Federated Learning-Based Approach for Effective Educational Resource Recommendation

In recent years, technology-enhanced learning has affected the whole educational process by introducing innovative learner-centric tools; among them are educational recommender systems. It helps to access educational resources on the web in a flexible, suggestion-based way. This paper focuses on improving the accuracy of recommendations by proposing a novel hybrid approach based on the aggregation of several techniques and models. The proposed approach is grounded in the federated learning model, which involves replicating servers running different recommendation techniques and setting up a coordination mechanism to achieve consensus among them. Finally, it’s obvious that the proposed approach holds promise for overcoming the main recommendation hurdles, such as cold starts and data sparsity, and consequently improving its accuracy.