Code

COMPM012 (Also taught as: COMPGI07)

Year

4

Prerequisites

Part A: Proficiency in at least one high-level programming language such as C, C++, Pascal or Java etc as the course does not teach programming skills from scratch. Prior programming expertise is needed so that the new languages can be introduced quickly. Part B: although this course is not mathematically rigorous students should be at home in multivariate calculus and linear algebra.

Term

1

Taught By

Mark Herbster (Part A) (50%)
Massi Pontil (Part B) (50%)

Aims

The aims of Part A are: to provide a practical knowledge and understanding of the programming environments of Matlab and Mathematica. Knowledge of and proficiency in these languages/packages form an essential component of courses M057 (Evolutionary & Natural Computation) and M055 (Supervised Learning). The aim of Part B is to introduce the students to those elements of linear algebra which play an important role in machine learning and data analysis.

Learning Outcomes

Part A - Programming Issues: To be able to program in Matlab and Mathematica and to recognize their strengths and weaknesses versus conventional programming languages. Part B - To become confident with vector/matrix computations and basic concepts of linear algebra which are important for the development of machine learning techniques.

Part A - Programming Issues

Matlab

Matrix Operations
Control flow
I/O
Script files
2d and 3d graphics

Mathematica

Symbolic mathematics
List processing
Control flow
Functions
Graphics

Part B - Elements of Linear Algebra and its Applications

1. Vector and matrix operations, Orthogonality, Norms, Singular Value Decomposition (SVD).

2. Applications of SVD to Machine Learning and data analysis.

Principal component analysis, Least squares and regularization, Kernel Methods.

3. Elements of Numerical Optimization.

Part A: Lecture presentations with intensive programming coursework which is oriented towards machine learning. Part B: Lecture presentations with associated class problems. Part A and Part B are equally weighted. Full details of coursework to be set will be provided by the lecturers for this course.

The course has the following assessment components:

• Coursework Section (1 piece, 100 %)

To pass this course, students must:

• Pass the coursework section
• Pass the Coursework (50% or more)

Mastering MATLAB 6: A Comprehensive Tutorial and Reference by Duane Hanselman and Bruce R. Littlefield, Prentice Hall,

The Mathematica Book, Stephen Wolfram, Cambridge University Press, ISBN 0-521-64314-7.

The Definitive Guide to Project Management, Sebastian Nokes et al, Financial Times Prentice Hall, 2003

ISBN 0 273 66397 6

The Mythical Man-Month, Fredrick P Brooks, Addison-Wesley, 1995 Anniversary edition, Addison -Wesley

ISBN 0 201 83595 9

Leading Change, John P. Kotter, Harvard business School Press, 1996

ISBN 0 87584 747 1

Agile Software Development Ecosystems, Jim Highsmith, Addison-Wesley, 2002

ISBN 0 201 76043 6

In addition during the course you will be provided with both printed copies and references for research articles.

Lecture notes

General Information on Matlab and Mathematica

MSc Machine Learning Homepage