About

The RecSys Challenge 2018 will be organized by Spotify, The University of Massachusetts, Amherst, and Johannes Kepler University, Linz. Spotify is an online music streaming service with over 140 million active users and over 30 million tracks. One of its popular features is the ability to create playlists, and the service currently hosts over 2 billion playlists.

This year’s challenge focuses on music recommendation, specifically the challenge of automatic playlist continuation. By suggesting appropriate songs to add to a playlist, a Recommender System can increase user engagement by making playlist creation easier, as well as extending listening beyond the end of existing playlists.

As part of this challenge, Spotify will be releasing a public dataset of playlists, consisting of a large number of playlist titles and associated track listings. The evaluation set will contain a set of playlists from which a number of tracks have been withheld. The task will be to predict the missing tracks in those playlists.

A detailed description of the challenge can be found on https://recsys-challenge.spotify.com. Accepted contributions will be presented during the RecSys Challenge Workshop in 2018.

Publications ^top

"Current Challenges and Visions in Music Recommender Systems Research" M. Schedl, H. Zamani, C.-W. Chen, Y. Deldjoo, M. Elahi.

"Recsys Challenge 2018: Automatic Music Playlist Continuation" C.-W. Chen, P. Lamere, M. Schedl, H. Zamani.

"Proceedings of the ACM Recommender Systems Challenge 2018"

"An Analysis of Approaches Taken in the ACM RecSys Challenge 2018 for Automatic Music Playlist Continuation" H. Zamani, M. Schedl, P. Lamere, C.-W. Chen.

Participation and Data

Update for 2020:In Sept 2020, Spotify released an identical challenge and dataset named The Million Playlist Dataset Challenge. The dataset is available for non-commercial research use, and submissions to the challenge will be graded on an ongoing basis. Please visit the Million Playlist Dataset Challenge website for the terms and conditions.

The data for this year's challenge is provided by Spotify and can be downloaded from http://recsys-challenge.spotify.com. You can access the data once you create an account and agree with the terms and conditions. The information on the challenge participation can also be found on this website.

Timeline ^top

Note: the timeline is subject to slight modifications.

When?	What?
January 2018	Release of the Million Playlist Dataset The data can be downloaded from http://recsys-challenge.spotify.com.
March 2018	RecSys Challenge starts. Submission system will be available for offline evaluation. Each team can submit at most 1 run per day. A daily leaderboard will be available once the challenge starts. The submissions are evaluated based on 50% of a private test set.
June 30, 2018	Final submission due (11:59 PM in the UTC-12:00 time zone).
Early July 2018	Announcement of the final leaderboard. The daily leaderboard is recalculated using the private test set. The test set will be made publicly available.
July 15, 2018	Paper submission due Note: paper submission is necessary for the winners. Page limit: 4-6 pages (ACM SIG Format) Submission website: EasyChair
August 13, 2018	Paper Acceptance Notifications The submitted papers will be evaluated based on the novelty, the clarity, and the empirical results. Each paper will be reviewed by at least three PC members. Accepted papers should be presented in the RecSys Challenge Workshop.
August 27, 2018	Camera-ready due for the accepted papers.
October 7, 2018	Workshop will take place as part of the ACM RecSys conference in Vancouver, Canada.

