From Algorithm to Application: Enhancing Federated Learning with Adaptive Aggregation and Gradient Protection

HU, YI

doi:10.23889/SUThesis.71338

E-Thesis 33 views 11 downloads

From Algorithm to Application: Enhancing Federated Learning with Adaptive Aggregation and Gradient Protection / YI HU

Swansea University Author: YI HU

PDF | E-Thesis – open access

Copyright: the author, Yi Hu, 2025. Distributed under the terms of a Creative Commons Attribution 4.0 License (CC BY 4.0)
Download (6.5MB)

DOI (Published version): 10.23889/SUThesis.71338

Abstract

Federated Learning (FL) has emerged as a promising paradigm for decentralised machine learning, allowing multiple clients to collaboratively train models without sharing their private data. This distributed approach is particularly relevant in domains with stringent data privacy requirements, such a...

Full description

Published:	Swansea University 2025
Institution:	Swansea University
Degree level:	Doctoral
Degree name:	Ph.D
Supervisor:	Xie, X
URI:	https://cronfa.swan.ac.uk/Record/cronfa71338

first_indexed	2026-01-29T12:40:20Z
last_indexed	2026-01-30T06:53:11Z
id	cronfa71338
recordtype	RisThesis
fullrecord	<?xml version="1.0"?><rfc1807><datestamp>2026-01-29T12:50:48.6696046</datestamp><bib-version>v2</bib-version><id>71338</id><entry>2026-01-29</entry><title>From Algorithm to Application: Enhancing Federated Learning with Adaptive Aggregation and Gradient Protection</title><swanseaauthors><author><sid>c724e9f97d8ed76a8c0b3ea26bbc43a3</sid><firstname>YI</firstname><surname>HU</surname><name>YI HU</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2026-01-29</date><abstract>Federated Learning (FL) has emerged as a promising paradigm for decentralised machine learning, allowing multiple clients to collaboratively train models without sharing their private data. This distributed approach is particularly relevant in domains with stringent data privacy requirements, such as ﬁnance, healthcare, and edge computing. However, despite its advantages, FL faces three critical challenges: (1) efﬁcient model aggregation, (2) protection against privacy leakage, and (3) real-world applicability in complex domains. This thesis addresses these challenges by proposing novel strategies to optimise aggregation mechanisms, enhance its privacy protection capabilities, and explore its application in ﬁnancial modelling. First, we introduce Element-Wise Weights Aggregation for FL (EWWA-FL), a novel optimisation technique that improves global model convergence by adopting element-wise adaptive weighting. Traditional FL aggregation methods, such as FedAvg, assign a single proportion to each local model without considering the varying importance of individual model parameters. In contrast, EWWA-FL assigns unique proportions to each element within local model weights, ensuring more precise updates that account for dataset heterogeneity among clients. Experimental results demonstrate that EWWA-FL signiﬁcantly improves both convergence speed and ﬁnal model accuracy, outperforming FedAvg, FedOpt, and FedCAMS across multiple benchmark datasets. By incorporating an element-wise approach, EWWA-FL provides a more adaptive and ﬁne-grained aggregation strategy that enhances FL’s performance in both Independent and Identically Distributed (IID) and Non-IID settings. Second, we propose AdaDefence, a privacy-preserving defence mechanism against gradient leakage attacks in FL. While FL eliminates the need for raw data sharing, recent attacks have demonstrated that an adversary can reconstruct private training data from shared gradients, posing a severe privacy risk. To counteract this, AdaDefence introduces a gradient stand-in approach, wherein local clients replace actual gradients with modiﬁed gradients before sending them to the server. This method prevents attackers from reconstructing private data while maintaining model utility. AdaDefence effectively defends against state-of-the-art attacks such as Deep Leakage from Gradients (DLG), Generative Regression Neural Network (GRNN), and Inverting Gradient (IG) without signiﬁcantly compromising model accuracy. Our extensive empirical analysis shows that AdaDefence provides strong privacy guarantees while ensuring minimal performance degradation, making it a practical and scalable solution for real-world FL deployments. Finally, we explore the real-world application of FL in ﬁnancial modelling, particularly in Cross-Stock Trend Integration (CSTI) for enhancing stock price prediction. Traditional ﬁnancial models suffer from data fragmentation, where different ﬁnancial institutions and stock markets operate in silos, limiting predictive power. To overcome this, we develop a FL-based approach that enables multiple ﬁnancial institutions to collaboratively train stock prediction models without exposing sensitive trading data. This approach leverages cross-stock