原创力文档- 0
- 0
- 约6.07千字
- 约 16页
- 2026-05-08 发布于北京
- 举报
Review:PaperReading
RepresentationLearning-AReviewandNewPerspectives范老师2020/05/17
BackgroundUnsupervisedfeaturelearninganddeeplearningProbabilisticmodelsAuto-encodersManifoldlearningDeepnetworksRepresentationlearningDensityestimationManifoldlearning
1IntroductionDeeplearningmethods:compositionofmultiplenon-lineartransformations,foryieldingmoreabstractandusefulrepresentations.Answerquestions:whatmakesonerepresentationbetterthananother?Howshouldwecomputeisrepresentation,i.e.featureextraction?Whatareappropriateobjectivesforlearninggoodrepres
文档评论（0）