"I would like to watch something like 'The Terminator'…" Cooperative Query Personalization Based on Perceptual Similarity

摘要: In this paper, we showcase a privacy-preserving query personalization system for experience items like movies, music, games, or books. Personalizing queries such is notoriously difficult as meaningful attributes are either missing in the database would require extensive domain knowledge not available to most users. For reason, state-of-the-art content provision platforms e.g., Netflix Amazon usually rely on recommender systems support their users, and often working parallel with traditional SQL-style queries. Unfortunately, have several shortcomings example high barriers new users joining system, which first setup preference profile lengthy process, inability pose beyond recommendations matching personal profile, severe privacy concerns due storing rating data all long-term. order provide an alternative, present demonstration paper powerful intuitive query-by-example (QBE) interaction system. Bayesian Navigation used personalize user’s fly. The central challenge when using QBE selection of features represent database. Here, high-dimensional feature space was mined from large number allowing us measure perceived similarity between steer process. This also addresses many issues our capabilities can be by any user anonymously drive-by fashion. proposed demo, try never before presented hands-on, use it discover interesting movies tailored preferences pleasantly simple enjoyable experience.