trend integration, allowing predictive models to learn patterns from multiple stocks while preserving privacy.Our experimental results demonstrate that federated cross-stock learning improves predictive accuracy and model robustness, outperforming conventional single-stock prediction methods.By enabling secure, multi-institution collaboration, this work highlights the potential of FL in advancing ﬁnancial modelling while ensuring regulatory compliance and data conﬁdentiality. By addressing these fundamental aspects, optimisation and protection, this thesis makes signiﬁcant contributions to the ﬁeld of FL. The proposed methodologies collectively enhance FL’s efﬁciency, security, and real-world applicability. Through EWWA-FL, FL models achieve faster and more reliable convergence. By introducing AdaDefence, FL gains stronger privacy protections against gradient-based attacks. Finally, by demonstrating FL’s potential in cross-stock trend integration, this thesis showcases how FL can be deployed in privacy-sensitive ﬁnancial applications. These contributions pave the way for more efﬁcient, secure, and scalable FL systems, advancing their adoption in a wide range of domains, including healthcare, autonomous systems, and ﬁnancial technology.</abstract><type>E-Thesis</type><journal/><volume/><journalNumber/><paginationStart/><paginationEnd/><publisher/><placeOfPublication>Swansea University</placeOfPublication><isbnPrint/><isbnElectronic/><issnPrint/><issnElectronic/><keywords>Deep Learning, Federated Learning, Finance</keywords><publishedDay>7</publishedDay><publishedMonth>11</publishedMonth><publishedYear>2025</publishedYear><publishedDate>2025-11-07</publishedDate><doi>10.23889/SUThesis.71338</doi><url/><notes/><college>COLLEGE NANME</college><CollegeCode>COLLEGE CODE</CollegeCode><institution>Swansea University</institution><supervisor>Xie, X</supervisor><degreelevel>Doctoral</degreelevel><degreename>Ph.D</degreename><apcterm/><funders/><projectreference/><lastEdited>2026-01-29T12:50:48.6696046</lastEdited><Created>2026-01-29T12:35:37.6438235</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Mathematics and Computer Science - Computer Science</level></path><authors><author><firstname>YI</firstname><surname>HU</surname><order>1</order></author></authors><documents><document><filename>71338__36140__55aad54dc2ce443ebb020f4d633cc188.pdf</filename><originalFilename>2025_Hu_Y.final.71338.pdf</originalFilename><uploaded>2026-01-29T12:50:17.8875470</uploaded><type>Output</type><contentLength>6815072</contentLength><contentType>application/pdf</contentType><version>E-Thesis – open access</version><cronfaStatus>true</cronfaStatus><documentNotes>Copyright: the author, Yi Hu, 2025. Distributed under the terms of a Creative Commons Attribution 4.0 License (CC BY 4.0)</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language><licence>https://creativecommons.org/licenses/by/4.0/</licence></document></documents><OutputDurs/></rfc1807>
spelling	2026-01-29T12:50:48.6696046 v2 71338 2026-01-29 From Algorithm to Application: Enhancing Federated Learning with Adaptive Aggregation and Gradient Protection c724e9f97d8ed76a8c0b3ea26bbc43a3 YI HU YI HU true false 2026-01-29 Federated Learning (FL) has emerged as a promising paradigm for decentralised machine learning, allowing multiple clients to collaboratively train models without sharing their private data. This distributed approach is particularly relevant in domains with stringent data privacy requirements, such as ﬁnance, healthcare, and edge computing. However, despite its advantages, FL faces three critical challenges: (1) efﬁcient model aggregation, (2) protection against privacy leakage, and (3) real-world applicability in complex domains. This thesis addresses these challenges by proposing novel strategies to optimise aggregation mechanisms, enhance its privacy protection capabilities, and explore its application in ﬁnancial modelling. First, we introduce Element-Wise Weights Aggregation for FL (EWWA-FL), a novel optimisation technique that improves global model convergence by adopting element-wise adaptive weighting. Traditional FL aggregation methods, such as FedAvg, assign a single proportion to each local model without considering the varying importance of individual model parameters. In contrast, EWWA-FL assigns unique proportions to each element within local model weights, ensuring more precise updates that account for dataset heterogeneity among clients. Experimental results demonstrate that EWWA-FL signiﬁcantly improves both convergence speed and ﬁnal model accuracy, outperforming FedAvg, FedOpt, and FedCAMS across multiple benchmark datasets. By incorporating an element-wise approach, EWWA-FL provides a more adaptive and ﬁne-grained aggregation strategy that enhances FL’s performance in both Independent and Identically Distributed (IID) and Non-IID settings. Second, we propose