RECSYS CHALLENGE 2018: AUTOMATIC PLAYLIST CONTINUATION

Published in International Society for Music Information Retrieval, 2017

Recommended citation: Markus Schedl, Hamed Zamani, Ching-Wei Chen, Yashar Deldjoo, Mehdi Elahi International Society for Music Information Retrieval, 2017 (IMSIR LBD 2017).

[Proceedings] [PDF] [bibtex]

Abstract

In recent years, considerable attention has been given to studies on the role of playlists in music consumption. A study carried out in 2016, by the Music Business Association, showed that playlists accounted for 31% of music listening time among listeners in the USA. Another study, conducted by MIDiA, revealed that as many as 55% of streaming music service subscribers create playlists. Accordingly, music streaming services such as Spotify currently host over 2 billion playlists. This evidence may indicate the growing importance of playlists as a mode for music consumption, and indeed the crucial necessity of developing algorithms for automatic playlist continuation, which is the focus of the ACM Recommender Systems Challenge 2018. In this paper, we — the organization team of this challenge — briefly discuss the particular task we defined for the participating teams. We also provide some information on the overall Challenge process.