Scholarlink

Federated Learning (FL) has emerged as a promising solution within recommender systems, driven by the growing user demand for privacy-centric design. FL not only meets privacy needs but also enables precise, personalized recommendations while reducing the exposure of user preferences and interactions to external servers without explicit consent. This chapter provides a comprehensive review of literature on the use of FL in recommendation applications to safeguard user privacy. It highlights several groundbreaking methods that have transformed federated recommender systems. These include the use of cryptographic techniques to protect against private information leakage from adversarial attacks, the adoption of blockchain architectures for decentralization, and the implementation of graph-based models to better understand complex user-item interactions, thereby improving recommendation accuracy. The chapter discusses the horizon of FL in recommendation systems, examining open problems and limitations. This exploration casts a light on the potential future directions and innovations that can further fortify the synergy between FL and recommender systems.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Federated Learning for Recommender Systems: Advances and Perspectives

  • 章节

  • 发表日期：

    2025年

查看全文
  • Vasileios Perifanis,
  • Nikolaos Pavlidis,
  • Andreas Sendros,
  • Pavlos S. Efraimidis

摘要

Federated Learning (FL) has emerged as a promising solution within recommender systems, driven by the growing user demand for privacy-centric design. FL not only meets privacy needs but also enables precise, personalized recommendations while reducing the exposure of user preferences and interactions to external servers without explicit consent. This chapter provides a comprehensive review of literature on the use of FL in recommendation applications to safeguard user privacy. It highlights several groundbreaking methods that have transformed federated recommender systems. These include the use of cryptographic techniques to protect against private information leakage from adversarial attacks, the adoption of blockchain architectures for decentralization, and the implementation of graph-based models to better understand complex user-item interactions, thereby improving recommendation accuracy. The chapter discusses the horizon of FL in recommendation systems, examining open problems and limitations. This exploration casts a light on the potential future directions and innovations that can further fortify the synergy between FL and recommender systems.