AdaDefence, a privacy-preserving defence mechanism against gradient leakage attacks in FL. While FL eliminates the need for raw data sharing, recent attacks have demonstrated that an adversary can reconstruct private training data from shared gradients, posing a severe privacy risk. To counteract this, AdaDefence introduces a gradient stand-in approach, wherein local clients replace actual gradients with modiﬁed gradients before sending them to the server. This method prevents attackers from reconstructing private data while maintaining model utility. AdaDefence effectively defends against state-of-the-art attacks such as Deep Leakage from Gradients (DLG), Generative Regression Neural Network (GRNN), and Inverting Gradient (IG) without signiﬁcantly compromising model accuracy. Our extensive empirical analysis shows that AdaDefence provides strong privacy guarantees while ensuring minimal performance degradation, making it a practical and scalable solution for real-world FL deployments. Finally, we explore the real-world application of FL in ﬁnancial modelling, particularly in Cross-Stock Trend Integration (CSTI) for enhancing stock price prediction. Traditional ﬁnancial models suffer from data fragmentation, where different ﬁnancial institutions and stock markets operate in silos, limiting predictive power. To overcome this, we develop a FL-based approach that enables multiple ﬁnancial institutions to collaboratively train stock prediction models without exposing sensitive trading data. This approach leverages cross-stock trend integration, allowing predictive models to learn patterns from multiple stocks while preserving privacy.Our experimental results demonstrate that federated cross-stock learning improves predictive accuracy and model robustness, outperforming conventional single-stock prediction methods.By enabling secure, multi-institution collaboration, this work highlights the potential of FL in advancing ﬁnancial modelling while ensuring regulatory compliance and data conﬁdentiality. By addressing these fundamental aspects, optimisation and protection, this thesis makes signiﬁcant contributions to the ﬁeld of FL. The proposed methodologies collectively enhance FL’s efﬁciency, security, and real-world applicability. Through EWWA-FL, FL models achieve faster and more reliable convergence. By introducing AdaDefence, FL gains stronger privacy protections against gradient-based attacks. Finally, by demonstrating FL’s potential in cross-stock trend integration, this thesis showcases how FL can be deployed in privacy-sensitive ﬁnancial applications. These contributions pave the way for more efﬁcient, secure, and scalable FL systems, advancing their adoption in a wide range of domains, including healthcare, autonomous systems, and ﬁnancial technology. E-Thesis Swansea University Deep Learning, Federated Learning, Finance 7 11 2025 2025-11-07 10.23889/SUThesis.71338 COLLEGE NANME COLLEGE CODE Swansea University Xie, X Doctoral Ph.D 2026-01-29T12:50:48.6696046 2026-01-29T12:35:37.6438235 Faculty of Science and Engineering School of Mathematics and Computer Science - Computer Science YI HU 1 71338__36140__55aad54dc2ce443ebb020f4d633cc188.pdf 2025_Hu_Y.final.71338.pdf 2026-01-29T12:50:17.8875470 Output 6815072 application/pdf E-Thesis – open access true Copyright: the author, Yi Hu, 2025. Distributed under the terms of a Creative Commons Attribution 4.0 License (CC BY 4.0) true eng https://creativecommons.org/licenses/by/4.0/
title	From Algorithm to Application: Enhancing Federated Learning with Adaptive Aggregation and Gradient Protection
spellingShingle	From Algorithm to Application: Enhancing Federated Learning with Adaptive Aggregation and Gradient Protection YI HU
title_short	From Algorithm to Application: Enhancing Federated Learning with Adaptive Aggregation and Gradient Protection
title_full	From Algorithm to Application: Enhancing Federated Learning with Adaptive Aggregation and Gradient Protection
title_fullStr	From Algorithm to Application: Enhancing Federated Learning with Adaptive Aggregation and Gradient Protection
title_full_unstemmed	From Algorithm to Application: Enhancing Federated Learning with Adaptive Aggregation and Gradient Protection
title_sort	From Algorithm to Application: Enhancing Federated Learning with Adaptive Aggregation and Gradient Protection
author_id_str_mv	c724e9f97d8ed76a8c0b3ea26bbc43a3
author_id_fullname_str_mv	c724e9f97d8ed76a8c0b3ea26bbc43a3_***_YI HU
author	YI HU
author2	YI HU
format	E-Thesis
publishDate	2025
institution	Swansea University
doi_str_mv	10.23889/SUThesis.71338
college_str	Faculty of Science and Engineering
hierarchytype
hierarchy_top_id	facultyofscienceandengineering
hierarchy_top_title	Faculty of Science and Engineering
hierarchy_parent_id	facultyofscienceandengineering
hierarchy_parent_title	Faculty of Science and Engineering
