Dr Kyle Martin k.martin3@rgu.ac.uk
Lecturer
Dr Kyle Martin k.martin3@rgu.ac.uk
Lecturer
Professor Nirmalie Wiratunga n.wiratunga@rgu.ac.uk
Supervisor
Understanding similarity between different examples is a crucial aspect of Case-Based Reasoning (CBR) systems, but learning representations optimised for similarity comparisons can be difficult. CBR systems typically rely on separate algorithms to learn representations for cases and to compare those representations, as symbolised by the vocabulary and similarity knowledge containers respectively. Deep Metric Learners (DMLs) are a branch of deep learning architectures which learn a representation optimised for similarity comparison by leveraging direct case comparisons during training. In this thesis we explore the symbiotic relationship between these two fields of research. Firstly we examine what can be learned from traditional CBR research to improve the training of DMLs through training strategies. We then examine how DMLs can fill the traditionally separate roles of the vocabulary and similarity knowledge containers. We perform this exploration on the real-world problem of experience transfer between experts and non-experts on service provisioning for telecommunication organisations. This problem is also revealing about the requirements for practical applications to be explainable to their intended user group. With that in mind, we conclude this thesis with work towards the development of an explanation framework designed to explain the recommendations of similarity-based classifiers. We support this practical contribution with an exploration of similarity knowledge to support autonomous measurement of explanation quality.
MARTIN, K. 2021. Similarity and explanation for dynamic telecommunication engineer support. Robert Gordon University, PhD thesis. Hosted on OpenAIR [online]. Available from: https://doi.org/10.48526/rgu-wt-1447160
Thesis Type | Thesis |
---|---|
Deposit Date | Sep 7, 2021 |
Publicly Available Date | Sep 7, 2021 |
Keywords | Case-based reasoning (CBR); Deep metric learners (DMLs); Machine learning; Semantic computing; Telecommunications |
Public URL | https://rgu-repository.worktribe.com/output/1447160 |
External URL | https://doi.org/10.48526/rgu-wt-1447160 |
Award Date | May 31, 2021 |
MARTIN 2021 Similarity and explanation
(8.3 Mb)
PDF
Licence
https://creativecommons.org/licenses/by-nc/4.0/
Copyright Statement
© The Author.
Evaluating explainability methods intended for multiple stakeholders.
(2021)
Journal Article
Assessing the clinicians’ pathway to embed artificial intelligence for assisted diagnostics of fracture detection.
(2020)
Presentation / Conference Contribution
Escaping traditional outreach: digital escape rooms to engage potential students.
(2022)
Presentation / Conference Contribution
Digital interpretation of sensor-equipment diagrams.
(2018)
Presentation / Conference Contribution
About OpenAIR@RGU
Administrator e-mail: publications@rgu.ac.uk
This application uses the following open-source libraries:
Apache License Version 2.0 (http://www.apache.org/licenses/)
Apache License Version 2.0 (http://www.apache.org/licenses/)
SIL OFL 1.1 (http://scripts.sil.org/OFL)
MIT License (http://opensource.org/licenses/mit-license.html)
CC BY 3.0 ( http://creativecommons.org/licenses/by/3.0/)
Powered by Worktribe © 2025
Advanced Search