Computer vision expert Adrian Rosebrock recently wrote a useful list of seven books for beginners to deep learning and, most importantly, the best books to read.

Some of these books are theoretical and focus on the mathematics and assumptions behind neural networks and deep learning. Others are entirely practical, teaching you deep learning through code rather than theory.

Then there are books that combine theory and practice, giving you the theoretical knowledge and implementing the algorithms to learn by yourself. (Who doesn’t love that?)

Below we talk about what each book is about, the target audience, and whether it’s the right book for you.

Before choosing a book, it’s best to assess your own personal learning style, which will enable you to make the most of it and get the most out of it.

Start by asking yourself the following questions:

What is the best way for me to learn? Do I prefer to get my knowledge from theoretical texts? Or would you rather learn from snippets and implementations?

Everyone has their own unique learning style, and your own best learning style will determine which books you should read.

For example, for some people, they prefer a balance between theory and practice, so it is good to read books that combine theory and practice. Deep learning books that are too theoretical or abstract will only make them feel boring, or they will fall asleep. On the other hand, if a deep learning book skips theory entirely and goes straight to concrete code implementation, the reader will miss the core theoretical foundation that can help us solve new deep learning problems or projects. As far as they are concerned, a good book needs to strike a balance between the two.

We need theories to help us understand the core foundations of deep learning, as well as applications and code to help us deepen what we have learned.

Book 1 — Deep Learning

If you want to write a blog about the best books on Deep Learning, you have to mention Deep Learning by Goodfellow, Bengio, and Courville. Deep Learning by Zhao Shenjian, Li Yujun, Fu Tianfan and Li Kai.

This is a college textbook that teaches the basic principles and theories of deep learning. Deep Learning, by Goodfellow et al., is a purely theoretical book, aimed at an academic audience, with no code in it.

The book begins with a discussion of the fundamentals of machine learning, including the applied mathematics necessary to learn deep learning (linear algebra, probability theory, information theory, etc.) from an academic perspective.

Then, modern deep learning algorithms and techniques are discussed in depth. In the end, the book focuses on the current research trends of deep learning and the new trends in the field of deep learning.

You can read the ebook for free on the book’s website, or purchase the physical book yourself.

You should read this book if:

  • You prefer theoretical knowledge to practical knowledge

  • Love academic works

  • You are a professor, undergraduate, or graduate student doing deep learning research

Book 2 — Neural Networks and Deep Learning

The second book on Deep Learning theory to recommend is Michael Nielsen’s Neural Networks and Deep Learning.

There are seven Python code pieces in the book, which use MNIST data sets to explain the basics of various machine learning, neural networks, and deep learning techniques and go a long way toward illustrating the theoretical concepts described in the book.

If you are new to machine learning and deep learning and eager to get into the theoretical realm, this book should be your first choice.

The book is much easier to read than Goodfellow’s Deep Learning, and Nielsen’s writing style and code snippets make it easier to finish.

You can read the electronic version of this book for free on the official website, and you can find the corresponding Chinese version resources on the website.

You should read this book if:

  • You’re looking for a theory book on deep learning

  • You are new to the field of machine learning or deep learning and prefer to learn more about the field from an academic perspective

Book 3 — Deep Learning with Python

Francois Chollet, a Google AI researcher and author of Keras, a popular and popular Deep Learning library, wrote Deep Learning with Python in October 2017.

This book deals with deep learning from a practitioner’s point of view, and although there is some theoretical knowledge in the book, every few paragraphs will teach you how to use Keras to implement related technologies.

Francois provides many examples of applying deep learning to computer vision, text, sequences, etc. The book is very comprehensive for readers who want to learn about Keras as well as machine learning and deep learning.

Not only is the book concise and easy to read, but some of the author’s observations on the trends and history of deep learning are also impressive.

It is important to note that this book is not an in-depth deep learning book, but rather teaches you the basic concepts of deep learning by writing various practical deep learning examples using the Keras library.

You should read this book if:

  • You are very interested in the Keras library

  • You prefer to learn by doing

  • Do you want to quickly understand how deep learning is applied to different fields, such as computer vision, sequence learning, and text analysis

Book 4 — Hands-on Machine Learning with Scikit-Learn and TensorFlow

Some people buy Aurelien Geron’s Hands-on Machine Learning with SciKit-Learn and TensorFlow for the first time, not quite sure what to Learn, just as a basic introduction to Machine Learning, If it weren’t for “TensorFlow” in the title, you’d probably ignore it completely.