department_str	School of Mathematics and Computer Science - Computer Science{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Mathematics and Computer Science - Computer Science
document_store_str	1
active_str	0
description	Federated Learning (FL) has emerged as a promising paradigm for decentralised machine learning, allowing multiple clients to collaboratively train models without sharing their private data. This distributed approach is particularly relevant in domains with stringent data privacy requirements, such as ﬁnance, healthcare, and edge computing. However, despite its advantages, FL faces three critical challenges: (1) efﬁcient model aggregation, (2) protection against privacy leakage, and (3) real-world applicability in complex domains. This thesis addresses these challenges by proposing novel strategies to optimise aggregation mechanisms, enhance its privacy protection capabilities, and explore its application in ﬁnancial modelling. First, we introduce Element-Wise Weights Aggregation for FL (EWWA-FL), a novel optimisation technique that improves global model convergence by adopting element-wise adaptive weighting. Traditional FL aggregation methods, such as FedAvg, assign a single proportion to each local model without considering the varying importance of individual model parameters. In contrast, EWWA-FL assigns unique proportions to each element within local model weights, ensuring more precise updates that account for dataset heterogeneity among clients. Experimental results demonstrate that EWWA-FL signiﬁcantly improves both convergence speed and ﬁnal model accuracy, outperforming FedAvg, FedOpt, and FedCAMS across multiple benchmark datasets. By incorporating an element-wise approach, EWWA-FL provides a more adaptive and ﬁne-grained aggregation strategy that enhances FL’s performance in both Independent and Identically Distributed (IID) and Non-IID settings. Second, we propose AdaDefence, a privacy-preserving defence mechanism against gradient leakage attacks in FL. While FL eliminates the need for raw data sharing, recent attacks have demonstrated that an adversary can reconstruct private training data from shared gradients, posing a severe privacy risk. To counteract this, AdaDefence introduces a gradient stand-in approach, wherein local clients replace actual gradients with modiﬁed gradients before sending them to the server. This method prevents attackers from reconstructing private data while maintaining model utility. AdaDefence effectively defends against state-of-the-art attacks such as Deep Leakage from Gradients (DLG), Generative Regression Neural Network (GRNN), and Inverting Gradient (IG) without signiﬁcantly compromising model accuracy. Our extensive empirical analysis shows that AdaDefence provides strong privacy guarantees while ensuring minimal performance degradation, making it a practical and scalable solution for real-world FL deployments. Finally, we explore the real-world application of FL in ﬁnancial modelling, particularly in Cross-Stock Trend Integration (CSTI) for enhancing stock price prediction. Traditional ﬁnancial models suffer from data fragmentation, where different ﬁnancial institutions and stock markets operate in silos, limiting predictive power. To overcome this, we develop a FL-based approach that enables multiple ﬁnancial institutions to collaboratively train stock prediction models without exposing sensitive trading data. This approach leverages cross-stock trend integration, allowing predictive models to learn patterns from multiple stocks while preserving privacy.Our experimental results demonstrate that federated cross-stock learning improves predictive accuracy and model robustness, outperforming conventional single-stock prediction methods.By enabling secure, multi-institution collaboration, this work highlights the potential of FL in advancing ﬁnancial modelling while ensuring regulatory compliance and data conﬁdentiality. By addressing these fundamental aspects, optimisation and protection, this thesis makes signiﬁcant contributions to the ﬁeld of FL. The proposed methodologies collectively enhance FL’s efﬁciency, security, and real-world applicability. Through EWWA-FL, FL models achieve faster and more reliable convergence. By introducing AdaDefence, FL gains stronger privacy protections against gradient-based attacks. Finally, by demonstrating FL’s potential in cross-stock trend integration, this thesis showcases how FL can be deployed in privacy-sensitive ﬁnancial applications. These contributions pave the way for more efﬁcient, secure, and scalable FL systems, advancing their adoption in a wide range of domains, including healthcare, autonomous systems, and ﬁnancial technology.
published_date	2025-11-07T05:33:52Z
_version_	1856805832315895808
score	11.09611

From Algorithm to Application: Enhancing Federated Learning with Adaptive Aggregation and Gradient Protection / YI HU

Similar Items