Jun Ma (College of Computer Science & Technology of Shandong University) will present recent research by himself and his team on recommendation systems for smart TV. Wednesday October 5, 16.00hrs, Room G5.29.
A smart TV is a platform which enables users to watch the programs of CATV, retrieve CATV programs broadcasted in recent days, watch the videos and movies from web sites under agreement with the TV maker, watch TV shows, education programs and pay-channels or movies in the cloud storages of TV makers. In the near future, smart TV will be an important platform for e-business as well. Our research is being carried out with Hisense, the biggest TV maker in China. The function of a recommendation system (RS) for a smart TV is to let the general platform become a personalized one. The main difference or challenges between smart TV recommendation systems and recommender systems for web video sites are: 1) No rating information is available, only the click sequence on TV program selection and the time periods of the TV programs that user watched; 2) One TV ID usually corresponds to a family, the family maybe of three generations. Their common interests usually may be empty. 3) The TV watching time distributions of different TV users are very different, which is the basic information in designing temporal context RS. In this talk, I will introduce our current research progress in designing and implementing recommender systems on Hisense smart TVs.