Renewal: an online competition platform for news recommender systems

Julien Hay (1) , Erik M. Bray , Bich-Liên Doan (2, 3) , Fabrice Popineau (2) , Anne-Catherine Letournel (2) , Ouassim Ait Elhara

1 LaHDAK - LRI - Données et Connaissances Massives et Hétérogènes (LRI)
2 LISN - Laboratoire Interdisciplinaire des Sciences du Numérique
3 LaHDAK - Données et Connaissances Massives et Hétérogènes - LISN

This paper presents the Renewal platform and the new features it will provide for the online evaluation of news recommender systems. The platform is currently under development. In future organized competitions, it will allow the collection of real-time feedback from users reading news stories through a dedicated mobile application in order to evaluate competing recommender systems in online settings. We discuss the benefits over offline settings and the differences with existing online platforms dedicated to this task. We also present the results of a simulation allowing us to determine under which conditions, notably in terms of number of users and number of evaluation days, it is possible to organize a Renewal competition.

Mots clés: News Recommendation; Online Evaluation; Recommender System

Domaines: Recherche d'information [cs.IR]; Intelligence artificielle [cs.AI]

Julien Hay, Erik M. Bray, Bich-Liên Doan, Fabrice Popineau, Anne-Catherine Letournel, et al.. Renewal: an online competition platform for news recommender systems. Challenges in Machine Learning at NeurIPS 2020, Chklovski Tara; Banifatemi Amir; Stolovitzky Gustavo; Mendrik Adriënne, Dec 2020, Online, United States. ⟨hal-04421282⟩