Federated Multi-view Matrix Factorization for Personalized Recommendations

We introduce the federated multi-view matrix factorization method that learns a multi-view model without transferring the user's personal data to a central server. The method extends the federated learning framework to matrix factorization with multiple data sources. As far as we are aware, this is the first federated model to provide recommendations using multi-view matrix factorization. In addition, it is the first method to provide federated cold-start recommendations. 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, we also demonstrate the usefulness of the proposed method for the challenging prediction task of cold-start federated recommendations.