For example, some people think that adding TensorFlow to the end of an already long title is a marketing gimmick to increase circulation, since there are so many people interested in deep learning, right?

But it would be wrong to think so. It’s a very good book, and you can’t judge it by its cover.

The book is divided into two main parts.

  • The first part covers basic algorithms for machine learning, such as support vector machines, decision trees, random forests, integration methods, and some basic unsupervised learning algorithms, each with accompanying Scikit-Learn examples.

  • The second part explains the basic concepts of deep learning through TensorFlow library.

You should read this book if:

  • You are new to machine learning and want to introduce yourself to the core principles of machine learning through code examples

  • Interested in the popular SciKit-Learn machine learning library

  • Want to quickly learn how to complete basic deep learning tasks using the TensorFlow library

TensorFlow Deep Learning Cookbook

This book is completely hand by hand and is a very good reference book for TensorFlow. It does not teach deep learning, but shows you how to use the TensorFlow library in deep learning.

Don’t get me wrong — you’ll definitely learn new deep learning concepts, techniques, and algorithms with this book, but it takes a more hands-on approach: lots of code and explanations of it.

The only downside is that there are a lot of typos, which is to be expected in a code-centric book. Typos cannot be avoided, so be careful when reading.

You should read this book if:

  • You’ve learned the basic concepts of deep learning

  • Interested in the TensorFlow library

  • A hands-off approach that likes to provide code that solves problems but doesn’t care about the underlying theory

Book 6 — Deep Learning: A Practitioners Approach

In the first few chapters of the book, Gibson and Patterson discuss the basics of machine learning and deep learning, while the rest of the book covers Java deep learning code written using the DL4J library.

You should read this book if:

  • You need to use the Java language in your daily work and study

  • Your company or organization uses Java primarily for programming

  • Do you want to know how to use DL4J libraries

Book 7 — Deep Learning for Computer Vision with Python

Deep Learning for Computer Vision with Python, written by Computer Vision expert Adrian Rosebrock, is rated as one of the best Deep Learning and Computer Vision resources available today.

Francois Chollet, AI researcher at Google and author of the Keras Library, describes the book as an excellent, deep and practical deep learning exercise in computer vision. I found it very easy to read and understand: the explanations were clear and detailed. You’ll find a lot of practical advice that you won’t find in other books or college courses. For practitioners and beginners, I highly recommend Francois Chollet

If you’re interested in applying deep learning to computer vision (image classification, object detection, image understanding, etc.), this is a great book.

In this book, you will be able to: Advanced deep learning technologies, including object detection, multi-GPU training, transfer learning, generative adversarial networks, etc. Includes ResNet, SqueezeNet, VGGNet, and others that exist in the ImageNet dataset

In addition, there is a balance between theory and practice. For each deep learning theory, there is an associated Python implementation to help you consolidate your understanding and what you have learned.

You should read this book if:

  • You have a particular interest in applying deep learning to computer vision and image understanding

  • Your preferred way of learning is to combine theory with practice

  • You want a deep learning book that makes complex algorithms and techniques easy to understand

  • You want to have a clear book that guides you through the mysteries of deep learning

conclusion

Here we discuss seven books in the field of deep learning and who might read them.

Of course, if you want to check out other deep learning resources besides books, don’t miss the deep Learning in Depth column series on Jizhi Main website:

Shallow deep learning series

And the Deep Learning series:

Play deep Learning

If you are truly white, but want to learn about AI in the most effective way, then read the following text carefully:

The 6th online training course for AI Engineers of Jizhi will begin soon!

  • Weekly live teaching, with consolidation of homework, weekend correction + explanation, help you from zero to master, down-to-earth AI engineer.

  • Famous teachers teach. Engineers from domestic first-line enterprises, technical leaders, doctors from prestigious universities, and reviewers from international journals gave lectures.

  • Applicable people: junior high school or above with mathematics background, any programming background is acceptable. We have every confidence that you can go from an AI novice to a tech expert.

For details and lecture clips, please click herehere, detailed understanding!

Seats are limited! Limited time group purchase! The recommendation has the prize!

Now or never, waiting for you!