Understanding Machine Learning: From Theory to Algorithms

by Shai Shalev-Shwartz, Shai Ben-David

Machine learning is one of the fastest growing areas of computer science, with far-reaching applications. The aim of this textbook is to introduce machine learning, and the algorithmic paradigms it offers, in a principled way. The book provides a theoretical account of the fundamentals underlying machine learning and the mathematical derivations that transform these principles into practical algorithms. Following a presentation of the basics, the book covers a wide array of central topics unaddressed by previous textbooks. These include a discussion of the computational complexity of learning and the concepts of convexity and stability; important algorithmic paradigms including stochastic gradient descent, neural networks, and structured output learning; and emerging theoretical concepts such as the PAC-Bayes approach and compression-based bounds. Designed for advanced undergraduates or beginning graduates, the text makes the fundamentals and algorithms of machine learning accessible to students and non-expert readers in statistics, computer science, mathematics and engineering.

Genres: Computer Science, Artificial Intelligence, Textbooks, Mathematics, Science, Programming, Technical

416 pages, Kindle Edition

First published April 30, 2014

Reviews

[Aroosha Dehghan]

این کتاب خوبیه
اگر همزمان کلاس‌های شای بن دیوید تو یوتیوب رو هم ببینید خیلی بهتره.
همین کتاب رو تدریس می‌کنه

[Theivendiram Pranavan]

Most of the core theories in ML are covered. It is suitable for CS enthusiastic people.

[Fasil]

Definitely my machine learning bible. Far too many fields in this blossoming field have too much emphasis on the application and construction of algorithms and very few actually go in depth in the theory behind what makes a ML algorithm work. I would say every ML researcher should read this book once or at least have it as a reference.

[Abbas Mehrabian]

Excellent introduction to Machine Learning Theory. I have written a Persian review: http://bi-nahayat.persianblog.ir/post...

[Alireza Aghamohammadi]

مخاطب
کتاب درک یادگیری ماشین به مباحث نظری یادگیری ماشین می‌پردازد. مخاطبین این کتاب افرادی هستند که علاقمند به درک مباحث ریاضی پشت الگوریتم‌های یادگیری هستند.

محتوا
کتاب ابتدا با مفهوم یادگیری آغاز می‌کند و اینکه چطور می‌توان آن را به صورت احتمالی و محاسباتی بیان کرد. به همین منظور مدلی به نام Probably Approximately Correct یا همان PAC معرفی می‌کند. در باقی فصل‌های کتاب الگوریتم‌های یادگیری ماشین مانند درخت تصمیم، نزدیک‌ترین همسایه و ... را از منظر مدل PAC بررسی می‌کند.

یکی از بخش‌های جالب کتاب قضیه ناهار مجانی و اثبات آن بود و نتیجه‌ای که از آن می‌توان گرفت (حداقل در دنیای یادگیری ماشین):

There ain't no such thing as a free lunch

نویسنده دوم کتاب (بن دیوید) ویدیو‌های تدریس خود در دانشگاه واترلو را قرار داده است که از این لینک می‌توانید مشاهده کنید. البته ویدیوها فقط نیمه اول کتاب را در بر می‌گیرد.

جمع‌بندی
کتاب متن سلیسی دارد و خواندنش علی رغم ریاضیات سنگین آن برای بنده بسیار لذت‌بخش بود. اگر می‌خواهید درک عمیق‌تری نسبت به یادگیری ماشین داشته باشید، این کتاب فوق‌العاده است.

[Aadesh]

Didn't get to finish all the chapters of the book in details but did read first few chapters minutely. The approach of the book was beneficial for me to grasp the concept of PAC and its implication in learning theory.

[Ivan]

This was a pretty long project and a lot of fun once I deferred working through the proofs (I'm curious to take another stab at a deeper level and won't be surprised to have appetite for one more after)

Desireable outcomes when designing the algorithms, practical consequences and the lineage are very approachable otherwise

[Rohan Dalvi]

Used for a class as supplementary material. Think it would have been too dense without the class, but I thought it did a great job of conveying the theory behind ML in a way that was not overly abstract.

[Lan]

very good fundamentals

[Lucas Chen]

it was pretty good i don't feel like i need to read esl