PhD Student Projects
The following are a set of projects that I am interested in pursuing with students as PhD students. If you have other ideas regarding proposals in these or closely related areas, then by all means drop me an email and we can discuss ways to move your proposal forward.
- Uncertainty in Digital Games
- Situational Awareness in Autonomous Vehicles
- Sensemaking in Information Seeking
- Problem Solving in Data Analytics Applications
- Influence of User Behaviour on Image Segmentation Applications
- Positive Impacts of Gaming on Well-being
- Mutation Testing of Automated Accessibility Tools
Uncertainty in Interactive Systems
Players of digital games are able to identify and describe uncertainty in very specific terms, and it has an influence on their overall experience in a game. When uncertainty is too high, users become frustrated and disengage, and when it is too low they become bored. This tends to indicate a complex relationship between uncertainty and the types of goal experiences of immersion and fun that game designers are seeking.
As part of their work, a PhD student undertaking this project would work with my Player Uncertainty in Games Scale (PUGS) to identify the relationships between uncertainty and many of these goal experiences, and relate them to specific design patterns in games.
As autonomous vehicles move towards the mainstream, there are increasing concerns about their safety, in particular when there are handoffs required between a human operator and the system. The ability for the human to gain awareness of the overall state of the system, problem solve and take control the system quickly and correctly could mean the difference between having an accident or not.
A student interested in this problem could investigate the link between my work on uncertainty on digital games and the concept of situational awareness, given that both relate to problem solving and decision making in complex systems. This work could be undertaken in a number of different contexts, likely using digital games or virtual reality platforms as tools for simulating environments to investigate operator uncertainty.
Information seeking in large archives of documents, images and other digital artefacts is challenging, with users needing to formulate their goals, translate them into queries and then navigate a myriad of heterogeneous data and metadata, and that is before they even start to process data. This process is riddled with uncertainty, both from the system and from the problem they are trying to solve.
In work with a previous student, I have developed a psychometric scale for capturing the uncertainty that people have in archives and the disorientation that ensues. A student undertaking studies with this scale could focus on understanding user experience in archives, linking uncertainty to higher level concepts like engagement. Alternatively, could focus on engineering good information seeking systems, looking at what design patterns that could improve user outcomes by reducing uncertainty. This work could build on my links with the National Archives or the Archaeological Data Service, or could occur in a new domain.
Users in data analytics applications such as financial planning or forecasting need to maintain mental models of complex, multidimensional relationships in their data in order to make decisions for the future. One of the challenges in this space is that the interfaces are often complex, requiring pivoting on different views of the data in order to make decisions, or choosing the correct visualisation in order to gain insight. This can result in users becoming disoriented and lost in their data, unable to make forward progress.
A student undertaking this as a PhD would transfer some of the work I have done in other domains, such as information seeking and digital games, into the data analytics space. This could focus on a specific domain like financial planning, leveraging my links with IBM Data Analytics, or it could be in a domain of the students’ choosing.
Image segmentation algorithms provide a means of separating foreground images from background images. Segmentation is used in a number of different domains, but one key application is in health where they can be used to identify tumour sites. This of course requires a high level of accuracy and reproducibility across a range of images. One way to improve these algorithms is to seed the algorithms with input from users, who mark up images to indicate foreground and background.
Research in this area often discounts the variability of human behaviour as an input to the system. Evaluations with algorithms are low on detail in terms of instructions given to participants, the types of strokes needed and even in the image sets that were marked up. Recent work by myself and my colleague Dr. Mark Eramian at the University of Saskatchewan over a series of experiments has shown that these factors do have a substantial effect on the outcomes, and on user certainty as to whether they have done the task correctly. A student undertaking this project will follow up on this work, and begin to explore the full range of variables that impact these algorithms. Dr. Eramian will serve as an external co-supervisor to this work.
A lot of the research in the sphere of digital games is focussed on the negative impacts that games have on people. Isolation, addiction and other negative experiences get a lot of focus in both research and news cycles. However, we know that connections between people can have massive impacts on well-being outcomes, and on the life expectancy of individuals. There are strong arguments that can be made for using digital gaming to improve health and well-being over time.
A student undertaking this work will investigate different gaming experiences and how they might impact gamers in diverse populations. Students could work with people with disabilities, older adults or people who struggle with mental health issues, to understand how gaming plays a role in maintaining or improving their mental health.
This work builds on existing work with MSc. and PhD students that I have supervised in cooperation with Dr. Paul Cairns, who will serve as co-supervisor on projects in this area.
Automated testing remains one of the most relied on means of testing accessibility of websites. While it will never cover all of the problems encountered by people with disabilities, it is a key means in ensuring that web code has the qualities needed to support the assistive technologies used by people with disabilities, giving more options to access the web. However, there are indications from recent studies that these tools are not necessarily bug free, and that reports provided by these tools may be faulty, leading to false reporting by companies about the current accessibility of their websites.
A student undertaking this project will investigate different types of mutation testing for detecting bugs in accessibility testing software. The outcomes of this project could be oriented around a comparison of algorithms for testing these types of tools, or could be related to the common problems in these testing tools.
This work builds on existing work with MSc. students that I have supervised with Dr. Robert Alexander who will serve as co-supervisor on projects in this area.