School of Electronics and Computer Science:
COMP3008 Machine Learning

Basic Information

School	Dept- Electronics & Computer Science
Known as	COMP3008.
Status	This syllabus is still provisional.
Session and Semester	Semester Two, 2011 - 2012
Credit	10 Credit Points
Unit Leader	Dr Adam Prugel-Bennett
Moderators	Prof Mahesan Niranjan
Study	100 nominal hours
Assessment	Examination 70% Coursework 30%
Coursework	1 assignment
Teaching	24 Lectures
Prerequisites and Exclusions	Prerequisites: (COMP3032 - Intelligent Algorithms or MATH2022 - Mathematics For Electronic & Electrical Engineering or MATH2025 - Mathematics for Computer Engineering).
Referral	On referral, this unit will be assessed 100% by examination.
Syllabus Approved

Description

Aims

This unit aims to impart an understanding of the role of neural computing in computer science and artificial intelligence.

Learning Outcomes

Knowledge and Understanding

Having successfully completed the module, you will be able to demonstrate knowledge and understanding of:

1. classical neural network architectures such as the perceptron,MLP, RBF and SVM
2. handling data
3. theoretical concepts including issues of optimisation and generalisation including regularisation
4. the relationship of neural computing to neurophysiology
5. many of the common architectures and learning algorithms

Intellectual Skills

Having successfully completed the module, you will be able to:

1. how to analyse the performance of a neural network using bias-variance, structural risk minimisation or probability methods
2. derive formulas for training neural networks by minimising a cost function

General Transferable (key) Skills

Having successfully completed the module, you will be able to:

• Collect and pre-process data for learning
• Match a particular problem to a particular learning method
• Have an historical perspective on the subject

Topics Covered

• Historical Perspective
  • Biological motivations: the McCulloch and Pitts neuron, Hebbian learning.
  • The power of machine learning methods: what is a learning algorithm? what can they do?
  • Generalisation: What is learning?
• Classification Techniques (Supervised)
  • Perceptron Learning
  • Support Vector Machines
  • The multi-layer perceptron (MLP)
• Classification Techniques (Unsupervised)
  • K-Means Clustering
• Regression Techniques
  • Polynomial Fitting
  • Kernel Based Networks
• Data Handling
  • Types of Data
  • Data Partitioning
• Other Techniques
  • Radial Basis Function Networks, Bayesian Neural Networks, Support Vector Machines.
• Case Studies
  • Example applications from: optimisation, pattern recognition, and data modelling.

Resources

Background Resources

• Witten, IH and Frank, E, Data Mining: Practical Machine Learning Tools and Techniques with Java Implementations Second Edition, Morgan Kaufmann Publishers, 2005. [Library] [Shops]
• Aleksander I and Morton H, An Introduction to Neural Computing, Chapman and Hall, 1990
• Beale R and Jackson T, Neural Computing: An Introduction, Adam Hilger, 1990
• Bishop CM, Neural Networks for Pattern Recognition, Clarendon Press, 1995 [Library] [Shops]
• Haykin S, Neural Networks: A Comprehensive Foundation Second Edition., Prentice-Hall, 1999 [Library] [Shops]
• Hassoun, MH, Fundamentals of Artificial Neural Networks, MIT Press, 1995.
• Anderson JA, An Introduction to Neural Networks, MIT Press, 1995
• Cherkassky,V and Mulier, F. Learning from Data, John Wiley and Sons, 1998 [Library] [Shops]

Taught to

COMP3008

Non-existing cohort: "ceMEng3" (Optional)
Pt III BSc Computer Science (Optional)
Non-existing cohort: "csBScAi3" (Optional)
Non-existing cohort: "csBScDs3" (Optional)
Non-existing cohort: "csBScIm3" (Optional)
Computer Science Integrated PhD (Optional)
Pt III MEng Computer Science with Artificial Intelligence (Optional)
Pt III MEng Computer Science (Optional)
Pt III MEng Computer Science with Distributed Systems & Networks (Optional)
Pt III MEng Computer Science with Image and Multimedia Systems (Optional)
Pt III MEng Computer Science with Mobile and Secure Systems (Optional)
MSc in Artificial Intelligence (Optional)
Non-existing cohort: "csMScCo" (Optional)
Pt III BEng Electronic Engineering (Optional)
Pt III MEng Electronic Engineering (Optional)
Pt III MEng Electronic Engineering with Artificial Intelligence (Compulsory)
Pt III MEng Electronic Engineering with Computer Systems (Optional)
Pt III MEng Electronic Engineering with Nanotechnology (Optional)
Pt III MEng Electronic Engineering with Optical Communications (Optional)
Pt III MEng Electronic Engineering with Power Systems (Optional)
Pt III MEng Electronic Engineering with Mobile and Secure Systems (Optional)
Pt III MEng Electronic Engineering with Wireless Communications (Optional)
MSc in Systems and Signal Processing (Optional)
ECS Socrates Students (Optional)
Pt III Units offered to other Faculties (Optional)
Pt III BEng Software Engineering (Optional)
Pt III MEng Software Engineering (Optional)

Students who are not registered on an ECS approved programme may take this module subject to meeting its pre-requisites and the availability of resources. To confirm this, please can you contact the module leader (as listed above) in the first instance. They will then refer you on to the appropriate director of studies for formal approval of your selection.

Change Log

2012-01-17 10:41:39.647 - apb
2011-04-04 18:59:39.830 - Roll script