SubjectsSubjects(version: 963)
Course, academic year 2020/2021
  
Artificial Neural Networks - N445024
Title: Neuronové sítě
Guaranteed by: Department of Computing and Control Engineering (445)
Faculty: Faculty of Chemical Engineering
Actual: from 2019 to 2020
Semester: summer
Points: summer s.:5
E-Credits: summer s.:5
Examination process: summer s.:
Hours per week, examination: summer s.:2/2, C+Ex [HT]
Capacity: unknown / unknown (unknown)
Min. number of students: unlimited
State of the course: taught
Language: Czech
Teaching methods: full-time
Teaching methods: full-time
Level:  
Is provided by: M445004
Guarantor: Procházka Aleš prof. Ing. CSc.
Is interchangeable with: M445004
Examination dates   Schedule   
Annotation -
The subhect is devoted to construction, optimization and application of methematical models of neural networks. Application area covers the use of artificial neural networks for (i) rejection of noise signal components, (ii) time series prediction, and (iii) classification of signal and image components using matrix of pattern vectors. Real applications are restricted to biomedical multi-dimensional data and environmantal signals. All algorithms are studied in the MATLAB computational environment.
Last update: Procházka Aleš (25.07.2013)
Aim of the course -

Students will know how

(i) to model multilayer and reccurent neural networks,

(ii) to create pattern matrices forming inputs for neural networks,

(iii) to optimize neural networks for data classification,

(iv) to use neural networks for noise cancelling and signal prediction

Last update: Procházka Aleš (25.07.2013)
Literature -

R: S. Haykin: Neural Networks, Prentice Hall, 1999, ISBN 0132733501

Last update: TAJ445 (30.09.2013)
Syllabus -

1. Fundamentals of MATLAB environment, basic operations, data files processing

2. Visualization tools in MATLAB, symbolic mathematics, principles of SIMULINK

3. Basic mathematical models of neurons, transfer functions, threshold, error surface evaluation

4. Single-layer networks, perceptron learning rule, application to classification problems

5. Adaptive linear networks, Widrow-Hoff learning rule, training and coefficients optimization

6. Neural networks in adaptive noise cancellation

7. Multi-layer networks, error function, optimization of parameters, gradient descent method

8. Basic optimization methods (backpropagation, Levenberg-Marquardt algorithm)

9. Neural networks in signal prediction and system modeling, learning and generalization

10. Radial basis networks, transfer functions definition, network structure optimization

11. Associative learning rules, self-organizing networks and maps, Hebian learning, Kohonen rule

12. Neural networks in segmentation, feature extraction and classification, competitive learning

13. Simulation of artificial neural networks in the SIMULINK environment

14. Neural networks in system identification and control, supervised and unsupervised learning

Last update: Procházka Aleš (08.12.2005)
Learning resources -

http://uprt.vscht.cz/prochazka/pedag/lectures/SP0_MATLAB_2006EN.pdf

Last update: Procházka Aleš (25.07.2013)
Entry requirements -

Knowledge of MATLAB and digital signal and image processing methods

Last update: Procházka Aleš (25.07.2013)
Registration requirements -

Digital Signal and Image Processing

Last update: Procházka Aleš (25.07.2013)
Teaching methods
Activity Credits Hours
Konzultace s vyučujícími 0.5 14
Účast na přednáškách 1 28
Příprava na přednášky, semináře, laboratoře, exkurzi nebo praxi 0.5 14
Práce na individuálním projektu 1 28
Příprava na zkoušku a její absolvování 1 28
Účast na seminářích 1 28
5 / 5 140 / 140
Coursework assessment
Form Significance
Regular attendance 30
Report from individual projects 30
Oral examination 40

 
VŠCHT Praha