SGN-90206 International Doctoral Seminar in Signal Processing, 1 cr
Lessons
| Study type | P1 | P2 | P3 | P4 | Summer | Lecture times and places |
|
|
|
|
|
|
|
|
Description:
Lectures on: Learning dictionaries for sparse representations.This is a course given by Professor Bogdan Dumitrescu in Erasmus Program.
Person responsible:
Ioan Tabus
Target groups:
Jatkotutkinto-opiskelijat
Signaalinkäsittely ja tietoliikennetekniikka
Assessment scale:
Numerical evaluation scale (1-5) will be used on the course
Additional information of course implementation:
Activities: There will be 4 lectures, on 8, 10,15,17 March Requirements: For obtaining the 1 credit point, the students must attend the lectures and write a 2-3 pages free description of the problems that they consider more interesting, including potential applications. Connections with their own research interests will be highly appreciated. Performing an extensive computer project work within the topics of the lectures with a 10-15 page report will get additional 2 credit points. Contents: The lectures cover the topics: - Sparse representations and the dictionary learning (DL) problem - Standard DL algorithms - Online DL algorithms - Optimizing the dictionary size - Classification using dictionaries - Structured dictionaries - Kernel dictionary learning