1. SylabUZ
2. Wydział Informatyki, Elektrotechniki i Automatyki
3. semestr zimowy 2018/2019
4. WIEiA - oferta ERASMUS / Informatyka - Program Erasmus drugiego stopnia
5. Artificial intelligence techniques

Wygeneruj PDF dla tej strony

Artificial intelligence techniques - opis przedmiotu

Cel przedmiotu

• Familiarize students with the selected AI techniques, current trends and application areas of AI systems.
• Teach students how to select the appropriate method of artificial intelligence depending on the type of scientific or engineering problem.
• Teach students how to design, develop and train artificial intelligence models and how to interpret and assess their results.

Wymagania wstępne

Zapisz zmiany

Zakres tematyczny

Introduction to artificial intelligence. Social and biological inspirations. General assumptions. Learning and organization of data. Comparison of artificial intelligence techniques to analytical methods.

Feed forward neural networks. Structures and their properties. Training algorithms: backpropagation and quasi-Newton methods. Examples of neural network applications in modeling and image recognition.

Recurrent neural networks. Dynamic-feedback neural networks. Learning algorithms for feedback
neural networks.

Self-organizing neural networks. Kohonen self-organizing feature maps. Competitive learning.

Deep learning. Convolutional neural network. Restricted Boltzman Machine. Deep Belief Networks. Fast deep learning with GPU computations.

Fuzzy and neuro-fuzzy systems. Fuzzy sets and fuzzy logic. Operations on fuzzy sets. Fuzzy inference. Fuzzy rules. Neuro-fuzzy systems. Gradient descent based learning algorithm for neuro-fuzzy systems.  

Expert systems. Knowledge representation. Rule based systems. Inference methods. Forward reasoning. Backward reasoning. Mixed reasoning.

Evolutionary algorithms and swarm intelligence. Basic concepts. General schema of the evolutionary algorithm. Evolutionary strategies. Encoding methods. Evolutionary operators. Simple genetic algorithm.  Crossing operators. Holland's theorem on schemes. Preventing premature convergence to local optima. Genetic programming. Adaptation in non-stationary environment. Swarm intelligence algorithms. Particle swarm optimization. Ant algorithms.

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. Haykin S.: Neural Networks: A Comprehensive Foundation (2nd Edition), Prentice Hall, 1998..
2. Goodfellow I., Bengio Y., Courville A.: Deep Learning, MIT Press, 2016.
3. Russell S., Norvig P.: Artificial Intelligence: A Modern Approach, Prentice Hall, 2009.
4. Bishop C.M., Hinton G.: Neural Networks for Pattern Recognition, Clarendon Press, Oxford, 1995.
5. Zimmermann H-J.: Fuzzy Set Theory and Its Applications, Springer, 2006.
6. Rutkowska D.:Neuro-Fuzzy Architectures and Hybrid Learning, Springer, 2001.
7. Goldberg D.E.: Genetic Algorithms in Search, Optimization, and Machine Learning, Addison-Wesley, 1989.
8. Dasgupta D., Michalewicz Z.: Evolutionary Algorithms in Engineering Applications, Springer-Verlag, 2010
9. Eberhart R.C., Shi Y., Kennedy J.: Swarm Intelligence, Morgan Kaufmann, 2001.

Literatura uzupełniająca

 

1. Murphy K.P.: Machine Learning. A Probabilistic Perspective, MIT Press, 2013.
2. Theodoridis S.: Machine Learning. A Bayesian and Optimization Perspective. Academic Press, 2015. 
3. Hastie T., Tibshirani R., Friedman J: The Elements of Statistical Learning, Springer, 2001.

Uwagi

Zmodyfikowane przez dr hab. inż. Marek Kowal, prof. UZ (ostatnia modyfikacja: 27-03-2018 18:03)