Neural Networks and Neurocomputers

Semestr: Winter

Range: 2+2s


Credits: 4

Programme type: Undefined

Study form:

Course language:


Introduction, basic terminology. Hopfield's net. Perceptron, multilevel perceptron, Adaline/Madaline. Back-propagation. Kohonen's net, LVQ1, LVQ2, LVQ3. ART net, Carpenter-Grossberg classifier. Boltzmann machine. Applications: prediction by means of neural nets, expert systems, image processing, data compression. Neural hardware, neurochips, neural accelerators. Software simulators, introduction to NeuralWorks Professional II/Plus.


Course syllabus:

1. Introduction into neural nets
2. Hopfield´s net
3. Back-propagation neural net
4. Kohonen´s net
5. Simulation, SW products
6. Paradigm classification
7. Time series prediction
8. Neural nets and image processing
9. Neural regulators
10. Selforganizing neural nets
11. Data compression
12. HW accelerators
13. Boltzman machine
14. Reserve

Seminar syllabus:

1. - 13. Will be organized according the number of participants (Individual neural net project implemented in NEURALWORKS PROFESSIONAL II/PLUS)
14. Credit


[1] Haykin, S.: Neural Networks. IEEE Computer Society Press 1994