UnbiasedGradientEstimationinUnrolledComputationGraphswithPersistentEvolutionStrategiesPaulVicol12LukeMetz2JaschaSohl-Dickstein2AbstractZipser,1989;Tallec&Ollivier,2017a;Mujikaetal.,2018;Benzingetal...
OntheDifficultyofUnbiasedAlphaDivergenceMinimizationTomasGeffner1JustinDomke1AbstractExistingalpha-divergenceminimizationalgorithmscanbeclassifiedintotwobroadgroups:biasedmethods(Li&Severalapproxim...
UnbiasedRiskEstimatorsCanMislead:ACaseStudyofLearningwithComplementaryLabelsYu-TingChou1GangNiu2Hsuan-TienLin1MasashiSugiyama23Abstractetal.,2014;2015;Niuetal.,2016;Sakaietal.,2017;2018),unlabeled-...
AR-DAE:TowardsUnbiasedNeuralEntropyGradientEstimationJaeHyunLim12AaronCourville1234ChristopherPal154Chin-WeiHuang12Abstractcontrolthisquantityaspartoftheoptimizationobjective.Inlightofthis,wepropos...
TamingMAML:EfficientUnbiasedMeta-ReinforcementLearningHaoLiu1RichardSocher1CaimingXiong1Abstractreinforcementlearning(Meta-RL)(Wangetal.,2016;Duanetal.,2016;Mishraetal.,2018;Finnetal.,2017;NicholWh...
Noisin:UnbiasedRegularizationforRecurrentNeuralNetworksAdjiB.Dieng1RajeshRanganath2JaanAltosaar3DavidM.Blei1Abstractvationconditionalonitsstate.ThekeyelementofanRNNisitstransitionfunction.Thetransi...