This course is structured around algorithmic issues in machine learning. Modern machine learning systems are frequently built on top of algorithms with no provable guarantees, and it is debatable when and why they work. In this class, we will concentrate on developing algorithms whose performance can be rigorously evaluated for fundamental machine learning problems.
The book is based on the MIT course “Algorithmic Aspects of Machine Learning,” which was taught in the fall of 2013, spring of 2015, and fall of 2017.
This open book is distributed under the terms of the Open Publication License (OPL). You can get the Algorithmic Aspects of Machine Learning ebook in PDF format for free (1.5 MB).