Machine Learning Theory

Lecturer: James Worrell (University of Oxford).

About the lecturer | Course Summary | Slides | Assignment

About the lecturer: James Worrell is a Professor of Computer Science at the University of Oxford. His main research interests lie in quantitative verification, including probabilistic, real-time and hybrid systems, various forms of quantitative automata, linear recurrence sequences and linear differential equations. He currently holds an EPSRC Leadership Fellowship on the subject of Verification of Linear Dynamical Systems.

Course summary: Computational learning theory aims to develop rigorous mathematical foundations for machine learning, in order to provide guarantees about the behaviour of learning algorithms and to understand the inherent difficulty of learning problems.

This course will introduce a theoretical framework for classification problems in the supervised setting and will analyse the support vector machines algorithm within this framework. Specific topics include the PAC model, VC dimension, Rademacher complexity, support vector machines, and margin theory.

Lecture slides can be found here.

Assignment: You can find the problem set here. Please send the solutions to Bartek Klin (klin@you-know-what) by Thursday, March 17 (extended!).