Francois Maillet - "Creating transparent, steerable recommendations "

Music recommendation systems are increasingly important in the ever growing world of digital music.  However, most commercial music recommenders rely on collaborative filtering techniques to generate music recommendations. These type of recommendations lack two aspects that are important for recommendation.  First, they lack transparency - they cannot explain why an item was recommended beyond the trivial "Other people who listed to XX also listened to YY". Second, they lack steerability - there is no way for a user to interact with the recommender to steer it to more relevant content. In this talk, the Sun Labs Music Explaura is presented - a web-based recommender that provides transparent and steerable recommendations. The Music Explaura can offer a detailed explanation about why a particular item was recommended and will allow a user to steer the recommendations based upon attributes of the music.