Federated Neural Collaborative Filtering

In this work, we present a federated version of the state-of-the-art Neural Collaborative Filtering (NCF) approach for item recommendations. The system, named FedNCF, enables learning without requiring users to disclose or transmit their raw data. Data localization preserves data privacy and complies with regulations such as GDPR. Although model training local dissemination, transmission clients’ updates raises additional issues. To address challenge, incorporate privacy-preserving aggregation method that satisfies security requirements against an honest but curious entity. We argue theoretically experimentally existing algorithms are inconsistent latent factor updates. propose enhancement by decomposing step into matrix factorization neural network-based averaging. Experimental validation shows FedNCF achieves comparable recommendation quality original NCF while our proposed leads faster convergence compared methods. investigate effectiveness recommender system evaluate mechanism in terms computational cost.