Train AI without sharing data

The data we generate when using applications is used to train artificial intelligence, but privacy can be a problem. Federated studies are being carried out at Mondragon Unibertsitatea. Instead of sending data in federated learning, artificial intelligence is trained on our own device, and only mathematical variables are shared. In this way, privacy is guaranteed, the system is continuously improved and adapted to each user. In addition, collaboration between organizations that handle sensitive data becomes possible.