Federated Learning With Dataset Splitting and Weighted Mean Using Particle Swarm Optimization

Federated learning uses the concept of decentralized training of n number of local clients for a small number of epochs say 2-5, and then averaging the learned weights of all local clients, and evaluating on test dataset with the average weights loaded to a global model. The train dataset is split i...

Full description

Saved in:
Bibliographic Details
Main Authors: Mohit Agarwal, Garima Jaiswal, Rohit Kumar Kaliyar, Akansha Singh, Krishna Kant Singh, S. S. Askar, Mohamed Abouhawwash
Format: Article
Language:English
Published: IEEE 2024-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10731716/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850199186803261440
author Mohit Agarwal
Garima Jaiswal
Rohit Kumar Kaliyar
Akansha Singh
Krishna Kant Singh
S. S. Askar
Mohamed Abouhawwash
author_facet Mohit Agarwal
Garima Jaiswal
Rohit Kumar Kaliyar
Akansha Singh
Krishna Kant Singh
S. S. Askar
Mohamed Abouhawwash
author_sort Mohit Agarwal
collection DOAJ
description Federated learning uses the concept of decentralized training of n number of local clients for a small number of epochs say 2-5, and then averaging the learned weights of all local clients, and evaluating on test dataset with the average weights loaded to a global model. The train dataset is split into n clusters and each cluster acts as a distributed data for each local model. Each round of weight averaging and then uploading the average weights on each local client for further training is called communication round and it was observed that similar accuracy can be obtained with a lesser amount of training time. In this paper, instead of averaging the weights, a weighted mean concept was developed where the PSO vector helps to find the weight values for the best accuracy of a global model. It was found that PSO can help in two ways by bettering the accuracy and also reducing the training time. The proposed approach can enhance the performance of pre-trained models like AlexNet, VGG16, InceptionV3, and ResNet50 on CIFAR-10 and CIFAR-100 datasets. The maximum increase was found with VGG16 of around 26.01% for CIFAR-10 and 26.84% for CIFAR-100. Similarly, on the Tomato dataset, AlexNet accuracy can be increased by 28.56%. Multi-modal model accuracy on the fake news dataset was also enhanced by 8.21%.
format Article
id doaj-art-ca53c90719244b7ba72f1ce0117bbbe9
institution OA Journals
issn 2169-3536
language English
publishDate 2024-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-ca53c90719244b7ba72f1ce0117bbbe92025-08-20T02:12:41ZengIEEEIEEE Access2169-35362024-01-011216196816198110.1109/ACCESS.2024.348510010731716Federated Learning With Dataset Splitting and Weighted Mean Using Particle Swarm OptimizationMohit Agarwal0https://orcid.org/0000-0001-7216-3886Garima Jaiswal1https://orcid.org/0000-0002-2676-299XRohit Kumar Kaliyar2https://orcid.org/0000-0001-7233-872XAkansha Singh3https://orcid.org/0000-0002-5520-8066Krishna Kant Singh4https://orcid.org/0000-0002-6510-6768S. S. Askar5https://orcid.org/0000-0002-1167-2430Mohamed Abouhawwash6https://orcid.org/0000-0003-2846-4707School of CSET, Bennett University, Greater Noida, IndiaSchool of CSET, Bennett University, Greater Noida, IndiaSchool of CSET, Bennett University, Greater Noida, IndiaSchool of CSET, Bennett University, Greater Noida, IndiaDelhi Technical Campus, Greater Noida, IndiaDepartment of Statistics and Operations Research, College of Science, King Saud University, Riyadh, Saudi ArabiaDepartment of Mathematics, Faculty of Science, Mansoura University, Mansoura, EgyptFederated learning uses the concept of decentralized training of n number of local clients for a small number of epochs say 2-5, and then averaging the learned weights of all local clients, and evaluating on test dataset with the average weights loaded to a global model. The train dataset is split into n clusters and each cluster acts as a distributed data for each local model. Each round of weight averaging and then uploading the average weights on each local client for further training is called communication round and it was observed that similar accuracy can be obtained with a lesser amount of training time. In this paper, instead of averaging the weights, a weighted mean concept was developed where the PSO vector helps to find the weight values for the best accuracy of a global model. It was found that PSO can help in two ways by bettering the accuracy and also reducing the training time. The proposed approach can enhance the performance of pre-trained models like AlexNet, VGG16, InceptionV3, and ResNet50 on CIFAR-10 and CIFAR-100 datasets. The maximum increase was found with VGG16 of around 26.01% for CIFAR-10 and 26.84% for CIFAR-100. Similarly, on the Tomato dataset, AlexNet accuracy can be increased by 28.56%. Multi-modal model accuracy on the fake news dataset was also enhanced by 8.21%.https://ieeexplore.ieee.org/document/10731716/Federated learningparticle swarm optimizationoptimizationmodel performancemultimodal
spellingShingle Mohit Agarwal
Garima Jaiswal
Rohit Kumar Kaliyar
Akansha Singh
Krishna Kant Singh
S. S. Askar
Mohamed Abouhawwash
Federated Learning With Dataset Splitting and Weighted Mean Using Particle Swarm Optimization
IEEE Access
Federated learning
particle swarm optimization
optimization
model performance
multimodal
title Federated Learning With Dataset Splitting and Weighted Mean Using Particle Swarm Optimization
title_full Federated Learning With Dataset Splitting and Weighted Mean Using Particle Swarm Optimization
title_fullStr Federated Learning With Dataset Splitting and Weighted Mean Using Particle Swarm Optimization
title_full_unstemmed Federated Learning With Dataset Splitting and Weighted Mean Using Particle Swarm Optimization
title_short Federated Learning With Dataset Splitting and Weighted Mean Using Particle Swarm Optimization
title_sort federated learning with dataset splitting and weighted mean using particle swarm optimization
topic Federated learning
particle swarm optimization
optimization
model performance
multimodal
url https://ieeexplore.ieee.org/document/10731716/
work_keys_str_mv AT mohitagarwal federatedlearningwithdatasetsplittingandweightedmeanusingparticleswarmoptimization
AT garimajaiswal federatedlearningwithdatasetsplittingandweightedmeanusingparticleswarmoptimization
AT rohitkumarkaliyar federatedlearningwithdatasetsplittingandweightedmeanusingparticleswarmoptimization
AT akanshasingh federatedlearningwithdatasetsplittingandweightedmeanusingparticleswarmoptimization
AT krishnakantsingh federatedlearningwithdatasetsplittingandweightedmeanusingparticleswarmoptimization
AT ssaskar federatedlearningwithdatasetsplittingandweightedmeanusingparticleswarmoptimization
AT mohamedabouhawwash federatedlearningwithdatasetsplittingandweightedmeanusingparticleswarmoptimization