Course detail

Machine Learning Fundamentals

FEKT-MPA-MLFAcad. year: 2022/2023

The course deals with classical machine learning methods, such as support vector machines or principal component analysis, as well as with the approaches based on artificial neural networks, including convolutional or recursive networks. In addition to lectures, an important part of the course are exercises focused both on understanding the basic principles and on the use of machine learning in the field of radio communications, ranging from the classification of radio transmitters to a complete transmission system based on machine learning.

Language of instruction

English

Number of ECTS credits

Mode of study

Not applicable.

Guarantor

prof. Ing. Roman Maršálek, Ph.D.

Department

Department of Radio Electronics (UREL)

Offered to foreign students

Of all faculties

Learning outcomes of the course unit

The graduate of the course will be able to (a) use basic machine learning methods for classification (b) use methods based on artificial neural networks (c) correctly choose a suitable machine learning method for the given task (d) discuss the use of machine learning methods in radio communications

Prerequisites

A student who enrolls in a course should:
- understand basic mathematical methods at the bachelor's degree level
- be able to write a simple program in the Matlab environment and one of the higher programming languages

Co-requisites

Not applicable.

Planned learning activities and teaching methods

Not applicable.

Assesment methods and criteria linked to learning outcomes

PC labs: 16 points

Project: 50 points

Exam (written test): 24 points

Exam (oral interview, optional): 10 points

Course curriculum

1 – Course organization, introduction to machine learning, motivation
2 – Basics of linear algebra needed for machine learning
3 - Support vector Machines
4 - Principal component analysis
5 – Introduction to neural networks, representation, classification
6 – Training of neural networks (linear regression, gradient methods, polynomial regression …)
7 - Convolutional neural networks
8 - Recursive neural networks
9 – Tuning of hyperparameters, batch normalization, frameworks
10 – Unsupervised learning
11 – Generative networks, autoencoders, GAN
12 – Machine learning in large scale
13 – Use of machine learning in radio communications

Work placements

Not applicable.

Aims

The aim of the course is to make students familiar with the basic concepts of machine learning, the most important individual methods and tasks of machine learning, and to show the use of machine learning in the field of radio communications. The aim of the computer exercises is to gain practical experience in the implementation of machine learning methods on real data sets.

Specification of controlled education, way of implementation and compensation for absences

Not applicable.

Recommended optional programme components

Not applicable.

Prerequisites and corequisites

Not applicable.

Basic literature

Goodfellow, I., Bengio, Y., Courville, A., Deep Learning (Adaptive Computation and Machine Learning series), MIT Press, 2016, ISBN ‎ 0262035618 (EN)
Smola, A., Vishwanathan, S.V.N., Introduction to Machine Learning, Cambridge University press, available at https://alex.smola.org/drafts/thebook.pdf (EN)
Mueller, J.P., Massaron, L. Machine Learning‎ For Dummies; 1st edition, 2016, ISBN : ‎ 1119245516 (EN)