High-PerformanceLarge-ScaleImageRecognitionWithoutNormalizationAndrewBrock1SohamDe1SamuelL.Smith1KarenSimonyan1AbstractFigure1.ImageNetValidationAccuracyvsTrainingLatency.Allnumbersaresingle-model,...
LossFunctionSearchforFaceRecognitionXiaoboWang1ShuoWang1ChengChi2ShifengZhang2TaoMei1AbstractGenerally,theCNNsareequippedwithclassificationlossfunctions(Liuetal.,2017;Wangetal.,2018f;Chenetal.,Infa...
DeepIsometricLearningforVisualRecognitionHaozhiQi1ChongYou1XiaolongWang12YiMa1JitendraMalik1Abstract(optional)b=1?(?)skipconnectionb=-1Initialization,normalization,andskipconnectionsb=-3arebeliev...
Data-EfficientImageRecognitionwithContrastivePredictiveCodingOlivierJ.He´naff1AravindSrinivas2JeffreyDeFauw1AliRazavi1CarlDoersch1S.M.AliEslami1AaronvandenOord1Abstract2xfewerlabelsHumanobserversc...
Context-AwareZero-ShotLearningforObjectRecognitionE´loiZablocki1PatrickBordes1BenjaminPiwowarski1LaureSoulier1PatrickGallinari12Abstractaclassforwhichnosuperviseddataisavailable,byusingknowledgeac...
DeepPredictiveCodingNetworkforObjectRecognitionHaiguangWen1KuanHan1JunxingShi1YizhenZhang1EugenioCulurciello12ZhongmingLiu12Abstractfeaturesbyafeedforwardpassthroughstacksofconvolu-tionallayers.Bas...
MultichannelEnd-to-endSpeechRecognitionTsubasaOchiai1ShinjiWatanabe2TakaakiHori2JohnR.Hershey2Abstractintoasingleneuralnetwork.Specifically,anattention-basedencoder-decoderframework(Chorowskietal.,...