Advanced Privacy-Preserving Federated Relationship Recommendation

Federated relationship recommendation is a growing technology trend in future intelligent system. Over the recent years, federated recommendation has recorded an exponential growth, leading to millions of intelligent equipments, and still increasing. However, privacy-preserving is an intractable problem in modern digital systems. In this paper, a comprehensive review of privacy-preserving federated recommendation is presented. Particularly, three kinds of federated recommendation methods are reviewed based on matrix factorization, neural network and other algorithms. Moreover, five authoritative datasets and three general evaluation metrics are mainly described. Simultaneously, a serious of outstanding federated recommender systems with related research discussions are provided. Finally, some potential future works will be tried to improve the application performance.