AGIC: Approximate Gradient Inversion Attack on Federated Learning

Jin Xu, Chi Hong, Jiyue Huang +2

Federated learning is a private-by-design distributed learning paradigm where clients train local models on their own data before a central server aggregates their local updates to…