KL-FedDis: A federated learning approach with distribution information sharing using Kullback-Leibler divergence for non-IID data

Data Heterogeneity or Non-IID (non-independent and identically distributed) data identification is one of the prominent challenges in Federated Learning (FL). In Non-IID data, clients have their own local data, which may not be independently and identically distributed. This arises because clients i...

Full description

Saved in:
Bibliographic Details
Main Authors: Md. Rahad, Ruhan Shabab, Mohd. Sultan Ahammad, Md. Mahfuz Reza, Amit Karmaker, Md. Abir Hossain
Format: Article
Language:English
Published: Elsevier 2025-03-01
Series:Neuroscience Informatics
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S277252862400027X
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850026469702500352
author Md. Rahad
Ruhan Shabab
Mohd. Sultan Ahammad
Md. Mahfuz Reza
Amit Karmaker
Md. Abir Hossain
author_facet Md. Rahad
Ruhan Shabab
Mohd. Sultan Ahammad
Md. Mahfuz Reza
Amit Karmaker
Md. Abir Hossain
author_sort Md. Rahad
collection DOAJ
description Data Heterogeneity or Non-IID (non-independent and identically distributed) data identification is one of the prominent challenges in Federated Learning (FL). In Non-IID data, clients have their own local data, which may not be independently and identically distributed. This arises because clients involved in federated learning typically have their own unique, local datasets that vary significantly due to factors like geographical location, user behaviors, or specific contexts. Model divergence is another critical challenge where the local models trained on different clients, data may diverge significantly but making it difficult for the global model to converge. To identify the non-IID data, few federated learning models have been introduced as FedDis, FedProx and FedAvg, but their accuracy is too low. To address the clients Non-IID data along with ensuring privacy, federated learning emerged with appropriate distribution mechanism is an effective solution. In this paper, a modified FedDis learning method called KL-FedDis is proposed, which incorporates Kullback-Leibler (KL) divergence as the regularization technique. KL-FedDis improves accuracy and computation time over the FedDis and FedAvg technique by successfully maintaining the distribution information and encouraging improved collaboration among the local models by utilizing KL divergence.
format Article
id doaj-art-3c9d05dae92349cdb978ccbb4265fa9a
institution DOAJ
issn 2772-5286
language English
publishDate 2025-03-01
publisher Elsevier
record_format Article
series Neuroscience Informatics
spelling doaj-art-3c9d05dae92349cdb978ccbb4265fa9a2025-08-20T03:00:32ZengElsevierNeuroscience Informatics2772-52862025-03-015110018210.1016/j.neuri.2024.100182KL-FedDis: A federated learning approach with distribution information sharing using Kullback-Leibler divergence for non-IID dataMd. Rahad0Ruhan Shabab1Mohd. Sultan Ahammad2Md. Mahfuz Reza3Amit Karmaker4Md. Abir Hossain5Computer Science & Engineering, Mawlana Bhashani Science and Technology University, Tangail, 1902, BangladeshComputer Science & Engineering, Mawlana Bhashani Science and Technology University, Tangail, 1902, BangladeshComputer Science & Engineering, Mawlana Bhashani Science and Technology University, Tangail, 1902, BangladeshComputer Science & Engineering, Mawlana Bhashani Science and Technology University, Tangail, 1902, BangladeshInstitute of Information & Communication Technology, Bangladesh University of Engineering and Technology, Dhaka, 1205, BangladeshInformation & Communication Technology, Mawlana Bhashani Science and Technology University, Tangail, 1902, Bangladesh; Institute of Information & Communication Technology, Bangladesh University of Engineering and Technology, Dhaka, 1205, Bangladesh; Corresponding author at: Information & Communication Technology, Mawlana Bhashani Science and Technology University, Tangail, 1902, Bangladesh.Data Heterogeneity or Non-IID (non-independent and identically distributed) data identification is one of the prominent challenges in Federated Learning (FL). In Non-IID data, clients have their own local data, which may not be independently and identically distributed. This arises because clients involved in federated learning typically have their own unique, local datasets that vary significantly due to factors like geographical location, user behaviors, or specific contexts. Model divergence is another critical challenge where the local models trained on different clients, data may diverge significantly but making it difficult for the global model to converge. To identify the non-IID data, few federated learning models have been introduced as FedDis, FedProx and FedAvg, but their accuracy is too low. To address the clients Non-IID data along with ensuring privacy, federated learning emerged with appropriate distribution mechanism is an effective solution. In this paper, a modified FedDis learning method called KL-FedDis is proposed, which incorporates Kullback-Leibler (KL) divergence as the regularization technique. KL-FedDis improves accuracy and computation time over the FedDis and FedAvg technique by successfully maintaining the distribution information and encouraging improved collaboration among the local models by utilizing KL divergence.http://www.sciencedirect.com/science/article/pii/S277252862400027XFederated learningFedDisKL divergenceRegularization
spellingShingle Md. Rahad
Ruhan Shabab
Mohd. Sultan Ahammad
Md. Mahfuz Reza
Amit Karmaker
Md. Abir Hossain
KL-FedDis: A federated learning approach with distribution information sharing using Kullback-Leibler divergence for non-IID data
Neuroscience Informatics
Federated learning
FedDis
KL divergence
Regularization
title KL-FedDis: A federated learning approach with distribution information sharing using Kullback-Leibler divergence for non-IID data
title_full KL-FedDis: A federated learning approach with distribution information sharing using Kullback-Leibler divergence for non-IID data
title_fullStr KL-FedDis: A federated learning approach with distribution information sharing using Kullback-Leibler divergence for non-IID data
title_full_unstemmed KL-FedDis: A federated learning approach with distribution information sharing using Kullback-Leibler divergence for non-IID data
title_short KL-FedDis: A federated learning approach with distribution information sharing using Kullback-Leibler divergence for non-IID data
title_sort kl feddis a federated learning approach with distribution information sharing using kullback leibler divergence for non iid data
topic Federated learning
FedDis
KL divergence
Regularization
url http://www.sciencedirect.com/science/article/pii/S277252862400027X
work_keys_str_mv AT mdrahad klfeddisafederatedlearningapproachwithdistributioninformationsharingusingkullbackleiblerdivergencefornoniiddata
AT ruhanshabab klfeddisafederatedlearningapproachwithdistributioninformationsharingusingkullbackleiblerdivergencefornoniiddata
AT mohdsultanahammad klfeddisafederatedlearningapproachwithdistributioninformationsharingusingkullbackleiblerdivergencefornoniiddata
AT mdmahfuzreza klfeddisafederatedlearningapproachwithdistributioninformationsharingusingkullbackleiblerdivergencefornoniiddata
AT amitkarmaker klfeddisafederatedlearningapproachwithdistributioninformationsharingusingkullbackleiblerdivergencefornoniiddata
AT mdabirhossain klfeddisafederatedlearningapproachwithdistributioninformationsharingusingkullbackleiblerdivergencefornoniiddata