BidirectionalLearningforTime-seriesModelswithHiddenUnitsTakayukiOsogami1HiroshiKajino1TaroSekiyama1Abstractdenunits,however,makeitdifficulttolearntheparametersofthosemodels.Forexample,theBaum-Welch...
AutomatedCurriculumLearningforNeuralNetworksAlexGraves1MarcG.Bellemare1JacobMenick1Re´miMunos1KorayKavukcuoglu1Abstractmenttothenexttask,alongwithafixedprobabilityofre-turningtoearliertasks,toprev...
AnAlternativeSoftmaxOperatorforReinforcementLearningKavoshAsadi1MichaelL.Littman1AbstractAnidealsoftmaxoperatorisaparameterizedsetofoperatorsthat:Asoftmaxoperatorappliedtoasetofvaluesactssomewhatli...
AdaNet:AdaptiveStructuralLearningofArtificialNeuralNetworksCorinnaCortes1XavierGonzalvo1VitalyKuznetsov1MehryarMohri21ScottYang2Abstractoflayersandunitsspecifiedsincethereneedstobeatleastonepaththr...
ActiveLearningforTop-KRankAggregationfromNoisyComparisonsSoheilMohajer1ChanghoSuh2AdelElmahdy1Abstractcantitems,saytop-K,areoftendesiredtoberetrieved.Weexploreanactivetop-KrankingproblemInaneffortt...