Dec 10, 2019   0:56 a.m. Radúz
Academic information system

Course syllabus I-SUNS - Machine learning and Neural networks (FEEIT - WS 2019/2020)


     Information sheet          ECTS          Syllabus          


     Slovak          English          


University: Slovak University of Technology in Bratislava
Faculty: Faculty of Electrical Engineering and Information Technology
Course unit code: I-SUNS
Course unit title: Machine learning and Neural networks
Mode of delivery, planned learning activities and teaching methods:
lecture2 hours weekly (on-site method)
seminar2 hours weekly (on-site method)

 
Credits allocated: 6
 
Recommended semester/trimester: Applied Informatics - master (compulsory), 1. semester
Level of study: 2.
Prerequisites for registration: none
 
Assesment methods:
Active participation in the exercises and successful fulfillment of tasks. Successful completion of the exam.
Total mark in the examination will be derived from a point expressed as follows:
-exercises max. 45 points
-written exam max. 55 points
Exam is successful provided the following criteria are met:
-20 points for exercise
-20 points for the written exam
Evaluation of the subject is according to the grading scale according to STU Study Regulations (Article 16, paragraph 3):
Success criteria (percentage of results in the evaluation of the subject) for classification grades are as follows:
a) A – 92 to 100%
b) B - 83 to 91%
c) C - 74 to 82%
d) D – 65 to 73%
e) E – 56 to 64%
f) FX – 0 to 55%
 
Learning outcomes of the course unit:
The subject Machine Learning and Neural Networks aims to apprise the students of the theory and applications of machine learning. The subject offers systematic approach to the best known methods of machine learning with the emphasis on neural networks, boosting, kernel methods, support vector machines, clustering, deep learning. Following the study of theoretical foundations, a student is able to utilize these modern tools in various areas of information and communication technologies. The subject contains also machine learning applications (in the area of pattern recognition, biometrics, communication networks, signal processing etc.).
 
Course contents:
Concepts and principles (artificial intelligence, machine learning, computational intelligence, intelligent systems, knowledge discovery, neural network).
Supervised learning, unsupervised learning, learning theory, reinforcement learning.
Neurocomputing, differences to the classical approach, range of applications, analogies and differences in biological and artificial neural networks (ANN).
Models of a neuron, activation function, ANN architectures, learning process, ANN and mapping.
Multilayer perceptron, backpropagation algorithm, generalization, approximation of functions.
RBF (Radial-Basis Function) network, regularization theory.
Self-organizing systems based on competitive learning and the Hebbian learning.
Recurrent networks.
Ensembles - boosting, strong and weak learning, AdaBoost.
Margins, kernel methods, support vector machines.
Clustering.
Deep learning, deep neural networks.
Applications of machine learning in pattern recognition and biometrics, in communications networks, in signal processing.
 
Recommended or required reading:
Basic:
MARSLAND, S. Machine Learning: An Algorithmic Perspective. Boca Raton : CRC Press, 2009. 390 p. ISBN 978-1-4200-6718-7.
MITCHELL, T M. Machine learning. New York : McGraw-Hill, 1997. 572 p. ISBN 0-07-115467-1.
KECMAN, V. Learning and Soft Computing: Support Vector Machines, Neural Networks, and Fuzzy Logic Models. Massachusetts : MIT Press, 2001. 541 p. ISBN 0-262-11255-8.
HAYKIN, S. Neural Networks: A Comprehensive Foundation. New York : Macmillan College Publ.Co.Inc, 1994. 696 p. ISBN 0-02-352761-7.
BISHOP, C. Neural Networks for Pattern recognition. Oxford: Clarendon Press, 1996.
ORAVEC, M. Metódy strojového učenia na extrakciu príznakov a rozpoznávanie vzorov. 1. diel: Neurónové siete na extrakciu príznakov, kompresiu a rozpoznávanie obrazu. Bratislava : Nakladateľstvo STU, 2012. 150 p. ISBN 978-80-227-3691-6.
ORAVEC, M. -- PAVLOVIČOVÁ, J. Metódy strojového učenia na extrakciu príznakov a rozpoznávanie vzorov 2: Rozpoznávanie tvárí v biometrii. Bratislava: vydavateľstvo Felia, 2013. 179 p. ISBN 978-80-971512-0-1.
ORAVEC, M. -- POLEC, J. -- MARCHEVSKÝ, S. Neurónové siete pre číslicové spracovanie signálov. Bratislava: Faber, 1998. ISBN 80-967503-9-9.
ORAVEC, M. -- FÉDER, M. -- ZELINA, M. Strojové učenie a neurónové siete: učebné texty. Bratislava : RT Systems, 2013. ISBN 978-80-970519-5-2.
KVASNIČKA, V. Úvod do teórie neurónových sietí. Bratislava: IRIS, 1997. ISBN 80-88778-30-1.

Recommended:
Goodfellow,I., Bengio,Y., Courville,A.: Deep Learning, MIT Press book, 2016 online http://www.deeplearningbook.org/
Li Deng, Dong Yu: Deep Learning: Methods and Applications, Now Publishers, 2014
Oravec,M., Féder,M., Zelina,M.: Strojové učenie a neurónové siete– učebné texty, RT systems Bratislava, 2013, ISBN 978-80-970519-5-2, http://ibooks.sk/publ/suns/
UFLDL (Unsupervised Feature Learning and Deep Learning) Tutorial, http://deeplearning.stanford.edu/wiki/index.php/UFLDL_Tutorial

 
Language of instruction: slovak or english
 
Notes:
 
Courses evaluation:
Assessed students in total: 477

ABCDEFX
11,7 %19,7 %29,6 %23,3 %11,1 %4,6 %
Name of lecturer(s): Ing. Zuzana Bukovčiková (examiner, instructor, tutor) - slovak, english
Ing. Pavol Marák (examiner, instructor, tutor) - slovak, english
prof. Dr. Ing. Miloš Oravec (examiner, instructor, lecturer, person responsible for course, tutor) - slovak, english
Ing. Dominik Sopiak, PhD. (examiner, instructor, tutor) - slovak, english
 
Last modification: 9. 5. 2019
Supervisor: prof. Dr. Ing. Miloš Oravec and programme supervisor


Last modification made by RNDr. Marian Puškár on 05/09/2019.

Type of output: