65. THE NATURAL LANGUAGE OF PLAYLISTS

Department: Computer Science & Engineering
Research Institute Affiliation: California Institute for Telecommunications and Information Technology (Calit2)
Faculty Advisor(s): Gert Lanckriet

Primary Student
Name: Brian R McFee
Email: bmcfee@ucsd.edu
Phone: 000-000-0000
Grad Year: 2012

Abstract
We propose a simple, scalable, and objective evaluation procedure for music playlist generation algorithms. By drawing on standard techniques from the statistical natural language processing literature, we characterize playlist algorithms as generative models of strings of songs belonging to some unknown language. To demonstrate the evaluation procedure, we compare several playlist algorithms which are derived from audio content, semantic annotations, and meta-data. We then develop an efficient learning algorithm to construct an optimal ensemble model from a collection of simple playlist algorithms. Experiments on a large collection of naturally occurring playlists demonstrate the efficacy of the evaluation procedure and learning algorithm.

Related Links:

  1. http://cseweb.ucsd.edu/~bmcfee/playlists/

« Back to Posters or Search Results