Privacy-Preserving Three-Factors Authentication and Key Agreement for Federated Learning

Privacy-Preserving Three-Factors Authentication and Key Agreement for Federated Learning

연구 분야: Analysis

학회: International Conference on Machine Learning for Cyber Security

With the development of big data and distributed computing, federated learning has received widespread attention for protecting data privacy. Federated Learning is a distributed machine learning framework that allows multiple participants to collaboratively train shared models with locally retained data. However, the issues of communication security and authentication between participants have emerged as key challenges affecting the security of federal learning systems. To address this challenge, we propose an authentication and key agreement protocol for federated learning environment. The proposed protocol enables authentication between edge devices and server, which establishes a secure session key to secure communication between participants. Additionally, we verify the security of the session key by the verification tool. Finally, experimental evaluation results show that the protocol performs well in terms of computational overhead and communication latency and is suitable for large-scale federated learning systems with resource-constrained devices.