Gradient-Congruity Guided Federated Sparse Training

arXiv:2405.01189v1 Announce Type: new
Abstract: Edge computing allows artificial intelligence and machine learning models to be deployed on edge devices, where they can learn from local data and collaborate to form a global model. Federated learning (FL) is a distributed machine learning technique that facilitates this process while preserving data privacy. However, FL also faces challenges such as high computational and communication costs regarding resource-constrained devices, and poor generalization performance due to the heterogeneity of data across edge clients and …

abstract artificial artificial intelligence arxiv computing cs.ai cs.lg data data privacy devices distributed edge edge computing edge devices federated learning form global gradient however intelligence learn machine machine learning machine learning models privacy process training type while