FedADMM-InSa: An Inexact and Self-Adaptive ADMM for Federated Learning

arXiv:2402.13989v1 Announce Type: new
Abstract: Federated learning (FL) is a promising framework for learning from distributed data while maintaining privacy. The development of efficient FL algorithms encounters various challenges, including heterogeneous data and systems, limited communication capacities, and constrained local computational resources. Recently developed FedADMM methods show great resilience to both data and system heterogeneity. However, they still suffer from performance deterioration if the hyperparameters are not carefully tuned. To address this issue, we propose an inexact and self-adaptive FedADMM …

abstract algorithms arxiv challenges communication computational cs.cr cs.dc cs.lg data development distributed distributed data federated learning framework math.oc privacy resilience resources show systems type