PrincipalBitAnalysis:AutoencodingwithSchur-ConcaveLossSourbhBhadane1AaronB.Wagner1JayadevAcharya1Abstractandgareselectedthroughtrainingfromtheclassoffunc-Weconsideralinearautoencoderinwhichthetions...
ABitMoreBayesian:Domain-InvariantLearningwithUncertaintyZehaoXiao1JiayiShen1XiantongZhen12LingShao2CeesG.M.Snoek1AbstractLietal.,2018a)orspecificarchitectures(Lietal.,2017;D’Innocente&Caputo,2018)...
BeingBayesian,EvenJustaBit,FixesOverconfidenceinReLUNetworksAgustinusKristiadi1MatthiasHein1PhilippHennig12Abstract2016).Themostobviouscasesaretestpointsthatlie“faraway”fromthetrainingdata.Manyme...