P2P Federated Learning using SAC

This project is an implementation of the algorithm proposed in [1].The goal is to simulate peer-to-peer Federated Learning implementing an n-out-of-n secret sharing schema and Secure Average Computation algorithm to let the peers collaborate in the training of a Neural Network, without the risk of direct or indirect training data theft, allowing the presence of semi-honest peers. The project will showcase some emblematic cases probing strengths
and weaknesses of the approach.

[1] Tobias Wink and Zoltan Nochta. An approach for peer-to-peer federated learning. In 2021 51st Annual IEEE/IFIP International Conference on Dependable Systems and Networks Workshops (DSN-W), pages 150–157, 2021.