Part 3/9:
At the core of this research is a revolutionary approach known as Open Federated Learning (Open FL), pioneered by Intel. Unlike traditional AI training methods, where data is pulled into a central server to train models, federated learning brings the model to the data. This shift allows multiple data contributors—such as the Mayo Clinic, NASA’s Gene Lab, and various collaborators handling human and mouse data—to train models locally, sharing only model updates with a central aggregator.
Open FL addresses major concerns around data privacy and security, enabling scientists to develop powerful, generalizable AI models without violating regulations like HIPAA. This is especially critical when working with sensitive health data across multiple institutions and jurisdictions.