The course covers algorithms which gets its computational capabilities by training from examples. There is thus no need to explicitly provide rules and instead training using measured data is performed. Learning can be done either by providing the correct answer, or be totally autonomous.
The courser also covers principles of representation of data in neural networks. The course also includes principles of hardware architectures (euro chips and neuro computers) and shows how ANN can be used in robotics. We also show applications of learning systems in areas like pattern recognition, combinatorial optimization, and diagnosis.
After the course the student should be able to
- explain the function of artificial neural networks of the Back-prop, Hopfield, RBF and SOM type
- explain the difference between supervised and unsupervised learning
- describe the assumptions behind, and the derivations of the ANN algorithms dealt with in the course
- give example of design and implementation for small problems
- implement ANN algorithms to achieve signal processing, optimization, classification and process modeling
so that the student
- achieves an understanding of the technical potential and the advantages and limitations of the learning and self organizing systems of today
- can apply the methods and produce applications in their working life
