TractableStructuredNatural-GradientDescentUsingLocalParameterizationsWuLin1FrankNielsen2MohammadEmtiyazKhan3MarkSchmidt14AbstractFinally,manyrobustorglobaloptimizationtechniquesem-ployq(w)tosmootho...
TowardsOpenAdHocTeamworkUsingGraph-basedPolicyLearningArrasyRahman1NiklasHo¨pner2FilipposChristianos1StefanoV.Albrecht1AbstractPrioradhocteamworkapproaches(Barrett&Stone,2015;Albrechtetal.,2016;Ba...
TowardsCertifying∞RobustnessUsingNeuralNetworkswith∞-distNeuronsBohangZhang1TianleCai23ZhouLu4DiHe5LiweiWang16Abstract1.IntroductionItiswell-knownthatstandardneuralnetworks,Modernneuralnetworksar...
SparsewithinSparseGaussianProcessesUsingNeighborInformationGia-LacTran12DimitriosMilios1PietroMichiardi1MaurizioFilippone1Abstract&Ghahramani,2005;Naish-Guzman&Holden,2007;Tit-sias,2009;Hensmanetal...
Solvinghigh-dimensionalparabolicPDEsUsingthetensortrainformatLorenzRichter123LeonSallandt4NikolasNu¨sken5AbstractonstochasticrepresentationsofthePDEsunderconsidera-tion(Eetal.,2017;Raissietal.,201...
SkillDiscoveryforExplorationandPlanningUsingDeepSkillGraphsAkhilBagaria1JasonSenthil1GeorgeKonidaris1AbstractWeintroduceanewskill-discoveryalgorithmthatbuildsadiscretegraphrepresentationoflargecon-...
SegmentingHybridTrajectoriesUsingLatentODEsRuianShi12QuaidMorris123Abstracttioncharacteristics,suchastherateofencounter,andrateofsuccessfulpredationperencounter.WhentheseparametersSmoothdynamicsint...
SafeReinforcementLearningUsingAdvantage-BasedInterventionNolanWagener1ByronBoots2Ching-AnCheng3AbstractFigure1.Advantage-basedinterventionofSAILRandconstruc-tionofthesurrogateMDPM.InM,wheneverthepo...
RobustReinforcementLearningUsingLeastSquaresPolicyIterationwithProvablePerformanceGuaranteesKishanPanaganti1DileepKalathil1AbstractThismismatchbetweenthetrainingandtestingenvironmentparameterscansi...
QuantifyingtheBenefitofUsingDifferentiableLearningoverTangentKernelsEranMalach1PritishKamath2EmmanuelAbbe3NathanSrebro2AbstractCanallthesuccessofdeeplearningbeexplainedUsingtheNTK?Thiswouldimplytha...
PsiPhi-Learning:ReinforcementLearningwithDemonstrationsUsingSuccessorFeaturesandInverseTemporalDifferenceLearningAngelosFilos1ClareLyle1YarinGal1SergeyLevine2NatashaJaques23GregoryFarquhar4Abstract...
PriorImage-ConstrainedReconstructionUsingStyle-BasedGenerativeModelsVarunA.Kelkar1MarkA.Anastasio1Abstractqualityofimagesgenerated,invertibilityandimagerepre-sentation,andameaningfulcontrolofsemant...
PolicyAnalysisUsingSyntheticControlsinContinuous-TimeAlexisBellot12MihaelavanderSchaar123Abstractoraverageinaneighbourhoodofcontrols)oftenprovidesamoreinformativecomparisonfortreatmenteffectestimat...
PersonalizedFederatedLearningUsingHypernetworksAvivShamsian∗1AvivNavon∗1EthanFetaya1GalChechik12Abstractwhileminimizingcommunication.Unfortunately,learningasingleglobalmodelmayfailwhenthedatadist...
ParallelDropletControlinMEDABiochipsUsingMulti-AgentReinforcementLearningTung-CheLiang1JinZhou1Yun-ShengChan2Tsung-YiHo3KrishnenduChakrabarty1Chen-YiLee2Abstract1.IntroductionMicrofluidicbiochipsar...
OnRecoveringfromModelingErrorsUsingTestingBayesianNetworksHaiyingHuang1AdnanDarwiche1Abstracttoprovideguaranteesonthelearnedfunctions.Yet,model-basedsupervisedlearninghasbeenoutperformedbytheWecons...
MorphVAE:GeneratingNeuralMorphologiesfrom3D-WalksUsingaVariationalAutoencoderwithSphericalLatentSpaceSophieLaturnus1PhilippBerens123Abstractputsitreceivesandwherethecomputedoutputsaresentto(Hilleta...
LearningNodeRepresentationsUsingStationaryFlowPredictiononLargePaymentandCashTransactionNetworksCiwanCeylan12SallaFranze´n2FlorianT.Pokorny1AbstractannualGDP(UnitedNationsofficeondrugsandcrime,201...
IntermediateLayerOptimizationforInverseProblemsUsingDeepGenerativeModelsGiannisDaras1JosephDean1AjilJalal1AlexandrosG.Dimakis1Abstractthatrelyonapre-trainedgenerator,buildingonthegeneralframeworkin...