Workshop Program and Accepted Papers

Time	Session
09:00 - 09:10	Opening: Welcome and Challenge statistics, Workshop organizers.
09:10 - 10:40	Session 1: Neighborhood-based approaches: Efficient K-NN for Playlist Continuation, Domokos M. Kelen, Dániel Berecz, Ferenc Béres, András A. Benczúr. [Code] Effective Nearest-Neighbor Music Recommendations, Malte Ludewig, Iman Kamehkhosh, Nick Landia, Dietmar Jannach. [Code] Automatic Music Playlist Continuation via Neighbor-based Collaborative Filtering and Discriminative Reweighting/Reranking, Lin Zhu, Bowen He, Mengxin Ji, Cheng Ju, Yihong Chen. [Code] Efficient similarity based methods for the playlist continuation task, Guglielmo Faggioli, Mirko Polato, Fabio Aiolli. [Code]
10:40 - 11:00	Coffee Break
11:00 - 12:30	Session 2: Different approaches: Automatic Playlist Continuation using Subprofile-Aware Diversification, Mesut Kaya, Derek Bridge. [Code] Random Walk with Restart for Automatic Playlist Continuation and Query-Specific Adaptations, Timo van Niedek, Arjen de Vries. [Code][Slides] Automatic playlist continuation using a hybrid recommender system combining features from text and audio, Andres Ferraro, Dmitry Bogdanov, Jisang Yoon, Kwangseob Kim, Xavier Serra. [Code][Slides] A Line in the Sand: Recommendation or Ad-hoc Retrieval?, Surya Kallumadi, Bhaskar Mitra, Tereza Iofciu. [Code]
12:30 - 14:00	Lunch Break
14:00 - 15:30	Neural network approaches: TrailMix: An Ensemble Recommender System for Playlist Curation and Continuation, Xing Zhao, Qingquan Song, James Caverlee, Xia Hu. [Code][Slides] An Ensemble Approach of Recurrent Neural Networks using Pre-Trained Embeddings for Playlist Completion, Diego Monti, Enrico Palumbo, Giuseppe Rizzo, Pasquale Lisena, Raphaël Troncy, Michael Fell, Elena Cabrio, Maurizio Morisio. [Code][Slides] Towards Seed-Free Music Playlist Generation: Enhancing Collaborative Filtering with Playlist Title Information, Jaehun Kim, Minz Won, Cynthia C.S. Liem, Alan Hanjalic. [Code][Slides] Using Adversarial Autoencoders for Automatic Playlist Continuation, Iacopo Vagliano, Lukas Galke, Florian Mai, Ansgar Scherp. [Code][Slides]
15:30 - 16:00	Coffee Break
16:00 - 17:30	Top-performing approaches: A hybrid two-stage recommender system for automatic playlist continuation, Vasiliy Rubtsov, Mikhail Kamenshikov, Ilya Valyaev, Vasiliy Leksin, Dmitry I. Ignatov. [Code] Artist-driven layering and user’s behaviour impact on recommendations in a playlist continuation scenario, Sebastiano Antenucci, Simone Boglio, Emanuele Chioso, Ervin Dervishaj, Shuwen Kang, Tommaso Scarlatti, Maurizio Ferrari Dacrema. [Code][Slides] MMCF: Multimodal Collaborative Filtering for Automatic Playlist Continuation, Hojin Yang, Yoonki Jeong, Minjin Choi, Jongwuk Lee. [Code][Slides] Two-stage Model for Automatic Playlist Continuation at Scale, Maksims Volkovs, Himanshu Rai, Zhaoyue Cheng, Ga Wu, Yichao Lu, Scott Sanner. [Code]
17:30 - 17:40	Closing

Program Committee ^top

Himan Abdollahpouri, DePaul University, USA
Qingyao Ai, University of Massachusetts Amherst, USA
Dmitry Bogdanov, Universitat Pompeu Fabra, Spain
Matthias Braunhofer, Free University of Bozen-Bolzano, Italy
Ching-Wei Chen, Spotify, USA
Marco De Gemmis, University of Bari, Italy
Yashar Deldjoo, Polytechnic University of Milan, Italy
Mehdi Elahi, Free University of Bozen-Bolzano, Italy
Bruce Ferwerda, Jönköping University, Sweden
Emilia Gomez, Universitat Pompeu Fabra, Spain
Balázs Hidasi, Gravity R&D, Hungary
Dietmar Jannach, AAU Klagenfurt, Austria
Peter Knees, TU Wien, Austria
David Massimo, Free University of Bozen-Bolzano, Italy
Vijai Mohan, Amazon A9, USA
Javier Parapar, University of A Coruña, Spain
Neil Rubens, University of Electro-Communications, Japan
Alan Said, University of Skövde, Sweden
Marko Tkalcic, Free University of Bozen-Bolzano, Italy
Yong Zheng, Illinois Institute of Technology, USA

Challenge Organizers ^top

Ching-Wei Chen, Spotify, New York, USA
Markus Schedl, Johannes Kepler University, Linz, Austria
Hamed Zamani, University of Massachusetts Amherst, MA, USA
Paul Lamere, Spotify, USA

Advisors ^top

Spotify Organizers ^top

Benjamin Carterette
Christophe Charbuillet
Cedric De Boom
Jean Garcia-Gathright
James Kirk
James McInerney
Vidhya Murali
Hugh Rawlinson
Sravana Reddy
Marc Romejin
Romain Yon
Yu Zhao

ACM RecSys Challenge 2018

About

Publications top

Participation and Data

Timeline top

Workshop Program and Accepted Papers

Program Committee top

Challenge Organizers top

Advisors top

Spotify Organizers top