Neural Networks and Deep Learning: first chapter goes live

I am delighted to announce that the first chapter of my book “Neural Networks and Deep Learning” is now freely available online here.

The chapter explains the basic ideas behind neural networks, including how they learn. I show how powerful these ideas are by writing a short program which uses neural networks to solve a hard problem — recognizing handwritten digits. The chapter also takes a brief look at how deep learning works.

The book’s landing page gives a broader view on the book. And I’ve written a more in-depth discussion of the philosophy behind the book.

Finally, if you’ve read this far I hope you’ll consider supporting my Indiegogo campaign for the book, which will give you access to perks like early drafts of later chapters.


  1. As best as I understand deep learning has now been overtaken by sum-product networks. Will your book cover them too?

  2. That’s a very strong statement — what makes you think that?

    Sum-product networks are an example of a deep architecture, but at the moment I don’t plan to focus much on them. The philosophy of the book is to explain a few core ideas well. It’s inevitable that over time the details of techniques in vogue will change. But people who really master core ideas such as backpropagation will be well-placed to keep up with such changes. So, for instance, although sum-product networks are quite different to traditional neural nets, they can be trained using backprop.

  3. Michael, your new book looks fascinating. Hope I can find time to read it. I much enjoyed Reinventing Discovery. Regarding your focus on core concepts for the deep learning book, are you familiar with the idea of threshold concepts ? I think these will gain ground as those of us in education are obliged to reconsider what exactly is core knowledge in a fluid, interdisciplinary landscape. All best,

  4. Michael, first, thank you for the initiative to write a book on this interesting topic. With your background in quantum computation, will you be including any thoughts on neural networks quantum algorithm?

  5. Hi Kirill – Unfortunately, I don’t know of any interesting results about neural nets and quantum algorithms, so I won’t have anything to say.

  6. Hi, first i’d like to say thank you for this awesome explanation about how neural networks work, i already had studied this topic on this semester but i had no idea of how this really works and why. I’m not finished the chapter but i have a question, the answer of exercises will be available? I have made some of them but i’d like to be sure if i understand the concept really well. I will understand if you prefe choose to dont give the solutions, sometimes i agree with the idea of if you understand the concept 100% you will be sure about the correctness of your solutions. Anyway, thank you again from Brazil

    Filipe T Amaral

  7. I’m so excited about this book. It believe it will really fill my desire of truly learning about deep learning, which most papers/monographs in the field have not yet accomplished. In my opinion, one never actually fully understand any machine learning concept until he writes a code snippet of the particular algorithm solving a particular toy problem. Therefore, I can’t wait for this book! Could you please provide an estimate of the date it will be available to purchase (the whole final book)?

    Fred Sampedro

  8. Hi Michael,
    First of all, I would like to thank you for coming up with an article that helps the human neural nets to understand about the artificial ones 😀 Can’t wait for the book! It would be nice if you could temme about the release date of the book . Many congratulations in advance 🙂

  9. Hi Adithya,

    As I said above, I hope sometime in 2014. I don’t have a more precise estimate.

    (Nor will I be terribly shocked if it takes longer. I know of one well-known popular physics book whose author made a bet that he could finish his draft book in a few months. The bet was made in the 1970s. The book came out in the 1990s. So while I think I’ll get this done this year, who knows…)

  10. Hi Michael,

    I’m a college student interested in learning about and testing my own neural networks. I was familiar with the concepts you described in the first chapter but your explanations are the most well detailed and understandable that I have come across. I’m very excited for the following chapters! 🙂

    Thanks a lot for making this resource openly available!

Comments are closed.