A survey on security and privacy of federated learning

综述

联邦学习

1. 联邦学习的基本流程
   
2. 联邦学习的应用

• Google文字预测：mobile applications for next word prediction on keyboards [15–18] like Gboard by Google on Android mobile phones, and wake word detection[19]
• 医疗领域：
  • In medical domains, FL can be utilized to keep patient data private and enhance ML capabilities in assisting medical practitioners similar to the work in [20] which demonstrates the benefits of FL in the medical domain.
  • As for application use-cases in the medical domain, the attack detection in Medical cyber–physical systems that maintain sensitive information on patient’s health records [33], and managing digital health record with FL [34] are more examples of FL applications.
• 通信领域：the research works in [21,22] summarize possible applications for wireless communications with FL by avoiding communication overheads
• 安全领域：malware classification [23], human activity recognition [24], anomaly detection [25], and intrusion detection [26], to name a few.
• 智能交通行业：autonomous cars and driving[27,28],For preventing data leakage in vehicular cyber–physical systems[29],traffic flow prediction[30],and the detection of attacks in aerial vehicles[31]

分类

• 按中心分类
  • 中心化或集群联邦学习
  • 完全去中心联邦学习

---

[1]

Mothukuri, V., Parizi, R. M., Pouriyeh, S., Huang, Y., Dehghantanha, A., & Srivastava, G. (2021). A survey on security and privacy of federated learning. Future Generation Computer Systems-the International Journal of Escience, 115, 619-640. https://doi.org/10.1016/j.future.2020.10.007