CWI team wins recommendation challenge

Researchers Arjen de Vries, Alan Said and Alejandro Bellogin from Centrum Wiskunde & Informatica in Amsterdam, scored the highest ranking in the news recommendation challenge that is part of the ACM Recommender System Conference.

Publication date: 17-10-2013

Researchers Arjen de Vries, Alan Said and Alejandro Bellogin from Centrum Wiskunde & Informatica in Amsterdam, scored the highest ranking in the news recommendation challenge that is part of the ACM Recommender System Conference. The winning CWI team achieved the best performance out of twenty international research teams that participated in the competition. The conference was held in Hong Kong from 12 to 16 October 2013 where the results of teams were presented.

Recommendations were featured on the web pages of a wide range of German online newspapers and magazines such as Tagesspiegel, Sport 1 and Computerwoche. To tune their recommendation algorithms, the researchers were provided with user and item features at the start of the competition. In the second stage of the challenge the researchers had the chance to directly interact with the real-time recommender system allowing them to improve the performance continuously. The recommendation algorithm that was submitted by the winning team in Amsterdam achieved most user clicks.

The challenge ran from August to September 2013 in collaboration with Plista, a pioneering German enterprise that provides the recommendations to the German media. At the ACM RecSys conference, the best solution will be awarded. ACM RecSys is the premier international forum for the presentation of new research results, systems and techniques in the broad field of recommender systems.