BettergeneralizationwithlessdatausingrobustgradientdescentMatthewJ.Holland1KazushiIkeda2Abstractthedefactostandardlearningstrategyfortacklingmostma-chinelearningproblems(Kearns&Schapire,1994;Bartle...
AUTOVC:Zero-ShotVoiceStyleTransferwithOnlyAutoencoderLossKaizhiQian1YangZhang23ShiyuChang23XuesongYang1MarkHasegawa-Johnson1Abstractingdata,i.e.speechpairswherethetwospeakersutterthesamesentences.O...
AsynchronousBatchBayesianOptimisationwithImprovedLocalPenalisationAhsanS.Alvi12BinxinRu1JanCalliess13StephenJ.Roberts123MichaelA.Osborne12Abstractexample,considertheoptimisationofthenumberofunitsin...
AutomatedModelSelectionwithBayesianQuadratureHenryChai1Jean-Franc¸oisTon2MichaelA.Osborne34RomanGarnett1AbstractformZ=f(Dθ)π(θ)dθ,whereθisavectorofmodelparameters,f(Dθ)isalikelihood,andπ(θ...
ApproximatingOrthogonalMatriceswithEffectiveGivensFactorizationThomasFrerix1JoanBruna2AbstracttionsofunitarymatricesintheformofGivensfactorization(Givens,1958).Givensrotationsarelocalizedinatwo-Wea...
AnalyzingandImprovingRepresentationswiththeSoftNearestNeighborLossNicholasFrosst1NicolasPapernot1GeoffreyHinton1Abstractthesoftnearestneighborloss(Salakhutdinov&Hinton,2007),whichweexploretomeasure...
AnomalyDetectionwithMultiple-HypothesesPredictionsDucTamNguyen12ZhongyuLou2MichaelKlar2ThomasBrox1Abstractdrivingsystem,wemayhaveatestcasewithabearorakangarooontheroad.Fordefectdetectioninmanufactu...
AdversarialOnlineLearningwithnoiseAlonResler1YishayMansour12Abstract&Lugosi,2006;Bubeck&Cesa-Bianchi,2012)).Wepresentandstudymodelsofadversarialon-Bothmodelsassumethattheobservedfeedbackisexact,lin...
AdversarialGenerationofTime-FrequencyFeatureswithapplicationinaudiosynthesisAndre´sMarafioti1NickiHolighaus1Nathanae¨lPerraudin2PiotrMajdak1Abstracttrainedsimultaneouslyinatwo-playermin-maxgame:T...
Vallossvs.testerrorAddressingtheLoss-MetricMismatc(ah)1616withAdaptive0.9Loss(b)2525Alignment0.9Classificationerror(%)1414Crossentropyloss0.7Crossentropyloss0.7Train✅Lossvalue2020Loss✅1-AUCPR(%)M...
ActiveLearningwithDisagreementGraphsCorinnaCortes1GiuliaDeSalvo1ClaudioGentile1MehryarMohri12NingshanZhang3Abstractheinteractivelyselectspointstolabel.Intheon-linesetting,thelearnerreceivesasequenc...
Zeno:DistributedStochasticGradientDescentwithSuspicion-basedFault-toleranceCongXie1OluwasanmiKoyejo1IndranilGupta1Abstractvarianceandmagnitude,makingthemhardtodistinguish.Itisalsopossiblethatindiff...
TrainingNeuralMachineswithTrace-BasedSupervisionMatthewB.Mirman1DimitarDimitrov1PavleDjordjevic1TimonGehr1MartinVechev1Abstractadditionalamountsofsupervisionprovidedtotheseinter-pretablecomponentsd...
TheoreticalAnalysisofSparseSubspaceClusteringwithMissingEntriesManolisC.Tsakiris1Rene´Vidal2Abstractfoundnumerousapplicationsinmachinelearning,com-putervision,patternrecognition,bioinformaticsands...
TheoreticalAnalysisofImage-to-ImageTranslationwithAdversarialLearningXudongPan1MiZhang1DaizongDing1Abstractagenerator(i.e.anadaptivemodelthatmapsagaussiannoisetoafakesample)andadiscriminator(i.e.an...
TheGeneralizationErrorofDictionaryLearningwithMoreauEnvelopesAlexandrosGeorgogiannis1Abstracttosomepredefinedfamilyofmatrices.Fromthestatisti-callearningtheoryperspective,theaimistominimizetheThisi...
SubspaceEmbeddingandLinearRegressionwithOrliczNormAlexandrAndoni1ChengyuLin1YingSheng1PeilinZhong1RuiqiZhong1Abstractwherel:Rn→R+isthelossfunction.Whenl(y)=nWeconsiderageneralizationoftheclassicli...
StructuredOutputLearningwithAbstention:ApplicationtoAccurateOpinionPredictionAlexandreGarcia1SlimEssid1ChloéClavel1Florenced’Alché-Buc1Abstractetal.,2016)hasevolvedtowardsamoreinvolvedmachinelea...
StructuredEvolutionwithCompactArchitecturesforScalablePolicyOptimizationKrzysztofChoromanski1MarkRowland2VikasSindhwani1RichardE.Turner2AdrianWeller23Abstractagentviadirectpolicysearchcanbecastasam...
StrassenNets:DeepLearningwithaMultiplicationBudgetMichaelTschannen1AranKhanna2AnimaAnandkumar23Abstractandreducingthenumericalprecisionofweightsandactiva-tions(seeSection1.1foradetailedoverview).Al...
