Bayesian Federated Model Compression for Communication and Computation Efficiency

arXiv:2404.07532v1 Announce Type: new
Abstract: In this paper, we investigate Bayesian model compression in federated learning (FL) to construct sparse models that can achieve both communication and computation efficiencies. We propose a decentralized Turbo variational Bayesian inference (D-Turbo-VBI) FL framework where we firstly propose a hierarchical sparse prior to promote a clustered sparse structure in the weight matrix. Then, by carefully integrating message passing and VBI with a decentralized turbo framework, we propose the D-Turbo-VBI algorithm which can (i) reduce …

abstract arxiv bayesian bayesian inference communication compression computation construct cs.ai cs.dc cs.lg decentralized efficiency federated learning framework hierarchical inference paper prior promote turbo type