Computational Intelligence

Credits: 6

Semester: 3

Course: Elective

Lecturer

Michael Lees

Objectives

Students will learn: the basic tasks of machine learning, the basic concepts and basic mathematical apparatus of fuzzy logic, the basic concepts and principles of the operation of artificial neural networks, the main varieties of evolutionary algorithms, examples of the application of computational intelligence methods for solving problems; to set a task and develop an algorithm for solving it using the methods of computational intelligence, analyze scientific literature sources, analyze the solutions obtained, analyze the problem for choosing the best method of computational intelligence or a hybrid method suitable for a specific task, analyze the work of methods of computational intelligence with the identification of their strengths and weaknesses, to carry out analysis of the tuning parameters of neural networks, evolutionary algorithms and fuzzy method in.
Students will study the technology of applying the methods of computational intelligence to solve practical problems, methods of intellectual data analysis, the skills of creating and testing artificial neural networks, evolutionary algorithms and fuzzy methods in one of the high-level programming languages, to analyze the work of methods of computational intelligence with the identification of their strong and weak parties, to carry out analysis of the adjustment of parameters of neural networks, evolutionary algorithms and fuzzy methods, methods of obtaining data from different cing available sources.

Contents

Main topics of the discipline:

  • Biological and formal neuron. Architecture of neural networks. Classification of neural networks. Basic principles of training neural networks. Perceptron and multilayer neural network. Retraining of the network. Methods for calculating the output of an INS.
  • Neural networks with feedbacks. Hopfield Networks. Maps of Kohonen. Radial-basic neural networks. General information on the application of neural networks to solve problems of classification, approximation, modeling and control. Features of practical application of neural networks.
  • Evolutionary algorithm. Types of evolutionary algorithms. Function fitness. Integer and real coding of information. Basic operators of evolutionary search and their varieties. Application of evolutionary algorithms for solving optimization problems.
  • Parameters and adaptation of parameters. The theorem on the lack of free meals. Evolutionary strategies. Algorithms for estimating distributions. Systems of classifiers. Genetic programming. Algorithm of differential evolution. Features of the practical application of evolutionary computations.
  • Linguistic variable. Fuzzy sets. The membership function. Basic operations and relations of fuzzy logic. Algorithms of fuzzy inference Mamdani and Sugeno. A model of the singleton type. Fuzzy databases. Calculations with words. Comparison of fuzzy and probabilistic systems. Features of practical application of systems with fuzzy logic.
  • Neuroevolutionary algorithms. Neural networks. Evolutionary fuzzy systems. Sharing the methods of computational intelligence and machine learning
  • Solution of problems of classification, approximation, clustering, control. Open libraries and programs for methods of computing intelligence.

Format

Lectures and laboratory works.

Assessment

Examination.