Machine Learning: 2009-2010
Information
|
Lecturer |
|
|
Degrees |
|
|
Term |
Hilary Term 2010 (16 lectures) |
Overview
The development of intelligent systems is an area that is becoming increasingly important in applied Computer Science. The most popular view of such systems postulates that truly intelligent machines must include the ability to learn from experience and observation.This is an advanced course that will focus on multiple machine learning approaches including neural networks, fuzzy systems, evolutionary algorithms. The course will also discuss general issues and challenging problems in machine learning, as model selection, feature selection, computational complexity of learning. The practicals will be concerned with the application of machine learning techniques to a range of real-world problems, with particular preference for problems from the Bioinformatics area.
The course is addressed to MSc students in Computer Science, and it assumes some familiarity with basic concepts of probability theory as well as some basic programming skills. Students attending Intelligent Systems I or Intelligent Systems II would have a particular advantage, but the course is self-contained and does not require any previous knowledge on Artificial Intelligence.
Learning outcomes
On completion of the course students will be expected to:
- Have a good understanding of general issues and challenges in developing machine learning applications: model selection, feature selection, model complexity, etc.
- Have an understanding of the strengths and weaknesses of many of the machine learning approaches.
- Have achieved a good level of knowledge and expertise in the main areas of computational intelligence: neural networks, fuzzy systems, evolutionary approaches.
- Be able to design and implement various machine learning algorithms in a wide range of real-world applications.
Synopsis
- Introduction to Machine Learning (2 lectures): Overview of Machine Learning. General issues: Concept learning, Classification and Classifiers, Supervised and unsupervised learning. Instance-based learning (K-Nearest Neighbour). Clustering. Learning using decision trees.
- Neural networks (4 lectures): Perceptron networks, Multi-Layer Perceptron networks, Self-Organizing Maps, Hopfield networks, Radial Basis Function networks, Support Vector Machines.
- Fuzzy Sets and Systems (3 lectures): Fuzzy sets. Learning through fuzzy logic, Fuzzy inference, Fuzzy modelling and optimization. Learning based on rough sets.
- Evolutionary algorithms (3 lectures): Genetic algorithms, Evolutionary strategies, Genetic programming, Ant colony optimization algorithms, Particle swarm optimization.
- Hybrid intelligent methods (1 lecture): Neuro-fuzzy systems, Neuro-symbolic systems, Genetic algorithms-based hybrid systems, Other hybrid learning approaches.
- Inductive Logic Programming (1 lecture): Inductive learning. Applications of ILP to real-world problems.
- Ensemble methods (1 lecture): Bagging, Boosting, methods of combining classifiers, classifier diversity, topologies of multi-classifier systems.
- Evaluating models and algorithms. Computational Learning Theory (1 lecture): Model selection, Feature selection, ROC analysis. PAC Learnability, The Vapnik-Chervonenkis dimension.
Syllabus
General Machine Learning concepts, Neural networks, Fuzzy sets and systems, Evolutionary algorithms, Hybrid intelligent methods, Inductive Logic Programming, Ensemble methods, Evaluating models and algorithms, Computational learning theory."Reading list
Primary Text
- T.M. Mitchell, Machine Learning, McGraw-Hill, 1997.
- M. Negnevitsky, Artificial Intelligence - A Guide to Intelligent Systems, Addison Wesley, London, 2001 (or later edition).
Reading List
- I.H. Witten and E. Frank, Data Mining: Practical Machine Learning Tools and Techniques with Java Implementations, Morgan Kaufmann, 2000.
- S. Russell and P. Norvig, Artificial Intelligence: A Modern Approach, Prentice Hall, 1995.
- B.D. Ripley, Pattern Recognition and Neural Networks, Cambridge University Press, 1996.
- C.M. Bishop, Neural Networks for Pattern Recognition, Oxford University Press, 1995.
- T. Hastie, R. Tibshirani and J.H. Friedman, The Elements of Statistical Learning, Springer 2001.