FedRP: A Communication-Efficient Approach for Differentially Private Federated Learning Using Random Projection
The episode discusses a paper titled "FedRP: A Communication-Efficient Approach for Differentially Private Federated Learning Using Random Projection." The hosts explain how FedRP uses random projection and ADMM optimization to reduce communication costs while maintaining differential privacy. They conclude that this method offers high model accuracy, strong privacy guarantees, and is a viable path for scalable, secure global AI systems.