Artificial Intelligence for Humans, Volume 3: Deep Learning and Neural Networks

Definitly worth its price for the content and the GitHub examples. The content is pretty minimalistic in of itself and is insufficient to have a full understand of deep learning, but there are some really good pointers and simplified explanations to help you get there. I enjoyed the fact that each concept was presented seperatly which helped me get a better undestanding on how neural networks works and what is there underlying philosophie.

The book gives a brief description of the large number of types and techniques for applying neural networks, but never gives the reader any examples of their implementation and how effective they are at solving any of the classic problems. Instead the reader is fed trivia like the definitions of sgn(x), mean, standard deviation, partial derivative or the amazing fact that the second derivative is the derivative of the first derivative. There is nothing like a tutorial anywhere in the book.

The book is more like a quick compilation of a college student's note. Concepts are presented in a relatively isolated manner; connections between concepts are, for the large part, missing. Furthermore, if the materials presented are rather shallow like in this book, readers will expect to see a strong emphasis on, or hands on exercises of, practical applications. But this book doesn't seem to help much in that regard either, despite what the book claims. The book does give introduction to a bunch of models, which can be useful for a beginner. But at least this edition I wouldn't suggest any one to buy because of poor editing. For example, the pseudo code fragments listed in the book are often wrong, or different than what is described by the text. References to formulas, etc need some proof reading, too.

I ran through the book quickly (kindle) and it seems good although I expected much more mathematics involved (Integrals, derivatives, linear algebra). I hope that a new version or a new book will go more in depth with the actual problems...

Really like this book explaining the state of the art in Neural Networks. I first encountered Jeff Heaton's work when i was looking for a neural network to predict forex on the mt4 platform, we used encog at that time. Jeff does not only have a high level view over neural networks, he knows also how to build them from ground up.

Jeff is one of the best ANN experts out there. The content is easy to digest and not heavy on the math. Great primer to get used to concepts before diving deeper.

Simple introduction. But examples could be unclear. Eg when introducing convolution layers, lack of explanation on the choice of dimension.

A good one for beginner, but not enough if you really want to understand the ideas. I still recommend it.

I have very mixed feelings about this book.Having read most of Heaton's papers, articles and seen most of his videos, and I have to admit I was disappointed.The pros:Covers a wide range of networks, a few training algorithms, and so on.If you're a student just starting on ANN, then this is a good book to get your self acquainted with the concepts.You may even be able to write your own network if you do some follow-up searching.The bizarre stuff:This book reads like an ode to Heaton. Was it really necessary to say that Heaton did this and that?For example at some point the book claims that Heaton introduced convolutional nets, which is of course, not true.Later they mention LeNet and rectify the earlier statement.Would a quality book allow for such claims? I leave that up to you to decide.The mediocre stuff:All source code samples are in Python. I would have preferred Pseudo-code, but there are some source code listingswhich you can use to see basic versions of algorithms.Math formulae are not thoroughly explained, which is what you would expect from a beginners book.The print quality is poor, in fact the book mentions heat-maps and colours, whilst the actual print is in Black and White!The bad stuff:Heaton barely scratches the surface.This book reads like an introduction or summary to what neural nets are.Some subsections are a paragraph long (!!!) and then move on to the next subject.It feels as if they are trying to fill pages, rather than provide substantial material.I guess for the price I shouldn't really complain, but you could have read an e-book or online tutorial and gotten the same information.Overall, treat this book as a very gentle and brief introduction and you won't be disappointed.

This is a subject I have been personally researching and reading from a number of texts, both for programmers and for the general interest. Jeff Heaton's book help to under-cut many missed details in other types of texts. As I work through this book I feel I have the tools to now design and build my own artificial neural network to my requirements rather than ploughing through recipes in cookbooks that left wanting in the dark on a lot of details, unable to judge how to best to manipulate the pieces and components.This is not an advanced text for the professional; this is very much about building solid foundations to understanding.A thoroughly great book.

I felt compelled to write this review when I'm only about halfway through the book because it's that brilliant. For anyone who's has the ability to understand mathematical concepts, but doesn't necessarily have the formal education in mathematical notation, especially those from a programming background, will really appreciate this book. I've read screens upon screens of literature on AI and ANNs on the internet and they've all been, quite frankly, baffling. If that person is you too, this book will demystify these techniques and make you wonder why you were ever so intimidated in the first place.

Nice book.

Good