Exploring privacy mechanisms and metrics in federated learning
- This study presents current methods and strategies to preserve privacy in both centralized and decentralized federated learning(FL) environments. It provides comparative analysis of FL architectures.
- It discusses the critical role of privacy enhancing technologies like Differential privacy (DP), Homomorphic encryption (HE) and Secure multiparty computation (SMPC) in federated learning with evaluation metrics.
- It outlines the applications of FL in the field of Natural language processing (NLP), Healthcare and Internet of Things (IoT) with Edge computing and gives domain-specific insights
- It addresses metrics that are essential in analyzing privacy and performance aspects of the FL system like Correctness vs. Privacy Trade-off, Communication Overhead, Computational Cost, Scalability and Resistance to Attacks.
- It identifies the areas of concern in FL that can be explored as future research direction.
