Artificial Neural Networks

(Artificiella neurala nät)

Course homepage, Spring 2005


Original road imageBrain and neural networkRoad image segmented by a neural network

This is the homepage of the third year Computer Science course in Artificial Neural Networks at the Department of Technology, Örebro University.


2005-06-06 The exam from Wednesday with answers can be found here.


Contents

Introduction
Course Contents
Teachers
Literature
Web Links
Labs
Examinations
Past Exam Papers
Matlab
Previous course (2004)


Introduction

Biological neuron

Artificial neural networks are parallel computing devices consisting of many interconnected simple processors. They share many characteristics of real biological neural networks such as the human brain. Knowledge is acquired by the network from its environment through a learning process, and this knowledge is stored in the connections strengths (weights) between processing units (neurons). In recent years, neural computing has emerged as a practical technology with applications in many fields. The majority of these applications are concerned with problems in pattern recognition, for example, in automatic quality control, optimization and feedback control. The course deals with classical pattern recognition, supervised and unsupervised learning using artificial neural networks, genetic algorithms, and applications of neural computing in artificial intelligence and robotics. The theoretical parts of the course will be tested by a number of computer based laboratory sessions ("laborations") using MATLAB.


Course Contents

The course is sub-divided into two parts (you can collect the points for each part separately):

Part I: Artificial Neural Networks in Theory, 3p

This part of the course is assessed by the written examination (60 multiple choice questions).

Part II: Laboratory Work in Artificial Neural Networks, 2p

This part of the course consists of 6 laboratory sessions:


Teachers

Lectures:

Tom Duckett
E-mail: tom.duckett@tech.oru.se

Labs:

Malin Lindquist
E-mail: malin.lindquist@tech.oru.se
Henrik Andreasson
E-mail: henrik.andreasson@tech.oru.se
The lectures will be given in English, but you can ask questions in Swedish.

Course Literature

Official Course Literature

  1. Robert Callan, "The Essence of Neural Networks", 1999, Prentice-Hall.
  2. A handout on "Introduction to pattern recogition" (this will be a translation of the corresponding Swedish handout).
  3. Other material to be given out during the course.

Literature for those who want to know more

Dan W. Patterson, "Artificial Neural Networks: Theory and Applications", 1996, Prentice-Hall.
R. Beale and T. Jackson, "Neural Computing: An Introduction", 1990, Institute of Physics Publishing.
Christopher M. Bishop, "Neural Networks for Pattern Recognition", 1995, Oxford University Press.
Simon Haykin, "Neural Networks: A Comprehensive Foundation", 1998, Prentice-Hall.
M. Arbib, "The Handbook of Brain Theory and Neural Networks", 1995, MIT Press.
F. F. Soulié and P. Gallinari, "Industrial Applications of Neural Networks", 1998, World Scientific.
David E. Goldberg, "Genetic Algorithms in Search, Optimization & Machine Learning", 1989, Addisson-Wesley.


Links

  • Artificiella neurala nätverk - en kort introduktion
  • Neuronnät och lärande system (kurs in Linköping)
  • Genetiska algoritmer
  • Vektorer, matriser, nätverk - några elementa
  • Swedish-English dictionary
  • Mean Vector and Covariance Matrix
  • Introduction to Neural Networks (1)
  • Introduction to Neural Networks (2) (course webpage, covers backpropagation very well)
  • Neural Networks FAQ (comp.ai.neural-nets)
  • WEBSOM - Self-Organizing Maps for Internet Exploration
  • Introduction to Evolutionary Biology (biological background)
  • Sumeet's Neural Networks Page (contains many links)
  • (Please mail me if you find any more useful links. :-)


    Labs

    The are six obligatory labs which all students must pass by presenting their work to the teachers. Each lab must be finished by the start of the next lab session. The penalty for missing more than one lab deadline will to complete one extra lab (at the discretion of the course leader).

    Lab

    Name

    Instructions

    Data/M-files

    1

    Introduction to pattern recognition

    ann_lab1.pdf

    lab1.zip

    2

    Minimum distance classifier and Bayes optimal classifier

    ann_lab2.pdf

    lab2.zip

    3

    Simple neuron models and training algorithms

    ann_lab3.pdf

    lab3.zip

    4

    Multi-layer feedforward networks and backpropagation

    ann_lab4.pdf

    lab4.zip

    5

    Self-organising feature maps and Hopfield networks

    ann_lab5.pdf

    lab5.zip somtoolbox2.zip

    6

    Sensory data analysis with an electronic nose

    ann_lab6.pdf

    data files

    All labs are carried out in Matlab. If you want to know more about Matlab there are some links below.

    The labs should be done in teams of two people. There is one 4 hour lab session per week where the teachers will be present.


    Examinations

    For the exact date, time and location see http://stormsvala.oru.se.

    Week 12: ordinary examination.

    May-June: first re-sit examination.

    August: second re-sit examination.


    Past exam papers

    Past exam papers with answers in pdf format.

    Jun 2005

    Mars 2005

    Aug 2004

    Jun 2004

    March 2004

    Jun/Aug 2003

    March 2003

    March 2002

    March 2001

    March 2000


    Matlab

    Matlab should be installed on the department computers; to start it click on NTLinks - Matematik - Matlab 6.5.

    There is a lot of on-line documentation in the form of pdf-files, an introduction to the Matlab environment, and all the toolboxes. Start a web browser, and then select Help - Help Desk (HTML) in Matlab.

    Here's some links to Matlab related stuff:


    Biological neuron


    Last updated 2005-1-17 by Tom Duckett.