Publications

Artificial Intelligence (AI) in Education has been said to have the potential for building more personalised curricula, as well as democratising education worldwide and creating a Renaissance of new ways of teaching and learning. Millions of students are already starting to benefit from the use of these technologies, but millions more around the world are not. If this trend continues, the first delivery of AI in Education could be greater educational inequality, along with a global misallocation of educational resources motivated by the current technological determinism narrative. In this paper, we focus on speculating and posing questions around the future of AI in Education, with the aim of starting the pressing conversation that would set the right foundations for the new generation of education that is permeated by technology. This paper starts by synthesising how AI might change how we learn and teach, focusing specifically on the case of personalised learning companions, and then move to discuss some socio-technical features that will be crucial for avoiding the perils of these AI systems worldwide (and perhaps ensuring their success). This paper also discusses the potential of using AI together with free, participatory and democratic resources, such as Wikipedia, Open Educational Resources and open-source tools. We also emphasise the need for collectively designing human-centered, transparent, interactive and collaborative AI-based algorithms that empower and give complete agency to stakeholders, as well as support new emerging pedagogies. Finally, we ask what would it take for this educational revolution to provide egalitarian and empowering access to education, beyond any political, cultural, language, geographical and learning ability barriers.

Extracting useful information from the user history to clearly understand informational needs is a crucial feature of a proactive information retrieval system. Regarding understanding information and relevance, Wikipedia can provide the background knowledge that an intelligent system needs. This work explores how exploiting the context of a query using Wikipedia concepts can improve proactive information retrieval on noisy text. We formulate two models that use entity linking to associate Wikipedia topics with the relevance model. Our experiments around a podcast segment retrieval task demonstrate that there is a clear signal of relevance in Wikipedia concepts while a ranking model can improve precision by incorporating them. We also find Wikifying the background context of a query can help disambiguate the meaning of the query, further helping proactive information retrieval.

For many years, the biggest issue that causes discomfort and hygiene issues for patients with lower limb amputations have been the interface between body and prosthetic, the socket. Often made of an inflexible, solid polymer that does not allow the residual limb to breathe or perspire and with no consideration for the changes in size and shape of the human body caused by changes in temperature or environment, inflammation, irritation and discomfort often cause reduced usage or outright rejection of the prosthetic by the patient in their day to day lives. To address these issues and move towards a future of improved quality of life for patients who suffer amputations, Loughborough University formed the Next Generation Prosthetics research cluster. This work is one of four multidisciplinary research studies conducted by members of this research cluster, focusing on the area of Computer Aided Design (CAD) for improving the interface with Additive Manufacture (AM) to solve some of the challenges presented with improving prosthetic socket design, with an aim to improve and streamline the process to enable the involvement of clinicians and patients in the design process. The research presented in this thesis is based on three primary studies. The first study involved the conception of a CAD criteria, deciding what features are needed to represent the various properties the future socket outlined by the research cluster needs. These criteria were then used for testing three CAD systems, one each from the Parametric, Non Uniform Rational Basis Spline (NURBS) and Polygon archetypes respectively. The result of these tests led to the creation of a hybrid control workflow, used as the basis for finding improvements. The second study explored emerging CAD solutions, various new systems or plug-ins that had opportunities to improve the control model. These solutions were tested individually in areas where they could improve the workflow, and the successful solutions were added to the hybrid workflow to improve and reduce the workflow further. The final study involved taking the knowledge gained from the literature and the first two studies in order to theorise how an ideal CAD system for producing future prosthetic sockets would work, with considerations for user interface issues as well as background CAD applications. The third study was then used to inform the final deliverable of this research, a software design specification that defines how the system would work. This specification was written as a challenge to the CAD community, hoping to inform and aid future advancements in CAD software. As a final stage of research validation, a number of members of the CAD community were contacted and interviewed about their feelings of the work produced and their feedback was taken in order to inform future research in this area.

Bodkin, Troy L. (2017): Specifying a hybrid, multiple material CAD system for next-generation prosthetic design. Loughborough University. Thesis. https://hdl.handle.net/2134/25...;

