Hands-On Generative Adversarial Networks with Keras
上QQ阅读APP看书,第一时间看更新

Foreword

I have known and worked with Rafael for about two years. Rafael is an expert in machine learning and deep learning in various tasks, and is particularly well known for his work in speech synthesis. In this book, Rafael will take you into a new world full of interesting things a Deep Neural Network (DNN) can do.

For those who are not familiar with DNNs, this book explains the basics and helps you set up the environment to train your first neural network. It then goes on to introduce a very popular class of DNNs, called Generative Adversarial Networks (GANs). This book goes through the principal ideas behind GANs, how to train and evaluate your first GANs, and the problems you might encounter while training them. To improve your results and stabilize your training procedure, you might find several tricks listed in the book to be very useful. Finally, the book shows several applications of GANs in different areas, including computer vision, natural language processing, and speech processing. Experience the fun and joy of turning simple lines of code into tangible images or audio.

The book approaches the concepts of GANs in both a mathematical and an intuitive way. After you have completed the book, you will be familiar with the GANs that are commonly used nowadays, as well as their different use cases and their potential impacts. You will also get to know the mechanism of training GANs, the current limitations of GANs, and the future of GANs. Most importantly, the book provides extensive hands-on examples to really help you implement everything from the ground up, so you can learn from your own experiences.

Get ready for the journey to see the fascinating world built by GANs.

 

Ting-Chun Wang

Senior Research Scientist, NVIDIA