【導(dǎo)讀】生成式對(duì)抗網(wǎng)絡(luò)模型(GAN)是基于深度學(xué)習(xí)的一種強(qiáng)大的生成模型，可以應(yīng)用于計(jì)算機(jī)視覺(jué)、自然語(yǔ)言處理、半監(jiān)督學(xué)習(xí)等重要領(lǐng)域。牛津大學(xué)Jakub Langr 撰寫(xiě)了《GANs in Action》的著作，囊括了關(guān)于GAN的定義、訓(xùn)練、變體等，是學(xué)習(xí)生成式對(duì)抗網(wǎng)絡(luò)非常好的學(xué)習(xí)圖書(shū)，是關(guān)于GAN的最好的書(shū)籍之一。

網(wǎng)址：

https://www./books/gans-in-action

【實(shí)戰(zhàn)GAN代碼和第一章便捷下載】

請(qǐng)關(guān)注專(zhuān)知公眾號(hào)（點(diǎn)擊上方藍(lán)色專(zhuān)知關(guān)注）

• 后臺(tái)回復(fù)“GAN實(shí)戰(zhàn)” 就可以獲取《GAN實(shí)戰(zhàn)》的下載鏈接~ 

實(shí)戰(zhàn)GAN

深度學(xué)習(xí)系統(tǒng)在識(shí)別文本、圖像和視頻中的模式方面做得非常好。但事實(shí)證明，創(chuàng)建逼真的圖像、自然的句子和段落，或翻譯成地道英語(yǔ)的應(yīng)用程序是難以實(shí)現(xiàn)的。生成式對(duì)抗網(wǎng)絡(luò)(GANs)通過(guò)將兩個(gè)相互對(duì)抗的神經(jīng)網(wǎng)絡(luò)(一個(gè)生成內(nèi)容，另一個(gè)拒絕質(zhì)量較差的樣本)配對(duì)，為這些挑戰(zhàn)提供了一個(gè)有希望的解決方案。

GAN

GANs已經(jīng)取得了被認(rèn)為是人工系統(tǒng)不可能取得的顯著成果，例如能夠生成逼真的人臉、將涂鴉變成類(lèi)似照片的圖像、將馬的視頻片段變成奔跑的斑馬。最重要的是，GANs學(xué)得很快，不需要費(fèi)力地標(biāo)注大量訓(xùn)練數(shù)據(jù)。

2014年由谷歌的Ian Goodfellow發(fā)明的生成式對(duì)抗網(wǎng)絡(luò)(GANs)是深度學(xué)習(xí)中最重要的創(chuàng)新之一。在GANs中，一個(gè)神經(jīng)網(wǎng)絡(luò)(生成器)生成內(nèi)容(圖像、句子等等)，另一個(gè)神經(jīng)網(wǎng)絡(luò)(識(shí)別器)確定它們是否來(lái)自生成器，因此是“假的”，還是來(lái)自訓(xùn)練數(shù)據(jù)集，因此是“真實(shí)的”。在這兩個(gè)系統(tǒng)之間的相互作用中，當(dāng)生成器試圖欺騙鑒別器相信“贗品”是真實(shí)的時(shí)，它會(huì)產(chǎn)生更真實(shí)的輸出。其結(jié)果是一個(gè)可以生成逼真圖像或自然文本和語(yǔ)音的生成器，以及一個(gè)訓(xùn)練有素的識(shí)別器，可以精確地識(shí)別和分類(lèi)這類(lèi)內(nèi)容。

本書(shū)簡(jiǎn)介

實(shí)戰(zhàn)GAN:用生成對(duì)抗式網(wǎng)絡(luò)進(jìn)行深度學(xué)習(xí)可以教會(huì)你如何構(gòu)建和訓(xùn)練自己的生成對(duì)抗式網(wǎng)絡(luò)。首先，您將了解生成模型以及GAN的工作原理，并概述它們的潛在用途。然后，隨著您探索GAN體系結(jié)構(gòu)的基礎(chǔ):生成器和鑒別器網(wǎng)絡(luò)，您將開(kāi)始構(gòu)建自己的簡(jiǎn)單的對(duì)抗系統(tǒng)。

