| Study program / study programs: Advanced data analytics |
| Course: Neural networks and deep learning |
| Teacher(s): Zoran V. Ševarac, Dragan O. Đurić |
| Course status: Elective |
| ECTS points: 10 |
| Prerequisites: none |
| Course objective: To learn basic concepts and algorithms in the field of neural networks and deep learning, and methods for application of these technologies in various domains. |
| Learning outcomes: Students will learn basic neural network concepts, types and application procedures, and develop skills required for their practical application. |
| Course structure and content: Basic concepts: Artificial neurons, activation functions, types and architectures of neural networks, learning algorithms. Mathematical and theoretical models and analogies with biological systems. Error functions and optimization methods. Neural network architectures: Multilayer perceptrons, algorithms for learning multilayer perceptrons and their application. Convolutional neural networks and deep learning. Problem solving problem using neural networks and deep learning, problems in the practical applications. Methodology for different types of problems that are resolved using neural networks and deep learning: classification, clustering, prediction. Areas of application of neural networks and deep learning: medicine, finance, production, defense, software development. The course assumes undergraduate-level proficiency in linear algebra, probability, statistics and programming (Python basics). |
| Literature/Readings: Neural Networks for Pattern Recognition, Christopher Bishop, Oxford University Press, 1996. ISBN-13:978-0198538646, http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.679.1104&rep=rep1&type=pdf Neural Networks – A Systematic Introduction, Raul Rojas, Springer-Verlag, 1996. ONLINE: http://page.mi.fu-berlin.de/rojas/neural/index.html.html Fundamentals of Neural Networks: Architectures, Algorithms and Applications, Laurene Fausett, Pearson Education, 2006. Deep Learning, Ian Goodfellow, Yoshua Bengio, Aaron Courville, MIT Press, 2016 Documentation and examples from Neuroph project http://neuroph.sourceforge.net/ |
| The number of class hours per week: Lectures: 4 Labs: 0 Workshops: 0 Research study: 3 Other classes: 0 |
| Teaching methods: Individual and group work; lectures and labs |
| Evaluation/Grading (maximum 100 points): Pre-exam requirements (Practical project assignment): 30 Final exam (Full implementation of practical project assignment): 70 |
