home |  lectures |  tutorials |  assignments

Machine Learning and Pattern Recognition

INTRODUCTION TO THE COURSE

Lecturer: Charles Sutton
TA: Victor Hernandez-Urbina < j.v.hernandez-urbina AT ed.ac.uk >
Location: Tuesday, 09:00 - 09:50, Appleton Tower LT5
Location: Friday, 09:00 - 09:50, Appleton Tower LT3
Office Hours: Tuesday and Friday 09:50 - 10:50. You may ask me any question about the course. I will sit at a table in the cafe in the AT lobby, drinking some kind of caffeinated beverage.

Welcome! This course examines the principles, methods and application of machine learning: the business of utilising data to enable computers to perform tasks that would be infeasible to do well by direct programming. The course aims to provide a firm grounded understanding of the field and a knowledge of the methods that can be used to design machine learning tools. Alongside these fundamentals, I will teach about the theory and practice of a core set of machine learning tools, and students will gain practical experience using these tools. Please see this handout for more detailed information.

Preparatory Work

Students with little mathematical background would be well advised to do some preparatory work before the course begins.

Mathematics

Some of you may struggle with the mathematics in this course. Here is more information about the mathematical background assumed.

Discussion Forum

We will be using Nota Bene as the discussion forum for this course. You can sign up for the forum here. The TA and I will be monitoring the site regularly. I urge you to use the discussion site to ask any questions about the course.

To encourage you to use the site, we will have several new policies this year:

1. By default, I will respond to all questions I get over email about the course content by posting the question and answer anonymously onto the discussion site. If your question is of a clearly confidential nature, I will obviously not post it to the discussion forum! Additionally if you prefer that I answer your question privately, please mention this in your email and I will honour your request.
2. The lecture notes of the course will be made available only through NB. You are welcome to download them from NB to your own computer.
3. NB allows you to click on any of the slides and leave a short comment, e.g., "Could you explain this slide again?" I will monitor all of the topics that students have asked me to explain again, and every week I will pick one question to answer in class. So you can ask me to explain slide 33 again, and I will! I will aim to do this first thing on Tuesday's lecture, every week from Week 3 onwards.

Books

The course text will be Machine Learning: A Probabilistic Perspective. Kevin P Murphy, MIT Press, 2012.

A similar recent book, which is also excellent, is Bayesian Reasoning and Machine Learning. by David Barber, 2012. It is available as a free pdf online.

Either one of the above two books is sufficent to do well in this course. For those who want to read even more, I'd also suggest:

David MacKay's book, "Information Theory, Inference and Learning Algorithms" is a classic book that gives a different perspective on many of the methods in this course. It can be read online for free.

Finally, the book The Elements of Statistical Learning: Data Mining, Inference, and Prediction, Hastie, Tibshirani, and Friedman. (PDF available online.) is a perspective on machine learning by several leading statisticians. It is very much complementary to the course texts.

Exam

A mock exam paper is available. The length of this exam paper is a little longer than the real one would be, so you should not be expected to do this in the time available. You should use this paper for immediate study to familiarize yourself with the types of questions you might see and the newer subjects covered. I will also try to broadly follow the same types of questions used in previous years' exam papers.

These pages are maintained by Charles Sutton
.

Home : Teaching : Courses