通過(guò)本書(shū)迷人的示例和詳細(xì)的插圖，您將學(xué)習(xí)為不同的場(chǎng)景培訓(xùn)不同的GAN體系結(jié)構(gòu)。您將探索生成高分辨率圖像、圖像到圖像的轉(zhuǎn)換、對(duì)抗性學(xué)習(xí)以及目標(biāo)數(shù)據(jù)生成，從而使您的系統(tǒng)變得智能、有效和快速。

目錄：

1 INTRODUCTION TO GANS

1.1 Introduction

1.2 Prerequisites

1.3 What Are Generative Adversarial Networks?

1.3.1 GAN in Action

1.3.2 GAN Training

1.3.3 GAN Training Visualized

1.3.4 Reaching Equilibrium

1.3.5 The Pros and Cons of Studying GANs

1.4 Applications of GANs

1.5 Guide to this Book

1.6 Summary

2 AUTOENCODERS AS A PATH TO GANS

2.1 Why did we include this chapter?

2.1.1 Generative learning is a new area for most

2.1.2 Challenges of generative modelling

2.1.3 An Important part of the literature today

2.2 So what are autoencoders to GANs?

2.3 What are the reasons behind autoencoders?

2.4 Overview of Autoencoders

2.5 Usage of autoencoders

2.6 Unsupervised learning

2.7 New take on an old idea

2.8 Variational autoencoder (VAE)

2.9 Code is life

2.10 Summary

3 YOUR FIRST GAN: GENERATING HANDWRITTEN DIGITS

3.1 Introduction

3.1.1 Adversarial Training

3.1.2 The Generator and the Discriminator

3.1.3 GAN Training Algorithm

3.2 Tutorial: Generating Handwritten Digits

3.2.1 Import Statements

3.2.2 The Generator

3.2.3 The Discriminator

3.2.4 Build the Model

3.2.5 Training

3.2.6 Outputting Sample Images

3.2.7 Run the Model

3.2.8 Inspecting the Results

3.3 Conclusion

3.4 Chapter Summary

4 DEEP CONVOLUTIONAL GAN (DCGAN)

4.1 Introduction

4.2 Convolutional Neural Networks (ConvNets)

4.3 Brief History of the DCGAN

4.4 Batch Normalization

4.4.1 Computing Batch Normalization

4.5 Tutorial: Generating Handwritten Digits with DCGAN

4.5.1 Import Statements

4.5.2 The Generator

4.5.3 The Discriminator

4.5.4 Build & Run the DCGAN

4.5.5 Model Output

4.6 Conclusion

4.7 Chapter Summary

5 TRAINING & COMMON CHALLENGES: GANING FOR SUCCESS

5.1 Evaluation

5.1.1 Inception Score

5.1.2 Fréchet Inception Distance

5.2 Training challenges

5.2.1 Network depth

5.2.2 Game set-ups

5.2.3 Min-Max GAN (MM-GAN)

5.2.4 Non-Saturating GAN (NS-GAN)

5.2.5 Summary of game setups

5.2.6 Training hacks

5.3 Chapter summary

6 PROGRESSING WITH GANS

6.1 Latent space interpolation

6.2 They grow up so fast

6.2.1 Progressive Growing & Smoothing in of Higher Resolution Layers

6.2.2 Minibatch Standard Deviation

6.2.3 Equalized Learning Rate

6.2.4 Pixel-wise Feature Normalization

6.3 Summary of key innovations

6.4 Tensorflow Hub and hands-on

6.5 Practical Applications

6.6 Chapter summary

7 SEMI-SUPERVISED GAN

7.1 Introduction: “The GAN Zoo”

7.2 Semi-Supervised GAN (SGAN)

7.2.1 Why Semi-Supervised Learning?

7.2.2 What is Semi-Supervised GAN?

7.3 Tutorial: Implementing Semi-Supervised GAN

