13S053NM - Neural Networks

Course specification
Course title		Neural Networks
Acronym		13S053NM
Study programme		Software Engineering
Module
Type of study		bachelor academic studies
Lecturer (for classes)		professor PhD Goran Kvaščev professor PhD Aleksandar Rakić
Lecturer/Associate (for practice)		asst. MA Marija Novičić, electrical and computer engineer asst. professor PhD Sanja Vujnović asst. Aleksa Stojić, electrical and computer engineer
Lecturer/Associate (for OTC)		asst. MA Marija Novičić, electrical and computer engineer
ESPB		6.0	Status	elective
Condition		none
The goal		Concepts of neural networks technology, architectures, learning ability and distributed information processing capability. Introduction to convolutional neural networks, autoencoder networks, deep learning. Design neural network systems for typical engineering applications, pattern recognition, algorithms for signal processing, classification....
The outcome		Students will be able to independently analyze and synthesize different types of neural networks that are applied in many areas of engineering and learn to apply various algorithms for learning and training of neural networks and their implementation using MATLAB/Python
Contents
Contents of lectures		Overview of classic structures of neural networks, training, generalization and initialization of neural networks. Classification and clustering with neural networks. Convolutional neural networks. Autoencoder neural networks. Concepts of regularization, data augmentation, hyperparameters, activation functions, dropout, crossentropy... Architectures LeNet, AlexNet, VGG, Resnet...
Contents of exercises		Computer exercises with demonstrations and training algorithms for the design of neural networks. Solve practical problems in various fields of engineering with the help of neural networks using MATLAB/Python
Literature
Goodfellow, Ian. "Deep Learning-Ian Goodfellow, Yoshua Bengio, Aaron Courville- Google Books." (2016). (Original title) Michael A. Nielsen, "Neural Networks and Deep Learning", Determination Press, 2015 (Original title) Patterson, Josh, and Adam Gibson. Deep learning: A practitioner's approach. " O'Reilly Media, Inc.", 2017. (Original title) Neural Fuzzy Systems: A Neuro-Fuzzy Synergism to Intelligent Systems, Chin-Teng Lin, C. S. George Lee, Prentice Hall, 1996 Neural Networks for Pattern Recognition, Christopher Bishop, Oxford University Press, 2000
Number of hours per week during the semester/trimester/year
Lectures	Exercises	OTC	Study and Research	Other classes
2	2	1
Methods of teaching		Lectures, exercises on computers
Knowledge score (maximum points 100)
Pre obligations		Points	Final exam	Points
Activites during lectures		0	Test paper	30
Practical lessons		12	Oral examination	0
Projects		18
Colloquia		40
Seminars		0