A research team has developed a new one-shot federated learning artificial intelligence (AI) technique that enables efficient training of large-scale models without sharing personal information. This research outcome is extremely significant, as it demonstrates that privacy protection, learning efficiency, and model performance can be secured simultaneously in the field of medical image analysis.