Federated Learning under Covariate Shifts with Generalization Guarantees
It addresses data distribution shifts across clients in federated learning, which is an incremental improvement for privacy-sensitive applications.
The paper tackles covariate shifts in federated learning by proposing FTW-ERM, a method that improves generalization error over classical ERM, with experimental results showing superiority in imbalanced settings.
This paper addresses intra-client and inter-client covariate shifts in federated learning (FL) with a focus on the overall generalization performance. To handle covariate shifts, we formulate a new global model training paradigm and propose Federated Importance-Weighted Empirical Risk Minimization (FTW-ERM) along with improving density ratio matching methods without requiring perfect knowledge of the supremum over true ratios. We also propose the communication-efficient variant FITW-ERM with the same level of privacy guarantees as those of classical ERM in FL. We theoretically show that FTW-ERM achieves smaller generalization error than classical ERM under certain settings. Experimental results demonstrate the superiority of FTW-ERM over existing FL baselines in challenging imbalanced federated settings in terms of data distribution shifts across clients.