Wolf M. Mathematical Foundations of Supervised Learning 2023

Textbook in PDF format

What is machine learning? Machine learning is often considered part of the field of artificial intelligence, which in turn may largely be regarded as a subfield of computer science. The aim of achine learning is to exploit optimization techniques and, in most cases, vast amounts of data in order to devise complex models or algorithms in an automated way. Loosely speaking, it is about roducing computer programs without writing them. Instead, one sets up an optimization procedure, which in this context is often called ‘learning’ or ‘training’, that eventually leads to the sought omputer program. Machine learning techniques are typically used whenever large amounts of data are available and when one aims at a computer program that is (too) difficult to program ‘directly’. Standard examples are programs that recognize faces, handwriting, or speech, drive cars, recommend products, translate texts or play Go. These are hard to program from scratch so one uses machine learning algorithms that produce such programs from large amounts of data.
Introduction.
Learning Theory.
Neural networks.
Kernel methods.
Probability theory