2013-Deep Learning for Signal and Information Processing .pdf

《2013-Deep Learning for Signal and Information Processing .pdf》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、DEEPLEARNINGFORSIGNALANDINFORMATIONPROCESSINGLiDengandDongYuMicrosoftResearchOneMicrosoftWayRedmond,WA980521TableofContentsChapter1:Introduction.................................................................................................................

2、...............61.1DefinitionsandBackground..........................................................................................................61.2OrganizationofThisBook..................................................................................

3、...........................9Chapter2:HistoricalContextofDeepLearning........................................................................................10Chapter3:ThreeClassesofDeepLearningArchitectures...................................................

4、......................173.1AThree-CategoryClassification..................................................................................................173.2GenerativeArchitectures...........................................................................

5、.................................193.3DiscriminativeArchitectures.......................................................................................................233.4HybridGenerative-DiscriminativeArchitectures........................................

6、................................26Chapter4:Generative:DeepAutoencoder................................................................................................294.1Introduction...........................................................................

7、.....................................................294.2UseofDeepAutoencodertoExtractSpeechFeatures...............................................................304.3StackedDenoisingAutoencoder.............................................................

8、....................................404.4TransformingAutoencoder.........................................................................................................41Chapter5:Hybrid:Pre-TrainedDee