The Hundred-Page Machine Learning Book

The advantage of short books like this is that if they are well written the author has to think carefully about what to write and how to write it. That's certainly been done here.After a crash course in what ML is and some mathematical notation, a few popular ML algorithms are introduced, before Burkov takes a look at what a learning algorithm fundamentally does: optimising a particular function (normally by minimising a loss function).Other parts of the book go into ML practice, deep learning, practical problems and solutions, and tips and tricks for situations you might run into (e.g. handling multiple outputs). Unsupervised learning, word embeddings and ranking and recommendation systems are discussed. The book's conclusion talks about other areas to learn about which weren't present.The book is dense in parts, no doubt about it. Burkov lays down all the mathematical formulae but also explains things pretty well and touches on the intuition behind key ideas, along with useful pictures and diagrams.That is one of the things I liked the most: it is rigorous, concise, but not unclear. Another thing I really liked is that it touches on very practical problem discussed less frequently elsewhere (e.g. imbalanced datasets) and interesting approaches you won't find in more traditional resources (like one and zero shot learning).In contrast to what some other reviewers on the back of book say, I'd say that this book is probably not the best one for absolute beginners. It would be much more useful when you know what ML is and have done a project or two, at least.To sum up, if you want an information packed ML book that has both theory and useful practical tips, read this.

This book is one of the best books I have read on machine learning. It’s beautifully written with concise and clear explanations. The author does an amazing job in only communicating the necessary on such a broad and deep project. I got the hard copy and it’s a pleasure to have. Thank you

This books price is a shame. Aside from that the content is good for the most part. Sadly it doesnt explain back propagation which would have been nice and theres no gaussian section which seemed odd. The best part about this book for me is its one of the few that actually explains the notation properly. I find that this subject appears a lot more difficult because of the dense notation which many books go out of their way not to define. This one does a good job of making sure you understand what all the letters and subscripts mean, and for that I was very happy

This is not the book you get for sample code and immediate applications, but it is a fantastic resource to learn more of the theory behind machine learning methods. You will improve your use of models by learning the background in this book.

The book is extremely comprehensive with the knowledge, but it's more than enough to know the basics, better take this one, than much longer but empty in context books.

This is an excellent brief but in-depth introduction to the subject for complete beginners who have a mathematical background. In the first 6 pages it explains from very basic principles to producing a complete machine learning model using one technique. It then explains other techniques, including multi-level neural networks. It is a remarkably easy read considering the level of detail it goes into. I found it an excellent first book on the subject.

For the most part, I liked the short and concise explanations. They were so concise I found my self reading and rereading sentences simply because there was so much information condensed into them. I disliked the treatment of backpropagation, which was almost non-existent and the explanation of convolutional neural networks was difficult to follow -despite the fact that I know how these networks work. Overall I feel that this is a good book to read if you have already had a healthy introduction to machine learning from other sources but there is no getting away from the fact that it is a little too short. The price also is a little high for such a slim book.There are some dreadful books about machine learning doing the rounds at the moment. This book is not one of them.

Difficult to believe but this book describes a variety of machine learning concepts and algorithms in just 136 pages. Of course it lacks of applied machine learning paradigms but there are plenty of books out there to improve your practical skills - e.g. Hands On Machine Learning with Scikit-Learn, Keras & Tensorflow. If you are a beginner on the field this book looks challenging but after you grasp the key concepts you will know how thinks work! On the other hand experienced data scientist and machine learning engineers can refresh their knowledge or even self-improve. Lastly, I really enjoyed QR codes which provide additional material which is constantly up to date.