Probabilistic and Unsupervised Learning
Approximate Inference and Learning in Probabilistic Models

Dates Mondays & Thursdays
29 September - 11 December 2014
Time Mondays and Thursdays 11:00-13:00
Tutorials Wednesdays 9:00-10:20
Lecturer Maneesh Sahani
TAs Heiko Strathmann, Wittawat Jitkrittum, Gergo Bohner
Location Basement Seminar Room (B10), Alexandra House, 17 Queen Square, London WC1N 3AR
About the course

This course (offered as two successive modules to MSc students) provides an in-depth introduction to statistical modelling, unsupervised, and some supervised learning techniques. It presents probabilistic approaches to modelling and their relation to coding theory and Bayesian statistics. A variety of latent variable models will be covered including mixture models (used for clustering), dimensionality reduction methods, time series models such as hidden Markov models which are used in speech recognition and bioinformatics, Gaussian process models, independent components analysis, hierarchical models, and nonlinear models. The course will present the foundations of probabilistic graphical models (e.g. Bayesian networks and Markov networks) as an overarching framework for unsupervised modelling. We will cover Markov chain Monte Carlo sampling methods and variational approximations for inference. Time permitting, students will also learn about other topics in probabilistic (or Bayesian) machine learning.

The complete course forms a component of the Gatsby PhD programme, and is mandatory for Gatsby students. The two component modules are also available to students on the UCL MSc in Machine Learning and UCL MSc in Computational Statistics and Machine Learning. The first part (GI18: Probabilistic and Unsupervised Learning) may be used to fill a core requirement in each Masters programme. The second part (GI16: Approximate Inference and Learning in Probabilistic Models) is an optional module, but is only available to students who have completed GI18.

General attendance

Ordinarily, students, postdocs and faculty from outside the unit and these programmes are also welcome to attend. However space may be very limited this year and we will need to operate a waiting list. Please contact the unit to be added to the list and we will let you know if and when space becomes available in the lectures. Unless you are a Gatsby student or fellow, or a registered student in the ML or CSML MSc programmes, please do not attend the lectures until you have received confirmation from us that you may.


A good background in statistics, calculus, linear algebra, and computer science. You should thoroughly review the maths in the following cribsheet [pdf] [ps] before the start of the course. This quiz may help you check where you stand. You must either know Matlab or Octave, be taking a class on Matlab/Octave, or be willing to learn it on your own.


There is no required textbook. However, the following are excellent sources for many of the topics covered here.

David J.C. MacKay (2003) Information Theory, Inference, and Learning Algorithms, Cambridge University Press.
Christopher M. Bishop (2006) Pattern Recognition and Machine Learning, Springer-Verlag.
David Barber (2012) Bayesian Reasoning and Machine Learning, Cambridge University Press.
Kevin Patrick Murphy (2012) Machine Learning: a Probabilistic Perspective, MIT Press.
Daphne Koller and Nir Friedman (2009) Probabilistic Graphical Models, MIT Press. (This contains a more extensive treatment of graphical models, good for reference)
Reading material

Specific recommendations from previous years can be found here. New material may be assembled as course progresses.

Lecture schedule and slides (provisional) Lecture slides and assignments will be posted here as the course progresses. Note that the content and sequence of lectures may vary slightly from those of previous years. Also note that, following the usual Gatsby teaching pattern, there are lectures scheduled during "reading week".


29/9: Introduction to Probabilistic Learning / screen version 2/10: contd.
6/10: Latent Variables / screen version 9/10: contd.
13/10: Expectation Maximisation / screen version 16/10: Latent chain models / screen version
20/10: contd. 23/10: Graphical models / screen version
27/10: contd 30/10: Model selection and GPs / screen version
3/11: no lecture 6/11: extra lecture (if needed)
10/11: Intractable models / screen version 13/11: contd.
17/11: Variational methods / screen version 20/11: Expectation Propagation / screen version
24/11: contd. 27/11: Belief Propagation / screen version
1/12: Convexity-based Methods / screen version 4/12: Sampling / screen version
8/12: contd. 11/12: contd.
50% of the total mark (for each module) is based on coursework assignments (the other 50% being based on the final written examination).

All assignments (coursework) are to be handed in to the Gatsby Unit, not to the CS department. Please hand in all assignments at the beginning of lecture on the due date to the lecturer or to the TAs. Late assignments will be penalised. If you are unable to attend the lecture, you can also hand in assignments to Mr. Barry Fong at the Alexandra House 4th floor reception.

Late Assignment Policy: we will be using the official UCL-wide late policy. Marks will be reduced by 5 for assignments up to one day late, 15 for assignments between one and seven days late (or before the answers are discussed in tutorial, whichever comes first), and will receive no credit thereafter.

Assignment 0 - maths quiz to be discussed in tutorial 1 Oct 2014
Assignment 1        [binarydigits.txt,  bindigit.m] due 13 Oct 2014
Assignment 2        [geyser.txt ssm_spins.txt ssm_spins_test.txt ssm_kalman.m] due 27 Oct 2014
Assignment 3        [co2.txt] due 10 Nov 2014
Assignments 4 and 5        [images.jpg, genimages.m, Mstep.m] due 4 Dec 2014
Assignment 6        [message.txt, symbols.txt, asst6code.tgz] due 18 Dec 2014
To attend Please see the note above