WeST at European Conference on Artificial Intelligence
Multi-modal understanding of Social Media
Organisatorisches
Beschreibung
The key objective of this thesis will be development and evaluation of novel multi-modal probabilistic models for describing Social Content in Web 2.0 folksonomies (i.e. image features, source locations, timestamps, annotations, social relationships between users). The project will include analysis of Social Data (based on large-scale crawls of folksonomies like Flickr, Delicious, and Bibsonomy), definition of contend description models, as well as systematic large-scale evaluation with real data from Web 2.0 portals.
Existing close cooperations with Social Media systems (Bibsonomy@Kassel, Yahoo! Research in Barcelona) open challenging opportunities to discuss ideas with leading experts in Social Web mining and to experiment with "real" Social Web at large (e.g. in form of short-term research visits). This work will be done in the context and with support of the special interest group on Web Science (4cats) at ISWeb.
Kontakt