A team of students from Politecnico di Milano, coordinated by Maurizio Ferrari Dacrema, researcher at the Department of Electronics, Information and Bioengineering, and Andrea Pisani, student in the Information Technology Ph.D. programme, won the RecSys Challenge 2024 as the best academic team. The team is composed by Andrea Alari, Lorenzo Campana, Federico Giuseppe Ciliberto, Saverio Maggese, Carlo Sgaravatti, Francesco Zanella.
The RecSys Challenge 2024 lasted for three months, with its final phase in Bari from October 14 to 18, 2024 as part of the 18th ACM Conference on Recommender Systems, and was organized by the Danish newspaper Ekstra Bladet in collaboration with the Indian Institute of Management Visakhapatnam, the Technical University of Denmark, the University of Bari Aldo Moro and the Politecnico di Bari.
This year's challenge focused on online news recommendation, addressing both the technical and normative challenges related to designing effective and responsible recommender systems for the publishing industry. Tackling this problem requires to develop a complex personalization process, which opens several challenges related to modeling of the user behavior, managing the rapid obsolescence of news articles, utilizing large-scale machine learning as well as safeguarding user privacy.
Participants had access to a large real-world dataset provided by Ekstra Bladet, which included logs from over 2.7 million active users who interacted with the online news platform over a six-week period. The dataset contains more than 600 million article impressions and over 120,000 news articles, enriched with textual features such as titles, abstracts, and body content.
Approximately 150 teams participated in the competition, including companies such as Huawei's Noah’s Ark Lab, the Meituan Search and Content Intelligence Center (China), as well as leading universities like the University of Tokyo, Zurich, and Amsterdam. The team from Politecnico di Milano presented their solution in the paper “Exploiting Contextual Normalizations and Article Endorsement for News Recommendation”
