Machine Learning Refined: Foundations, Algorithms, and Applications

by Jeremy Watt, Reza Borhani, Aggelos Katsaggelos

With its intuitive yet rigorous approach to machine learning, this text provides students with the fundamental knowledge and practical tools needed to conduct research and build data-driven products. The authors prioritize geometric intuition and algorithmic thinking, and include detail on all the essential mathematical prerequisites, to offer a fresh and accessible way to learn. Practical applications are emphasized, with examples from disciplines including computer vision, natural language processing, economics, neuroscience, recommender systems, physics, and biology. Over 300 color illustrations are included and have been meticulously designed to enable an intuitive grasp of technical concepts, and over 100 in-depth coding exercises (in Python) provide a real understanding of crucial machine learning algorithms. A suite of online resources including sample code, data sets, interactive lecture slides, and a solutions manual are provided online, making this an ideal text both for graduate courses on machine learning and for individual reference and self-study.

Genres: Artificial Intelligence, Programming, Nonfiction, Mathematics, Science, Computer Science

544 pages, Hardcover

Published February 29, 2020

Reviews

[Romann Weber · Rating 5 out of 5]

Dense but reader friendly, "Machine Learning Refined" is a well-written and handsomely illustrated book that covers an enormous amount of fundamental material in a relatively light page count. In a number of ways, it is a spiritual cousin of and worthy complement to the also-excellent "Learning from Data" by Abu-Mostafa, Magdon-Ismail, and Lin, which covers similar material more from the standpoint of statistical learning theory (think VC dimension).

Rather than exhaustively cover every variation of the various machine-learning algorithms out there, the authors wisely choose to focus on themes common to them all. In particular, I appreciated the recurring theme of basis functions as they relate to both obvious cases (polynomial and Fourier bases, etc.) and more abstract ones (namely neural networks, where increasing depth increases the "flexibility" in one's bases).

There is indeed some heavy lifting in the book, making some sections (or even entire chapters) less essential for complete digestion on a first reading. That said, none of it is pedantic, and the extra space devoted to some of the lengthier derivations is bound to be appreciated by those who have encountered skipped steps or dreaded left-to-the-reader punting in other texts.

Were I to design the book, I might reorder some of the material (I'd not have put the dimensionality-reduction chapter last, for instance), but this is hardly a major complaint. I hope this book catches on and that the authors stay invested enough to prepare future editions.

[Kimiaki Shirahama · Rating 5 out of 5]

This is one of the greatest books to study the fundamentals of machine learning. Especially, from the perspective of minimising a cost function, the authors provide a consistent view of various machine learning methods such as regression, classification, feature learning (deep learning), and so on. In addition, each method is not only explained by intuitive figures but also demonstrated with python jupyter-notebook codes, so as to easily get the main idea/concept of the method. I think this book serves as a good starting point for machine learning beginners, and also it is useful for experienced people to refresh the core idea of machine learning.

[Estefano Palacios · Rating 5 out of 5]

Gold

There are hundreds of books on the topic of machine learning. They belong two sets: heavy on math or so lightweight that.machine learning seems like witchcraft. This books strikes a balance by teaching !machine learning rigorously but from first principles. It is self contained, and introduces all the math you'll need for the deed. I haven't finished it yet but I am pretty far on the book and so far It's been a gem. Look for the reviews of the previous edition and you'll get a.feel.on the.quality of this book.

[Jasneet Singh · Rating 5 out of 5]

2nd edition review: Easy to read. I am a beginner, and the book uses analogies to drive home the point. It has an appendix section at the end that goes into more depth on the mathematical foundations required. And a Github page from wherein we can have the boilerplate code for doing the exercises. This is my first book on Machine Learning, and I appreciate the authors for a gentle introduction. Now, I have confidence that I can navigate through ML concepts. It serves as a good starting point in the journey.

[Ogi Ogas · Rating 2 out of 5]

My ratings of books on Goodreads are solely a crude ranking of their utility to me, and not an evaluation of literary merit, entertainment value, social importance, humor, insightfulness, scientific accuracy, creative vigor, suspensefulness of plot, depth of characters, vitality of theme, excitement of climax, satisfaction of ending, or any other combination of dimensions of value which we are expected to boil down through some fabulous alchemy into a single digit.

[Purbesh Mitra · Rating 5 out of 5]

Good book for learning introductory Machine Learning. The mathematics is rigorous and yet lucidly enjoyable. Deservs full rating in my opinion.