The "Mobile Music Genius" (MMG) is an intelligent mobile music player
(currently only for Android devices) that aims at learning in which
situation and mood the listener want to listen to which kind of music.
Assuming that soneone who is relaxing at the beach would like to listen
to different music than someone (or even the same person) doing
exhausting sports, a machine learning component makes use of a wide
range of user context attributes gathered on the device, for instance,
acceleration, weather conditions, or current mood, to learn such music
style-user properties relations.
The
player then makes use of these learned relations to offer an "adaptive
playlist" mode in which, based on the user context, songs that are
likely to fit the current entertainment need of the user are
added to the music playlist. In addition, automated playlist generation
based on a social music similarity measure (tag weights from
Last.fm)
is provided. In its current version, the MMG player only supports local
music collections. We plan, however, to add support for streaming
services, such as
Spotify.
If you are a student and interested in this or similar work, please
contact
Markus
Schedl.
The player was implemented by Georg Breitschopf as part of his Master's
thesis.