Semestr: Winter
Range: 2+2c
Completion:
Credits: 4
Programme type: Undefined
Study form: Fulltime
Course language:
Mainly the basic neural paradigms are studied (perceptron and perceptron-like artificial neural nets, Hopfield net, Kohonen and ART selforganizing nets, Neocognitron, GMDH, etc.). Their applications in different tasks are pointed out. Fundamental ideas of HW accelerator design are mentioned. Applications like data prediction, image and sound neural processing, data compression, principal and independent component analysis are described.
Artificial neural nets, Perceptron, BP, Hopfield, Kohonen, ART, GMDH, Neocognitron, RBF, applications.
1. Artificial neural nets introduction
2. Hopfield net
3. Perceptron-like neural nets
4. Back-propagation nets, the principle of error back propagation
5. Self-organizing nets - Kohonen
6. Self-organizing nets - ART
7. GMDH nets
8. Neocognitron
9. Image processing by artificial neural nets
10. Data prediction
11. Data mining
12. Neural net accelerators
13. Boltzman machine
14. Fuzzy neural nets
1. Examples of artificial neural net application
2. Simulation tools
3. 1st laboratory task analysis
4. Consultations
5.
6. Result presentation
7. 2nd laboratory task analysis
8. Consultations
9.
10.
11.
12.
13. Result presentation
14. Evaluation
1. Haykin, S.: Neural Networks. IEEE Computer Society Press 1994