UnICORNN:ARecurrentmodelforlearningverylongtimedependenciesT.KonstantinRusch1SiddharthaMishra1AbstractsuchasinLSTMs(Hochreiter&Schmidhuber,1997)andGRUs(Choetal.,2014),wheretheadditivestructureofThe...
TrainingRecurrentNeuralNetworksviaForwardPropagationThroughTimeAnilKag1VenkateshSaligrama1Abstractempiricalriskfunction:Back-propagationthroughtime(BPTT)hasbeen[W∗,v∗]=argminL(W,v)=1NT(yi,yˆti)w...
Re-understandingFinite-StateRepresentationsofRecurrentPolicyNetworksMohamadH.Danesh1AnuragKoul1AlanFern1SaeedKhorram1Abstracttivehumaninterpretationoftheunderlying“strategicrole"oftheattended-toel...
VideoOneNet:BidirectionalConvolutionalRecurrentOneNetwithTrainableDataStepsforVideoProcessingZolta´nA´.Milacski1Barnaba´sPo´czos2Andra´sLo˝rincz1Abstractthestate-of-the-artonmanydataprocessin...
TransformationofReLU-basedRecurrentneuralnetworksfromdiscrete-timetocontinuous-timeZahraMonfared1DanielDurstewitz12AbstractMorerecently,inthenaturalsciences,biologyandphysicsinparticular,RNNswereal...
RecurrentHierarchicalTopic-GuidedRNNforLanguageGenerationDandanGuo1BoChen1RuiyingLu1MingyuanZhou2Abstractlatentrepresentation,theytypicallytreateachdocumentasabagofwords(BoW),ignoringwordorder(Grif...
ImprovingtheGatingMechanismofRecurrentNeuralNetworksAlbertGu1CaglarGulcehre2TomPaine2MattHoffman2RazvanPascanu2Abstract(Chungetal.,2014,GRUs).ThesegatedRNNshavebeenverysuccessfulinseveraldifferenta...
HowRecurrentnetworksimplementcontextualprocessinginsentimentanalysisNiruMaheswaranathan1DavidSussillo1AbstractRigorouslyunderstandinghownetworkssolveimportanttasksisacentralchallengeindeeplearning....
FrequentistUncertaintyinRecurrentNeuralNetworksviaBlockwiseInfluenceFunctionsAhmedM.Alaa1MihaelavanderSchaar12Abstract(a)Blockwiseresampling(b)SequentialconfidenceintervalsRNNPredictionsRecurrentne...
ApproximatingStackedandBidirectionalRecurrentArchitectureswiththeDelayedRecurrentNeuralNetworkJavierS.Turek1ShaileeJain2VyA.Vo1MihaiCapota˘1AlexanderG.Huth23TheodoreL.Willke1Abstractelement.Howeve...
UnderstandingandControllingMemoryinRecurrentNeuralNetworksDoronHaviv12AlexnaderRivkind234OmriBarak23Abstract1.IntroductionTobeeffectiveinsequentialdataprocessing,Re-RecurrentNeuralNetworks(RNN)aret...
State-RegularizedRecurrentNeuralNetworksChengWang1MathiasNiepert1AbstracttraininganditremainschallengingtodeterminewhethertheextractedDFAfaithfullymodelstheRNN’sstatetransi-Recurrentneuralnetworks...
RecurrentKalmanNetworks:FactorizedInferenceinHigh-DimensionalDeepFeatureSpacesPhilippBecker123HaritPandya4GregorGebhardt1ChengZhao5JamesTaylor6GerhardNeumann423Abstract1.IntroductionInordertointegr...
POPQORN:QuantifyingRobustnessofRecurrentNeuralNetworksChing-YunKo1ZhaoyangLyu2Tsui-WeiWeng3LucaDaniel3NgaiWong1DahuaLin2Abstractlini&Wagner,2018).Asmallperturbationthathumansaremostlyimmunetocanbec...
OvercomingMean-FieldApproximationsinRecurrentGaussianProcessModelsAlessandroDavideIalongo12MarkvanderWilk3JamesHensman3CarlEdwardRasmussen1Abstractthedistributionoftheoutcome(vonNeumannetal.,1944)....
OptimalKronecker-SumApproximationofRealTimeRecurrentLearningFrederikBenzing1MarceloMatheusGauy1AsierMujika1AndersMartinsson1AngelikaSteger1Abstractachievednotablesuccesses.However,fewimprovementsha...
ARecurrentNeuralCascade-basedModelforContinuous-TimeDiffusionSylvainLamprier1Abstractprocesstoextractthemaindynamicsofthenetworkfromobserveddiffusionepisodes.Manyworkshavebeenproposedintheliteratur...
RevivingandImprovingRecurrentBack-PropagationRenjieLiao123YuwenXiong12EthanFetaya13LisaZhang13KiJungYoon45XaqPitkow45RaquelUrtasun123RichardZemel136Abstractetal.,2016).Moreover,asthenumberofunrolli...
RecurrentPredictiveStatePolicyNetworksAhmedHefny1ZitaMarinho23WenSun2SiddharthaS.Srinivasa4GeoffreyGordon1Abstract1.IntroductionWeintroduceRecurrentPredictiveStatePolicyRecently,therehasbeensignifi...
ProbabilisticRecurrentState-SpaceModelsAndreasDoerr12ChristianDaniel1MartinSchiegg1DuyNguyen-Tuong1StefanSchaal23MarcToussaint4SebastianTrimpe2AbstractInrealsystems,thelatentstatextcantypicallynotb...