Solutions for Personalized T-learning Alberto Gil Solla Department of Telematic Engineering University of Vigo (Spain) EuroITV 2005: the 3rd European Conference on Interactive Television Aalborg, Denmark April 1, 2005
Introduction Migration from analogue to digital TV Interactive multimedia applications mixed with audiovisual contents Standards to normalize receivers (MHP) T-learning is gaining popularity: Life-long education is essential in the current global economy Engaging applications for digital TV are needed
T-learning vs e-learning T-learning is different from e-learning E-learning involves active users TV traditional passive attitude demands an edutainment approach A likely overwhelming increase in T-learning contents will disorient users Tools will be needed to assist them to find interesting personalized educational material
AVATAR AdVAnced Telematic search of Audiovisual contents by semantic Reasoning Framework to test recommendation strategies: Profiles matching (collaborative filtering) Semantic reasoning about the user preferences and TV programs (enhanced content-based techniques) The knowledge base is an OWL ontology about the TV domain, describing hierarchies of classes and properties. Specific instances are extracted from TV-Anytime program descriptions Extended to applying the same techniques to recommend personalized T-learning contents
User profile Watching habits Learning history Courses Contents Descriptive metadata Recommender agent TV-Anytime LOM LIP
IEEE LOM: Learning Object Metadata IEEE standard to describe educational material: contents, purposes, formats, level of difficulty, languages, authors, intended audience, dependencies... Enables search and discovery of contents Enhancements for t-learning needed
TV-Anytime To describe: TV programs Content segmentation Users’ personal profiles Users’ viewing habits
LIP: Learner Information Package IMS standard to describe any relevant information about the user Combined with LOM, it permits making access to a course dependent on having proved some knowledge
Conclusions Applying semantic web techniques can improve course targeting, so optimising advertisement investments Standards are essential