Transferring independently to and from their wheelchair is an essential routine task for many wheelchair users but it can be physically demanding and can lead to falls and upper limb injuries that reduce the person’s independence. New assistive technologies (ATs) that facilitate the performance of wheelchair transfers have the potential to allow wheelchair users to gain further independence. To ensure that users’ needs are addressed by ATs, the active involvement of wheelchair users in the process of design and development is critical. However, participation can be burdensome for many wheelchair users as design processes where users are directly involved often require prolonged engagement. This thesis makes two contributions to facilitate wheelchair users’ engagement in the participatory design process for ATs, while being mindful of the burden of participation. The first contribution is a framework that provides a modular structure guiding the participatory design process from initial problem identification and analysis to facilitating collaborations between wheelchair users and designers. The framework identifies four factors determining the need and adoption process for ATs: (i) People focuses on the target population, (ii) Person includes personal characteristics, (iii) Activity refers to the challenges associated with the task, and (iv) Context encompasses the effect of the environment in which the activity takes place. The second contribution constitutes a rich picture of personal and external elements influencing real world wheelchair transfers that emerged from four studies carried out to investigate the effect of the framework factors on the design process for ATs. A related outcome based on these contributions is a framing document to share knowledge between wheelchair users and designers to provide focus and promote an equal collaboration among participants.

Barbareschi, Giulia. “YouTransfer, YouDesign : a Participatory Approach to Design Assistive Technology for Wheelchair Transfers / Giulia Barbareschi.” Thesis (Ph.D.)--University College London, 2018., 2018. Print.

Wrist splints are a common treatment for rheumatoid arthritis, however their effectiveness is compromised by patients not wearing splints as often as prescribed. Previous research has identified a number of reasons for non-compliance, but typically lacks insights that could lead to improved splint design.

This thesis investigates the motivators for patients to wear and not wear their wrist splints and, the impact of personalisation of splint appearance on patient wear. The work is based on the premise that digital design and manufacturing processes, such as Computer-Aided Design (CAD) and 3D Printing, can produce bespoke splints on demand.

The research begins with a literature review across the core areas of: splinting, additive manufacture, product appearance and personalisation. This literature review identifies gaps in knowledge from which research questions are established for the work.

The research employs a qualitative, generative design research approach and, follows a codesign framework employing telling, making and enacting tools. The thesis is made up of three studies. The first study is a sensitisation study and uses design probes to prepare the participants for the research and begin exploring the problem space. The second is a comprehensive study into participants splint wear behaviour and uses context mapping and scenario picture card tools to investigate the motivators for participants to wear and not wear wrist splints, along with positive and negative outcomes or wearing/not wearing splints. The final study uses a personalisation toolkit to elicit patient needs for a future wrist splint design and investigate self-reported expectations regarding compliance of patients who used the toolkit.

The research finds that patient compliance is affected by practical and aesthetic limitations of current splints. It identifies 4 motivating factors to wear a splint and 10 motivating factors to not wear a splint. Additionally, it identifies 6 positive outcomes of wearing splints, 6 negative outcomes of wearing splints, 3 positive outcomes of not wearing splints and 3 negative outcomes of not wearing splints. Requirements for an improved splint design are established and form the basis of the design for a prototype personalisation toolkit. Testing of this toolkit reveals that patients are keen to own more than one splint and personalise splints to match the scenario in which it is to be worn. Patients reported that they expected to be more compliant with a personalised splint when compared to their current splint.

Pyatt, Charlotte (2018): Understanding wrist splint user needs and personalisation through codesign. Loughborough University. Thesis. https://doi.org/10.26174/thesi...;

Independent mobility is fundamental to the quality of life of people with impairment. Most people with severe mobility impairments, whether congenital, e.g., from cerebral palsy, or acquired, e.g., from spinal cord injury, are prescribed a wheelchair. A small yet significant number of people are unable to use a typical powered wheelchair controlled with a joystick. Instead, some of these people require alternative interfaces such as a head- array or Sip/Puff switch to drive their powered wheelchairs. However, these alternative interfaces do not work for everyone and often cause frustration, fatigue and collisions. This thesis develops a novel technique to help improve the usability of some of these alternative interfaces, in particular, the head-array and Sip/Puff switch. Control is shared between a powered wheelchair user, using an alternative interface and a pow- ered wheelchair fitted with sensors. This shared control then produces a resulting motion that is close to what the user desires to do but a motion that is also safe. A path planning algorithm on the wheelchair is implemented using techniques in mo- bile robotics. Afterwards, the output of the path planning algorithm and the user’s com- mand are both modelled as random variables. These random variables are then blended in a joint probability distribution where the final velocity to the wheelchair is the one that maximises the joint probability distribution. The performance of the probabilistic approach to blending the user’s inputs with the output of a path planner, is benchmarked against the most common form of shared control called linear blending. The benchmarking consists of several experiments with end users both in a simulated world and in the real-world. The thesis concludes that probabilistic shared control provides safer motion compared with the traditional shared control for difficult tasks and hard-to-use interfaces.

Ezeh, Chinemelu Ejiamatu Muoma. Shared Control for Wheelchair Interfaces. UCL (University College London), 2018. Print.