When Media Gets Wise: Collaborative Filtering with Mobile Media Agents

摘要

We present a model where media (e.g. music files) are autonomous entities that carry their own individual information. Our goal is to turn such files into autonomous, rule-following agents capable of building their own identities from interactions with other agents and users. We are exploring how collaborative filtering-like behaviour could emerge out of large ensembles of interacting agents, which are distributed over mobile devices in social networks. We have implemented a first version of the model in the form of a music player application for mobile devices, called Push!Music. This system takes advantage of active recommendations as well as implicit user activity to build a profile for each media file.