COMPM056 - Graphical Models

This database contains the 2017-18 versions of syllabuses. Syllabuses from the 2016-17 session are available here.

Note: Whilst every effort is made to keep the syllabus and assessment records correct, the precise details must be checked with the lecturer(s).

Code COMPM056 (Also taught as COMPGI08)
Year4 (Masters)
Prerequisites

Excellent understanding and abilities with Linear Algebra, Multivariate Calculus and Probability all at first year undergraduate mathematics level.

Assignments will require students to have familiarity with coding a high level language; some tools in Matlab and Julia are provided.

Term1
Taught ByDmitry Adamsky [Teaching Fellow] (100%)
David Barber [Module Lead]
Aims

To give an introduction to probabilistic modelling covering the broad theoretical landscape.

The course aims to cover much of the first 12 chapters of the course textbook www.cs.ucl.ac.uk/staff/d.barber/brml/

The emphasis is on probabilistic modelling of discrete variables.

Learning OutcomesAbility to construct probabilistic models, learn parameters and perform inference. This forms the foundation of many models in the wider sciences and students should be able to develop novel models for applications in a variety of related areas.

Content

Bayesian Reasoning

Bayesian Networks

Directed and Undirected Graphical Models

Inference in Singly-Connected Graphs

Hidden Markov Models

Junction Tree Algorithm

Decision Making under uncertainty

Markov Decision Processes

Learning with Missing Data

Approximate Inference using Sampling

If time permits we will also cover some deterministic approximate inference.

Method of Instruction

Lectures

Assessment

The course has the following assessment components:

  • Written Examination (2.5 hours, 50%)
  • Coursework (50%)

To pass this course, students must:

  • Obtain an overall pass mark of 50% for all sections combined
  • Obtain a minimum mark of 50% in each component.

Resources

Reading list available via the UCL Library catalogue.