DA34FL: a robust dynamic accumulator-based authentication and key agreement with preserving model training data integrity for federated learning.

Federated Learning (FL) offers a privacy-preserving distributed learning paradigm by enabling model training without direct access to raw data. However, FL remains vulnerable to unauthorized access during training and client-server exchanges. Authent…

Preview only. Read the full abstract at the source