Federated Multi-view Matrix Factorization for Personalized Recommendations
This work addresses privacy-preserving personalized recommendations for users in federated settings, but it appears incremental as it extends existing federated learning to multi-view matrix factorization.
The paper tackles the problem of personalized recommendations in federated learning by introducing a federated multi-view matrix factorization method that learns from multiple data sources without transferring personal user data to a central server, and it shows improved performance over simpler methods, particularly for cold-start scenarios.
We introduce the federated multi-view matrix factorization method that extends the federated learning framework to matrix factorization with multiple data sources. Our method is able to learn the multi-view model without transferring the user's personal data to a central server. As far as we are aware this is the first federated model to provide recommendations using multi-view matrix factorization. The model is rigorously evaluated on three datasets on production settings. Empirical validation confirms that federated multi-view matrix factorization outperforms simpler methods that do not take into account the multi-view structure of the data, in addition, it demonstrates the usefulness of the proposed method for the challenging prediction tasks of cold-start federated recommendations.