Analysing Utility Loss in Federated Learning with Differential Privacy | IEEE Conference Publication | IEEE Xplore