Synthetic Data for Anonymization in Secure Data Spaces for Federated Learning
Cecilio Ángulo, Cristóbal Raya Giner · Frontiers in artificial intelligence and applications · 2022
Federated learning implies the integration of shared data. Privacy-enforcing platforms should be implemented to provide a secure environment for federated learning. We are proposing the integration of real world data from local data lakes and the generation and use of general synthetic data to simplify, eventually avoid, encryption or differential learning and use general architectures for data spaces.