyKroshnin123DarinaDvinskikh41PavelDvurechensky41AlexanderGasnikov512NazariiTupitsa15Ce´sarA.Uribe6Abstractlearningandoptimizationcommunitie...
onSymmetricLossesforLearningfromCorruptedLabelsNontawatCharoenphakdee12JongyeongLee12MasashiSugiyama21AbstractFigure1.Examplesoflossesusedinthispaper.Thezero-oneloss,sigmoidloss,andunhingedlossares...
onSparseLinearRegressionintheLocalDifferentialPrivacyModelDiWang1JinhuiXu1Abstractsciences(Marascuilo&Serlin,1988),genomicsresearch(Bu˘žková,2013)andsignalrecovery(Bühlmann&VanInthispaper,westu...
onScalableandEfficientComputationofLargeScaleOptimalTransportYujiaXie1MinshuoChen1HaomingJiang1TuoZhao1HongyuanZha1AbstractearliestliteraturedatesbacktoMonge(1781).Sincethen,ithasattractedincreasin...