QuantizationAlgorithmsforRandomFourierFeaturesXiaoyunLi,PingLiCognitiveComputingLabBaiduResearch10900NE8thStBellevueWA98004USA{lixiaoyun996,pingli98}@gmail.comAbstract1.IntroductionThemethodofrando...
PolicyCacheswithSuccessorFeaturesMarkNemecek1RonaldParr1Abstracttaskswhichvaryintheirrewardfunctions,butwherethedynamicsremainthesame.Althoughlimitedinscope,thisTransferinreinforcementlearningisbas...
ExaminingandCombatingSpuriousFeaturesunderDistributionShiftChuntingZhou1XuezheMa2PaulMichel1GrahamNeubig1Abstract(a)ERM(b)GroupDRO[clean](average=93.3%robust=83.4%)(average=96.5%robust=95.8%)Acentr...
APS:ActivePretrainingwithSuccessorFeaturesHaoLiu1PieterAbbeel1Abstract2019;Vinyalsetal.,2019;Badiaetal.,2020a)andsolvingcomplexroboticcontroltasks(Andrychowiczetal.,2017;Weintroduceanewunsupervised...
ANewRepresentationofSuccessorFeaturesforTransferacrossDissimilarEnvironmentsMajidAbdolshah1HungLe1ThommenKarimpanalGeorge1SunilGupta1SantuRana1SvethaVenkatesh1Abstractintoindependentsub-domains.How...
SparseGaussianProcesseswithSphericalHarmonicFeaturesVincentDutordoir1NicolasDurrande1JamesHensman2AbstractyWeintroduceanewclassofinter-domainvari-biasationalGaussianprocesses(GP)wheredataismappedon...
ScalableandEfficientComparison-basedSearchwithoutFeaturesDaniyarChumbalov1LucasMaystre2MatthiasGrossglauser1Abstractameaningfulquery.However,thisisoftenanon-trivialtaskforahumanuser.Forexample,most...
PreferenceModelingwithContext-DependentSalientFeaturesAmandaBower1LauraBalzano2Abstractmodelsaddressthisissuewithmodelingnoise,ignoringitssystematicnature.Weobserve,asothershavebeforeusWeconsiderth...
GraphRandomNeuralFeaturesforDistance-PreservingGraphRepresentationsDanieleZambon1CesareAlippi12LorenzoLivi34AbstractWhendatacomeasrealvectors,theseminalpaperbyRahimi&Recht(2008a)providesanefficient...
GeneralisationerrorinlearningwithrandomFeaturesandthehiddenmanifoldmodelFedericaGerace1BrunoLoureiro1FlorentKrzakala2MarcMézard2LenkaZdeborová1Abstracttiveresearchsubject.Thetraditionallearningth...
DecentralisedLearningwithDistributedGradientDescentandRandomFeaturesDominicRichards1PatrickRebeschini1LorenzoRosasco234Abstractofnewdatapoints.Duetothegrowingsizeofmoderndatasetsandcomplexityofmany...
TowardsaUnifiedAnalysisofRandomFourierFeaturesZhuLi1Jean-FrançoisTon1DinoOglic2DinoSejdinovic1Abstractimplicitcomputationofaninnerproductbetweenrichfea-turerepresentationsofdatathroughthekerneleva...
Sample-OptimalParametricQ-LearningUsingLinearlyAdditiveFeaturesLinF.Yang1MengdiWang1Abstractthistheoretical-sharpresultdoesnotgeneralizetopracticalproblemswhereS,Acanbearbitrarilylargeorinfinite.Co...
HybridModelswithDeepandInvertibleFeaturesEricNalisnick1AkihiroMatsukawa1YeeWhyeTeh1DilanGorur1BalajiLakshminarayanan1Abstractdetection(Bishop,1994),semi-supervisedlearning(Drucketal.,2007),andinfor...
DiscoveringConditionallySalientFeatureswithStatisticalGuaranteesJaimeRoqueroGimenez1JamesZou2Abstractingsomestatisticalcontrolontherateoffalsediscoveries.Thisproblemhasbeenextensivelystudiedinthest...
AdversarialGenerationofTime-FrequencyFeatureswithapplicationinaudiosynthesisAndre´sMarafioti1NickiHolighaus1Nathanae¨lPerraudin2PiotrMajdak1Abstracttrainedsimultaneouslyinatwo-playermin-maxgame:T...
ProvableVariableSelectionforStreamingFeaturesJingWang1JieShen2PingLi3AbstractandonlyrelatedFeatureswillbeincludedintherecommen-dationmodel.Hence,inreal-worldapplications,FeaturesInlarge-scalemachin...
OntheSpectrumofRandomFeaturesMapsofHighDimensionalDataZhenyuLiao1RomainCouillet12AbstractFeaturesrelevanttosomegiventask.ThenonlinearitiesmaketheserepresentationsmoremightybutmeanwhileRandomfeature...
RandomFourierFeaturesforKernelRidgeRegression:ApproximationBoundsandStatisticalGuaranteesHaimAvron1MichaelKapralov2CameronMusco3ChristopherMusco3AmeyaVelingker2AmirZandieh2AbstractIntheabove,K2Rn⇥...