Aggregation Design for Personalized Federated Multi-Modal Learning over Wireless Networks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yin, Benshun, Chen, Zhiyong, Tao, Meixia
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916283561279488
author Yin, Benshun
Chen, Zhiyong
Tao, Meixia
author_facet Yin, Benshun
Chen, Zhiyong
Tao, Meixia
contents Federated Multi-Modal Learning (FMML) is an emerging field that integrates information from different modalities in federated learning to improve the learning performance. In this letter, we develop a parameter scheduling scheme to improve personalized performance and communication efficiency in personalized FMML, considering the non-independent and nonidentically distributed (non-IID) data along with the modality heterogeneity. Specifically, a learning-based approach is utilized to obtain the aggregation coefficients for parameters of different modalities on distinct devices. Based on the aggregation coefficients and channel state, a subset of parameters is scheduled to be uploaded to a server for each modality. Experimental results show that the proposed algorithm can effectively improve the personalized performance of FMML.
format Preprint
id arxiv_https___arxiv_org_abs_2406_07915
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Aggregation Design for Personalized Federated Multi-Modal Learning over Wireless Networks
Yin, Benshun
Chen, Zhiyong
Tao, Meixia
Information Theory
Signal Processing
Federated Multi-Modal Learning (FMML) is an emerging field that integrates information from different modalities in federated learning to improve the learning performance. In this letter, we develop a parameter scheduling scheme to improve personalized performance and communication efficiency in personalized FMML, considering the non-independent and nonidentically distributed (non-IID) data along with the modality heterogeneity. Specifically, a learning-based approach is utilized to obtain the aggregation coefficients for parameters of different modalities on distinct devices. Based on the aggregation coefficients and channel state, a subset of parameters is scheduled to be uploaded to a server for each modality. Experimental results show that the proposed algorithm can effectively improve the personalized performance of FMML.
title Aggregation Design for Personalized Federated Multi-Modal Learning over Wireless Networks
topic Information Theory
Signal Processing
url https://arxiv.org/abs/2406.07915