Federated Learning (FL) will be a common approach used for many AI applications, where different devices collaborate on the development of ML models without sharing their private data.

FL approach will enable different wireless networks to collaborate to devise more intelligent solutions, which can also be used to develop energy-efficient solutions thanks to the faster convergence.

However, this will bring several challenges, which at UPC we would like to investigate.

We have been involved in EU projects developing FL extensions to accelerator-based Deep Learning frameworks, 5G IoT systems using FL, etc.

Looking forward to devise the AI solutions of the future together...