7.3.1 Architecture Diagram

7.3.2 Implementation

7.3.3 Setup

7.3.4 The Dataset

7.3.5 The Generator

7.3.6 The Discriminator

7.3.7 Build the Model

7.3.8 Training

7.3.9 Train the Model

7.3.10 Model Training and Test Accuracy

7.3.11 Comparison to a Fully-Supervised Classifier

7.4 Conclusion

7.5 Summary

8 CONDITIONAL GAN

8.1 Introduction

8.2 Conditional GAN (CGAN)

8.2.1 What is Conditional GAN?

8.3 Tutorial: Implementing Conditional GAN

8.3.1 Implementation

8.3.2 Setup

8.3.3 The Generator

8.3.4 Build the Model

8.3.5 Training

8.3.6 Outputting Sample Images

8.3.7 Train the Model

8.4 Inspecting the Output: Targeted Data Generation

8.5 Conclusion

8.6 Summary

9 CYCLEGAN

9.1 Introduction

9.2 Image-to-Image Translation

9.3 Cycle Consistent Loss: there and back aGAN

9.4 Adversarial Loss

9.5 Identity Loss

9.6 Architecture

9.7 CycleGAN architecture: building the network

9.8 Generator architecture

9.9 Discriminator architecture

9.10 Object Oriented Design of GANs

9.11 Tutorial: CycleGAN

9.12 Building the network

9.13 Running CycleGAN

9.14 Expansions, augmentations and applications

9.15 Applications

9.16 Summary

10 ADVERSARIAL EXAMPLES

10.1 Introduction

10.2 Context of Adversarial Examples

10.3 Lies, Damned Lies and Distributions

10.4 Use and abuse of training

10.5 Signal and the noise

10.6 Not all hope is lost

10.7 Conclusion

10.8 Summary

11 PRACTICAL APPLICATIONS OF GANS

11.1 Introduction

11.2 GANs in Medicine

11.2.1 Using GANs to Improve Diagnostic Accuracy

11.3 GANs in Fashion

11.3.1 Using GANs to Design Fashion

11.4 Conclusion

11.5 Summary

12 LOOKING AHEAD

12.1 Introduction

12.2 Ethics

12.3 GAN Innovations

12.4 Relativistic GAN (RGAN)

12.4.1 Application

12.5 Self-Attention GAN (SAGAN)

12.5.1 Application

12.6 BigGAN

12.6.1 Application

12.7 Further reading

12.8 Looking Back & Closing Thoughts

12.9 Conclusion

12.10 Summary

APPENDIXES

APPENDIX A: TECHNICAL/DEPLOYMENTS

作者介紹

Jakub Langr畢業(yè)于牛津大學(xué)，并在牛津大學(xué)的計(jì)算服務(wù)院任教。他從2013年開(kāi)始從事數(shù)據(jù)科學(xué)工作，最近開(kāi)始擔(dān)任Filtered.com的數(shù)據(jù)科學(xué)技術(shù)主管和Mudano的研發(fā)數(shù)據(jù)科學(xué)家。Jakub是曼寧出版社出版的《GANs in Action》的合著作者，這本書(shū)是市場(chǎng)上目前唯一一本關(guān)于生成對(duì)抗網(wǎng)絡(luò)（GAN）的書(shū)。Jakub還在伯明翰大學(xué)參與設(shè)計(jì)并教授數(shù)據(jù)科學(xué)課程。

弗拉基米爾·博克(Vladimir Bok)是Intent Media的高級(jí)產(chǎn)品經(jīng)理。Intent Media是一家為領(lǐng)先旅游網(wǎng)站提供數(shù)據(jù)科學(xué)服務(wù)的公司，他幫助監(jiān)管該公司的機(jī)器學(xué)習(xí)研究和基礎(chǔ)設(shè)施團(tuán)隊(duì)。在此之前，他是微軟的項(xiàng)目經(jīng)理。弗拉迪米爾以?xún)?yōu)異成績(jī)畢業(yè)于哈佛大學(xué)計(jì)算機(jī)科學(xué)專(zhuān)業(yè)。他曾在早期的金融科技公司擔(dān)任軟件工程師，包括PayPal聯(lián)合創(chuàng)始人Max Levchin創(chuàng)辦的一家公司，以及Y Combinator初創(chuàng)公司的數(shù)據(jù)科學(xué)家。