Artificial Intelligence and Neural Networks in Applications
- Garant předmětu:
- Department of Instrumentation and Control Engineering
Theory of Problem Solving. Logic of First Order Language, Teorem proving, Resolution principle. Formal grammars, Abstract automata as syntactic analysers. Fuzzy Sets, Fuzzy Relational Calculus, Fuzzy Logic. Rule based systems.
Fuzzy controllers, Design of fuzzy controllers, fuzzy controllers of Mamdani and Takagi-Sugeno type. Implementation of fuzzy controllers in the MATLAB/Simulink environment and Fuzzy Toolbox for MATLAB. Examples and applications. Expert systems and their applications for engineering problems. Neural networks. Classification of neural networks. MLP (Multi-Layer Perceptron) and RBF (Radial Basis Function) neural networks. Examples of neural network with neural units of higher order (HONNU). Implementation of neural networks in MATLAB/Simulink and the Neural Network Toolbox for MATLAB. Examples and applications. Data mining of knowledge from databases.
Artificial Intelligence and Neural Networks 2016/2017
(Problem fields for exam.)
(prof. Bíla, dr. Oswald)
1.Theory for problem solving, types of problems, Generalised State Space (GSS) and Generalised Problem Solver (GPS), formal model of the synthesis of the solution.
2.Formal logic. The language and the calculus of first order predicates. The synthesis of a solution.
3.Automatic proving of theorems - resolution method. Description of the method and examples.
4.Formal languages and grammars. Essential types of grammars.
5.Abstract automata as syntactic analysers. Types of automata. State machines (finite automaton and pushdown automaton).
6.Fuzzy controllers. Mamdani and Takagi-Sugeno controller.
7.Fuzzy toolbox for MatLab/Simulink.
8.Synthesis of fuzzy controller in Fuzzy toolbox.
9. The example of the control of nonlinear dynamic system by fuzzy controller.
10. Genetic algorithms. Description of the algorithm. Ending condition.
11. Neural network - life cycle and the fields of the deployment of neural networks.
12. Types of neural networks (MLP, RBF , HONU).
13. Training and testing of neural networks.
14. Back propagation method for training of MLP networks.
15. Tuning of RBF networks.
16. Convergence of training process.
17. Identification of dynamic systems by neural networks.
- Syllabus of lectures:
ARTIFICIAL INTELLIGENCE from 2017/18
(prof. Bíla, dr. Oswald)
1. Introduction to Artificial Intelligence and Theory of problem solving.
2. Formal logic. The language and the calculus of predicates of the first order (FOL). Automatic theorem proving.
3. Formal grammar and formal languages. Abstract automata.
4. Abstract automata as syntactic analysers of formal languages.
5. Fuzzy sets.
6. Fuzzy logic and rule-based systems.
7. Problem solving by expert systems.
8. Fuzzy controllers, Fuzzy toolbox for MatLab/Simulink.
9. Synthesis of a fuzzy controller in Fuzzy toolbox for MatLab/Simulink.
10. Genetic algorithms.
11. Neural networks, theory and types (MLP, RBF, HONNU).
12. Neural networks - Deep learning.
12. Neural Networks - application in identification and signal processsing.
13. Data mining of knowledge from databases.
14. Final lecture and Assessments.
- Syllabus of tutorials:
The topics of the seminaries follow the topics of lectures.
Topics of the semester projects will be gradually assigned since 9.lecture.
Conditions for the assessment:
-50 % participation in seminaries.
-Accepted semester project.
- Study Objective:
- Study materials:
1.P.H. Winston: Artificial Intelligence. Addison-Wesley Publishing Company, Amsterdam, 1977.
2.R.B. Banerji: Artificial Intelligence.(Theoretical Approach.) North Holland, N.Y., 1986.
3. Nils J.Nilsson: Artidficial Inteligence: A New Synthesis. Morgan Kaufmann Publisher,Inc. , San Francisco, Cal., 2005.
- Further information:
- No time-table has been prepared for this course
- The course is a part of the following study plans: