Distributed machine learning on Edge computing systems - PhD thesis

Aug 5, 2026

Upload by Aria

Distributed machine learning (DML) systems that split training workloads across multiple nodes have grown rapidly over the past decade due to increasing dataset scale and computational demands. Training at the edge has also become attractive for privacy protection and for reducing communication costs tied to sending raw data to the cloud. These drivers motivate a DML-at-edge paradigm, while practical deployment must address limited, heterogeneous hardware, high communication overhead, and constraints that prevent effective on-device training on low-end devices.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages213

Reading Time8 h 57 min

AccessFree to download

Candidate's declaration

Supervisor's declaration

Permission for publication

Underpinning Research Data or Digital Outputs

Acknowledgements

Funding

Abstract

Distributed machine learning on Edge computing systems Di Wu A thesis submitted for the degree of PhD at the University of St Andrews 2024 Full metadata for this thesis is available in St Andrews Research Repository at: [https://research-repository.st-andrews.ac.uk/](https://research-repository.st-andrews.ac.uk/) [Identifier to use to cite or link to this thesis:](Identifier to use to cite or link to this thesis:)[ ](Identifier to use to cite or link to this thesis:)DOI: [https://doi.org/10.17630/sta/1164](https://doi.org/10.17630/sta/1164) This item is protected by original copyright This item is licensed under a Creative Commons Licence [https://creativecommons.org/licenses/by-nc/4.0/](https://creativecommons.org/licenses/by-nc/4.0/) Candidate's declaration I, Di Wu, do hereby certify that this thesis, submitted for the degree of PhD, which is approximately 55,000 words in length, has been written by me, and that it is the record of work carried out by me, or principally by myself in collaboration with others as acknowledged, and that it has not been submitted in any previous application for any degree. I confirm that any appendices included in my thesis contain only material permitted by the 'Assessment of Postgraduate Research Students' policy. I was admitted as a research student at the University of St Andrews in January 2022. I received funding from an organisation or institution and have acknowledged the funder(s) in the full text of my thesis. Date Signature of candidate Supervisor's declaration I hereby certify that the candidate has fulfilled the conditions of the Resolution and Regulations appropriate for the degree of PhD in the University of St Andrews and that the candidate is qualified to submit this thesis in application for that degree. I confirm that any appendices included in the thesis contain only material permitted by the 'Assessment of Postgraduate Research Students' policy. Date Signature of supervisor Permission for publication In submitting this thesis to the University of St Andrews we understand that we are giving permission for it to be made available for use in accordance with the regulations of the University Library for the time being in force, subject to any copyright vested in the work not being affected thereby. We also understand, unless exempt by an award of an embargo as requested below, that the title and the abstract will be published, and that a copy of the work may be made and supplied to any bona fide library or research worker, that this thesis will be electronically accessible for personal or research use and that the library has the right to migrate this thesis into new electronic forms as required to ensure continued access to the thesis. I, Di Wu, confirm that my thesis does not contain any third-party material that requires copyright clearance. The following is an agreed request by candidate and supervisor regarding the publication of this thesis: Printed copy No embargo on print copy. Electronic copy No embargo on electronic copy. Date Signature of candidate Date Signature of supervisor Underpinning Research Data or Digital Outputs Candidate's declaration I, Di Wu, hereby certify that no requirements to deposit original research data or digital outputs apply to this thesis and that, where appropriate, secondary data used have been referenced in the full text of my thesis. Date Signature of candidate Acknowledgements General acknowledgements First and foremost, I would like to express my deepest gratitude to my supervisor, Dr Blesson Varghese. His invaluable guidance, support, and encouragement have been instrumental in completing this research. I still vividly remember how, after I ﬁnished the drafts of my research articles, Dr Varghese provided comprehensive feedback and suggestions for revisions, signiﬁcantly improving the quality of the work. I am incredibly grateful for his mentorship throug

FAQs

What problem does the thesis address?

The thesis addresses how to run distributed machine learning systems at the edge, motivated by large-scale training demand, data privacy needs, and communication cost reduction.

Why is DML training at the edge challenging?

The thesis highlights four challenges, including limited and heterogeneous edge hardware leading to impractical training times, substantial communication costs, and limitations that make on-device training difficult on low-end devices.

What funding supported the research?

The work was funded by Rakuten Mobile Inc., Japan, under the project titled “CASE: Context-aware, Adaptive and Scalable Edge Management”.

Aug 5, 2026

Upload by Aria