Efficient Protocols for the Secure Transmission of Training Parameters in a Federated Network: Research Frontiers, Taxonomy, and Future Directions

Federated Learning (FL), which ensures data privacy while jointly increasing model accuracy, has emerged as a viable method for collaborative machine learning across decentralized and heterogeneous data sources. However, because of differing computational capacities and security requirements, securely communicating training parameters across heterogeneous participants continues to be a significant difficulty. The present study offers an extensive overview of the state of research in the field and employs a taxonomy based on cryptography techniques, including Differential Privacy (DP), Homomorphic Encryption (HE), Secure Multiparty Computation (SMPC), and Blockchain technology, to classify the current secure transmission protocols. Additionally, it suggests future lines of inquiry for optimization and hybrid approaches to improve security, scalability, and efficiency in heterogeneous federated networks. For scholars and practitioners looking to propel the field of FL into more robust and privacy-preserving collaborative learning settings, the insights and taxonomy offered provide a road map