Course detail

Soft Computing

FIT-SFCAcad. year: 2026/2027

This course covers methods of soft computing and computational intelligence, which enable the solution of complex problems through learning, approximation, optimization, and decision-making under conditions of uncertainty. Emphasis is placed on approaches that leverage tolerance for imprecision, incomplete information, and approximate solutions.

Students will explore the principles and applications of neural networks—including deep learning and language models—as well as fuzzy systems, reinforcement learning, evolutionary and swarm algorithms, and other computational intelligence methods. The course also covers probabilistic reasoning, rough sets, the fundamentals of chaos theory, and hybrid approaches that combine various methods.

Attention is directed toward the application of these techniques in areas such as classification, optimization, associative recall, decision-making, and control.

Language of instruction

Czech

Number of ECTS credits

Assignment to study programme types

Master's

Mode of study

Not applicable.

Guarantor

doc. Ing. Vladimír Janoušek, Ph.D.

Department

Department of Intelligent Systems (UITS)

Entry knowledge

Programming in C++ or Python languages.
Basic knowledge of differential calculus and probability theory.

Rules for evaluation and completion of the course

Mid-term written examination - 15 points.
Project - 30 points.
Final written examination - 55 points, miinimum 25.

Aims

The objective of this course is to provide students with knowledge of the fundamental principles of soft computing and computational intelligence, as well as the ability to navigate their potential applications for solving practical problems.

Students will gain an overview of the main categories of methods, understand their mathematical and algorithmic principles, and become familiar with their typical applications.

Upon completion of the course, students should be able to:
-Explain the fundamental principles of neural networks, fuzzy systems, reinforcement learning, evolutionary algorithms, and other methods covered.
- Distinguish between various approaches to representing and processing uncertainty, imprecision, and incomplete information.
- Characterize the capabilities, assumptions, and limitations of individual methods, and assess their suitability for specific types of tasks.
- Apply selected methods to solve practical problems, such as classification, optimization, associative recall, or control.
- Understand the possibilities for combining different methods within hybrid intelligent systems.
- Use the standard technical terminology—in both Czech and English—relevant to these fields.

Study aids

Informace pro studenty

Prerequisites and corequisites

Not applicable.

Basic literature

Kriesel, D.: A Brief Introduction to Neural Networks, 2005, Fundamentals of the New Artificial Intelligence, Springer-Verlag New York, Inc., 2008. ISBN 978-1-84628-838-8
Russel,S., Norvig,P.: Artificial Intelligence, Prentice-Hall, Inc., third edition 2010, ISBN 0-13-604259-7
Rutkowski, L.: Flexible Neuro-Fuzzy Systems, Kluwer Academic Publishers, 2004, ISBN 1-4020-8042-5

Type of course unit

Lecture

26 hours, optionally

Teacher / Lecturer

doc. Ing. Vladimír Janoušek, Ph.D.

Syllabus

Introduction. Computational intelligence.
Neural networks. Backpropagation.
Convolutional neural networks. Deep learning.
Recurrent neural networks for time series. Embedding.
Recurrent networks as associative memory, Hopfield networks.
Transformer and LLM.
Fuzzy sets, fuzzy k-means, fuzzy logic.
Fuzzy control. Adaptive neuro-fuzzy systems.
Markov decision process and reinforcement learning.
Genetic algorithms and genetic programming.
Swarm intelligence. Probabilistic inference.
Rough sets and their applications. Chaos theory.
Recurrent networks with continuous time. Hybrid approaches.

Project

26 hours, compulsory

Teacher / Lecturer

Ing. Petr Veigend, Ph.D.
doc. Ing. František Zbořil, CSc.
doc. Ing. Vladimír Janoušek, Ph.D.

Syllabus

An independent individual project focused on solving a specific problem using selected soft computing or computational intelligence methods. Typical application areas include classification, optimization, associative information processing, and control. The project entails implementing the chosen method and evaluating the results obtained.

E-learning texts

SFC 2020
Soft Computing 2020.pdf 7.1 MB

Elearning

eLearning: currently opened course