生成式深度学习(影印版)+2

内容简介

生成式建模(Generative modeling)是人工智能领域热门的研究课题之一。现在算法已经可以教一台机器在绘画、写作和作曲等人类活动中取得的表现。通过这本实用指南，机器学师和数据科学家们将学会如何通过生成式深度学重新创建一些令人印象深刻的程序示例，例如变分自编码器、生成对抗网络、编码器—解码器模型和世界模型。

目录

PrefacePart Ⅰ I troductio to Ge erative Deep Lear i g 1. Ge erative Modeli g What Is Ge erative Modeli g Ge erative Versus Discrimi ative Modeli g Adva ces i Machi e Lear i g The Rise of Ge erative Modeli g The Ge erative Modeli g framework Probabilistic Ge erative Models Hello Wrodl! Your First Probabilistic Ge erative Model aive Bayes Hello Wrodl! Co ti ued The Challe ges of Ge erative Modeli g Represe tatio Lear i g Setti g Up Your E viro me t Summary 2. Deep Lear i g Structured a d U structured Data Deep eural etworks Keras a d Te sorFlow Your First Deep eural etwork Loadi g the Data Buildi g the Model Compili g the Model Trai i g the Model evaluati g the Model Improvi g the Model Co volutio al Layers Batch ormalizatio Dropout Layers Putti g It All Together Summary 3. Variatio al Autoe coflers The Art Exhibitio Autoe coders Your First Autoe coder The E coder The Decoder Joi i g the E coder to the Decoder A alysis of the Autoe coder The Variatio al Art Exhibitio Buildi g a Variatio al Autoe coder The E coder The Loss Fu ctio A alysis of the Variatio al Autoe coder Usi g VAEs to Ge erate Faces Trai i g the VAE A alysis of the VAE Ge erati g ew Faces