2025.09.30

Research on Privacy Verification in Federated Recommender Systems Accepted for Presentation at RecSys 2025, the Leading International Conference on Recommender Systems

Toyota Motor Corporation
InfoTech
Kenji Shinoda, Takeyuki Sasai, Shintaro Fukushima

Toyota Motor Corporation has identified a new vulnerability in federated recommender systems. The research paper “Popularity-Bias Vulnerability: Semi-Supervised Label Inference Attack on Federated Recommender Systems” has been accepted to the Nineteenth ACM Conference on Recommender Systems (RecSys 2025), a leading international conference in the field. The paper was presented at the main conference in September 2025.

For further details, please download and review the materials below.