V ponedeljek, 11. maja 2020, bo ob 16.00 uri prek spletnih orodij na daljavo izvedeno predavanje v okviru PONEDELJKOVEGA SEMINARJA RAČUNALNIŠTVA IN INFORMATIKE Oddelkov za Informacijske znanosti in tehnologije UP FAMNIT in UP IAM.

ČAS/PROSTOR: 11. maj 2020 ob 16.00 na daljavo 

-------------------------------------------------
PREDAVATELJICA: Elham MOTAMEDI
-------------------------------------------------

Elham Motamedi is a PhD student at UP FAMNIT. She completed her Master's degree of networked systems at the University of Tehran, Iran and her Bachelor's degree of information technology at the Isfahan University of Technology, Iran. Her research interests include multimedia and recommender systems.

--------------------------------------------------------------------------------
NASLOV: User-centric multimedia characteristics: A Survey
--------------------------------------------------------------------------------

The properties of multimedia content are usually represented as content-centered metadata, such as type (e.g. video, audio, text, etc.), duration, resolution, etc. However, in user-centric applications, such as personalized recommender systems, it makes more sense to describe multimedia items with user-centric meta-data, such as genre, emotions, etc. For example, a recommender system would make a better job of explaining recommendations in the form of "You have been recommended this film because you liked comedies" than "You have been recommended this film because you liked 1:30 long films". The problem of such a characterization is that it usually requires humans to annotate with user-centric metadata, which is expensive. Hence,  research efforts have been made to develop automatic methods for mapping multimedia content to user-centric item characteristics.
In this presentation, I present some of the state-of-the-art studies in studying automatic approaches for predicting user-centric characteristics of multimedia contents and some raw ideas of future work.

---------------------------------------------------------------------------------------------------

Predavanje bo potekalo v angleškem jeziku prek spletnega orodja Zoom.
Do predavanja dostopate tako, da se povežete prek sledeče povezave:

https://us02web.zoom.us/j/297328207

Vabljeni!