Study of a Cross-Platform Solution for Applied Federated Learning

Federated Learning is an on-growing technique in the Machine Learning field: its two strong points are the data privacy granted by the collaborative learning paradigm and the possibility to train on edge devices directly. This topic is rapidly growing thanks to the importance of data privacy and also thanks to the chips mounted on edge devices that reach higher and higher performances. For this application the smartphone sounds like the perfect fit, it is a device that stores an incredible amount of data, that might help to build an Artificial Intelligence model, but cannot, due to privacy limitations. At the moment, there are no tools specific for edge training: in this project, there will be an extensive analysis of the available technologies and how they can be used for creating an application that can be used in a real-life scenario.