Cinca, Robert;
(2025)
Understanding the use of ML
Technologies: Initial Approaches and
Analyzing Biases.
Doctoral thesis (Ph.D), UCL (University College London).
Preview |
Text
Cinca_Thesis.pdf - Accepted Version Download (4MB) | Preview |
Abstract
The growing application of machine learning (ML) across various domains, such as healthcare and crime prevention, presents significant implementation challenges. Misusing ML technologies can lead to harmful consequences, including biased out comes that unfairly impact certain groups. As ML algorithms and platforms become more accessible, users with limited knowledge of their theoretical foundations and implementation may struggle to apply them effectively. This thesis investigates how novice users engage with ML and how more experienced users navigate and operationalize bias. Addressing these challenges and mitigating biases in ML remains a complex issue. Understanding how ML technologies are used is crucial, as it informs the design of ML tools that support users in applying ML effectively—potentially leading to better models with fairer outcomes. The studies in this thesis address three challenges commonly presented in users’ interaction with ML technologies: 1. The challenges of novice users when initially approaching ML problems; 2. The reflections of ML practitioners when operationalizing bias definitions; 3. Designing an interactive bias detection and mitigation tool. These challenges were approached through the design of a bias tool and three qualitative user studies. The findings showed that both novices and ML practitioners faced difficulties with problem selection and multi-dimensionality, yet they demon strated awareness of good model building practices and engaged in discussions about bias. Drawing from these needs, an interactive ML tool named XPlain was devised to provide ML users with a practical way of automating bias detection and mitiga tion directly in their code. This tool was validated both empirically and through an analytical evaluation, comparing the results to previous research on two popular datasets employed by the fairness community, COMPAS and German Credit.
| Type: | Thesis (Doctoral) |
|---|---|
| Qualification: | Ph.D |
| Title: | Understanding the use of ML Technologies: Initial Approaches and Analyzing Biases |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Copyright © The Author 2025. Original content in this thesis is licensed under the terms of the Creative Commons Attribution 4.0 International (CC BY 4.0) Licence (https://creativecommons.org/licenses/by/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request. |
| UCL classification: | UCL > Provost and Vice Provost Offices > UCL BEAMS UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science > Dept of Computer Science UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Brain Sciences > Div of Psychology and Lang Sciences > UCL Interaction Centre UCL |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10208075 |
Archive Staff Only
 |
View Item |
