Bachelor Artificial Intelligence

Course Media Understanding

This is the information of 2017

The course was originally given as a combined course with Informatiekunde, called Knowledge-based media systems.

Description

The description is available in the course catalogue with code MEUN6Y. The course is a 3rd year obligatory course in the Bachelor Artificial Intelligence Curriculum. More resources are available on Blackboard.

Media understanding is the science of identifying semantic structures in digital media objects such as audio, biosignals, images, text and videos. A intelligent program should be able to do what our senses and cognition do: immediate understanding of events as diverse as watching a bird or listening to a bird song. The course introduce state-of-the-art methods for media summarization and event categorization.

Learning objectives

In this course students will gain executable knowledge (theory) on:

general concepts of media understanding, including common and deviating characteristics between media (e.g., text, video, and biosignals);
models and algorithms for feature extraction, information filtering, and categorization;
similarity and distance (measures);
advanced topics such as local and spectral features, perception and psychophysics, and dynamic models of categorization; and
semantic methods for media description.

The course aims to be complementary to other courses, in particular Machine Learning and Computer Vision.

Practical skills

During course, the student will gain the following practical skills:

Multimedia Understanding is an empirical branch of computer science. As such design, engineering and academic skills are essential
Training in conducting scientific research. Starting with the definition of a verifiable research question to presenting research (ideas) and reporting the research's results.

As such, the course also prepares you for your BSc graduation project.

Literature

Horst Eidenberger, 'Fundamental Media Understanding', 2nd edition, Books on Demand GmbH, Norderstedt, Germany, 2011, ISBN: 978-3-842-37917-6.

Schedule

The schedule highlights the different parts of the book:

Week 1: Part 1: Context and ambience (Theory: Knowledge & Context)
Week 2: Part 2: Text (Theory & Application)
Week 3: Part 2: Image (Theory & Application)
Week 4: Part 3: Video (Theory & Application)
Week 5: Part 4: Audio (Theory & Application)
Week 6: Part 5: Biometry (Theory & Application)
Week 7: Conclusion: Latest developments

Note that these are (still) presentations of previous year. Ths presentations will be updated with the slides of experts on each of the subjects .

Exam

Practical Assignment

Here are some final reports on the practical assignment: building an interactive system which combines multiple media and brings the sensor input to a semantic level:

Sara Aerssens, Caitlin Lagrand, Putri van der Linden, Lina Murady, Tirza Soute and Leila Talha, "Personal Teaching Aid", project report, Universiteit van Amsterdam, April 2, 2017.
Houda Albert, Urja Khurana, Melissa Tjhia and Richard Olij, "Cookin': Interactive Cooking Assistant", project report, Universiteit van Amsterdam, April 2, 2017.
Jonathan Gerbscheid, Thomas Groot, Joram Wessels, Rijnder Wever and Wijnand Van Woerkom,"Personalized news conversations with the Softbank Pepper", project report, Universiteit van Amsterdam, March 30, 2017.
Suzanne Bardelmeijer, Karen Beckers, Sanne Eggengoor, Job van Gerwen and Roos Slingerland",Don't over think, I know your drink", project report, Universiteit van Amsterdam, April 2, 2017.

Evaluation

The course is this year evaluated by the participants with a 4.4:
.

Last updated May 10, 2017

Course Media Understanding