Improving Recommender Systems Beyond the Algorithm.

Recommender systems rely heavily on the predictive accuracy of the learningalgorithm. Most work on improving accuracy has focused on the learningalgorithm itself. We argue that this algorithmic focus is myopic. Inparticular, since learning algorithms generally improve with more and betterdata, we propose shaping the feedback generation process as an alternate andcomplementary route to improving accuracy. To this effect, we explore howchanges to the user interface can impact the quality and quantity of feedbackdata -- and therefore the learning accuracy. Motivated by information foragingtheory, we study how feedback quality and quantity are influenced by interfacedesign choices along two axes: information scent and information access cost.We present a user study of these interface factors for the common task ofpicking a movie to watch, showing that these factors can effectively shape andimprove the implicit feedback data that is generated while maintaining the userexperience.