Loss Gap Parity for Fairness in Heterogeneous Federated Learning

ArXi:2603.29818v1 Announce Type: new While clients may join federated learning to improve performance on data they rarely observe locally, they often remain self-interested, expecting the global model to perform well on their own data. This motivates an objective that ensures all clients achieve a similar loss gap -the difference in performance between the global model and the best model they could train using only their local data