Federated Recommendation

Federated recommendation aims to build personalized recommendation systems while preserving user privacy by performing model training on decentralized data. Current research focuses on enhancing privacy through techniques like differential privacy and blockchain-based traceability, addressing data sparsity and heterogeneity with methods such as hybrid retrieval augmented generation and knowledge graph enhancement, and improving efficiency via low-rank training and update prediction. This field is significant for enabling collaborative recommendation across multiple data sources while adhering to privacy regulations, impacting both the development of more ethical AI systems and the practical deployment of personalized services.