1. SylabUZ
2. Wydział Nauk Inżynieryjno-Technicznych
3. semestr zimowy 2024/2025
4. Informatyka - Program Erasmus pierwszego stopnia
5. Elements of artificial intelligence

Wygeneruj PDF dla tej strony

Elements of artificial intelligence - opis przedmiotu

Cel przedmiotu

• Introducing students to basic architectures of artificial neural networks and their learning algorithms.
• Introducing students to the theory of fuzzy sets and fuzzy reasoning.
• Introducing students to the fundamentals of expert systems operation.
• Introducing students to evolutionary algorithms.
• Introducing students to heuristic strategies for searching graphs and trees.
• Developing skills in the implementation and practical use of the artificial intelligence methods learned.

Wymagania wstępne

Zapisz zmiany

Principles of programming

Zakres tematyczny

Lectures

1. Introductory lecture. Discussion of the content presented in the course. Provision and discussion of literature that students should familiarize themselves with. Discussion of the course completion requirements. Presentation and discussion of learning outcomes.
2.  Definition of basic concepts in the field of artificial intelligence. Types of artificial intelligence (weak and strong artificial intelligence). History of artificial intelligence development. Examples of artificial intelligence applications.
3. Introduction to neural networks. Elements of neurobiology. Structure of a biological neuron. Mathematical model of a neuron. Simple perceptron and its limitations. Adaline model. Linear neuron and its limitations. Decision boundary of the perceptron. Training and testing data. General idea of single neuron learning. Types and properties of neuron activation functions. Perceptron rule and Widrow-Hoff rule.
4. Multilayer networks. Architecture of feedforward networks. Decision boundary of a two-layer perceptron. Typical tasks performed by neural networks. Approximation capabilities of a multilayer perceptron. Presentation of the neural network learning problem as an optimization task. Definition of objective function. Typical objective functions used for training feedforward networks.
5. Methods of training feedforward multilayer networks. Introduction to optimization using gradient methods. Presentation of the idea of using gradient methods for neural network training. Simple gradient method. Steepest descent gradient method. Conjugate gradients method. Adaptive learning step and momentum. Variable metric methods.
6. Backpropagation algorithm. Example of the backpropagation algorithm for a two-layer network and generalization of the algorithm for networks with any number of layers. Weight update using online and offline (batch) strategies.
7. Recurrent networks, Hopfield model, associative memory, network equilibrium points, energy function, Hebb's rule, BAM network (Bidirectional Associative Memory).
8. Self-organizing networks, Kohonen network, Kohonen network topology, neighborhood of neurons, learning of self-organizing networks.
9. Fuzzy systems, fuzzy sets and fuzzy logic, examples of applications of fuzzy systems, operations on fuzzy sets, extension principle, fuzzy rules, general idea of fuzzy reasoning, fuzzy reasoning for the Mamdani model, fuzzy reasoning for the Takagi-Sugeno model, neuro-fuzzy structures and their learning algorithms.
10. Expert systems, structure of expert systems, knowledge base, rule-based systems, reasoning methods.
11. Introduction to evolutionary algorithms, general scheme of an evolutionary algorithm, simple genetic algorithm.
12. Heuristic search methods, problem solving by searching, cost-based search method, uniform cost strategy, heuristic function, greedy search strategy, A* algorithm.

Laboratory

1. Discussion of the laboratory operation rules and applicable occupational safety and health regulations. Presentation of the laboratory exercises program. Discussion of the course completion requirements. Introduction to the software available in the laboratory. Presentation of literature.
2. Familiarization with the environment for building and simulating artificial intelligence models.
3. Implementation of a simple perceptron model, visualization of its decision boundary, and its learning method.
4. Implementation of a two-layer neural network model and its training method using the backpropagation algorithm.
5. Classification of images using a multilayer neural network.
6. Implementation of associative memory using a Hopfield network.
7. Clustering of data using a self-organizing Kohonen feature map.
8. Application of fuzzy logic in control.
9. Implementation of a sample expert system.
10. Application of a simple genetic algorithm to find the global optimum of a given objective function.
11. Use of the A* algorithm to find the shortest path.
12. Summary and completion of the laboratory course.

Metody kształcenia

Lecture, teaching laboratory classes.

Efekty uczenia się i metody weryfikacji osiągania efektów uczenia się

Opis efektu	Symbole efektów	Metody weryfikacji	Forma zajęć

Warunki zaliczenia

Lecture - the passing criterion is a sufficient mark from the final test.
Laboratory - the passing criterion are positive marks for laboratory exercises and tests.
Final mark components = lecture: 50% + teaching laboratory: 50%

Literatura podstawowa

1. Russell S., Norvig P.: Artificial Intelligence: A Modern Approach, 4th edition, Prentice Hall, 2019.
2. Bishop C.M., Hinton G. : Neural Networks for Pattern Recognition, Clarendon Press, Oxford, 1995.
3. Edelkamp S., Schroedl S.: Heuristic Search: Theory and Applications, Morgan Kaufmann, 2012. 
4. Zimmermann H-J.: Fuzzy Set Theory and Its Applications, Springer, 2006.
5. Eiben A.E, Smith J.E.: Introduction to Evolutionary Computing, Springer, 2015.
6. Goodfellow I, Bengio Y., Courville A.: Deep Learning, MIT Press, 2016.
7. Chollet F.: Deep Learning with Python, Second Edition, Manning Publications, 2021.
8. Geron A.: Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow, 3rd Edition, O'Reilly, 2022. 

 

 

Literatura uzupełniająca

1. Bishop C.: Pattern Recognition and Machine Learning, Springer Verlag, 2006.
2. Ross. T.: Fuzzy Logic with Engineering Applications, Wiley, 2004.

Uwagi

Zmodyfikowane przez dr hab. inż. Marek Kowal, prof. UZ (ostatnia modyfikacja: 19-04-2024 09:31)