Academic Posters Collection

Follow

• 

  How Can Personalised Feedback in Assignments Help Address Gender Balance in Computing Education?

  Alina Berry

  Personalised feedback is frequently used in computing assessments in higher education. Research has shown that personalised feedback positively influences persistence in computer science. Computing and related disciplines are known to show relatively low retention rates. This includes female students, who are strongly underrepresented in computing disciplines, so they can be considered as a particularly important group for retention-driven initiatives. Female science students are more likely to act upon feedback, and personalised feedback has increased intentions to persist among female top performing students in computing. Hence, providing personalised feedback can be considered as a promising gender initiative that has a potential to enhance gender balance in computing education. It is one of the initiatives (actions) of TechMate, the toolkit developed at TU Dublin to help address the issue of gender imbalance on third level. This poster presents a study that has an aim to get an understanding on how personalised feedback is provided in technical modules at TU Dublin, what are students’ opinions about it, and how it possibly affects female students. The study comprises of a series of interviews with lecturers, followed by surveying students of all genders and interviewing some of the female students in relevant modules.

• 

  Personalised Feedback on Assessments in Computing Modules - Gender Equality Action in Context

  Alina Berry

  Personalised feedback in computing higher education has been known to positively influence retention of women. The issue of gender inequality in computing field is well known and one of the efforts to address it is the development of a gender equality toolkit (TechMate), with one initiative (action) in the toolkit being personalised feedback. While all actions in the toolkit are research-driven, the aim of this work was to evaluate the action on personalised feedback in a local context.

  The study comprised of 10 semi-structured interviews with computing lecturers at TU Dublin who provide personalised feedback, a student survey with 68 responses and 4 female student interviews. Results (in progress) reveal that there are gender differences in ways personalised feedback is perceived and how it benefits students from different gender groups, as well as student recommendations have been collected on enhancement of existing feedback.

• 

  Feedback, Learning Outcomes and Mathematics Anxiety in a Digital Game Based Learning Approach in Mathematics Education

  André Almo

  Feedback is a crucial part of learning, and an essential element in digital game-based learning approaches, in which digital games - known as 'serious games' - are used to deliver educational content. Feedback features respond to players' actions within the game, providing them with information and guidance, as well as potentially impacting their learning, motivation and engagement. However, these features may be designed differently, since they include various distinct characteristics and dimensions. This work proposes a new taxonomy for feedback features in serious games, with an emphasis in game design aspects, in order to provide clearer descriptions and distinctions of different feedback characteristics.

• 

  Evaluation of Gender-Based Differences in Primary School Maths Education: the Potential of Digital Games

  Maíra Amaral

  Digital Game-Based Learning is shown to be a more effective instructional method than traditional instruction, however less effective than other technology-supported instruction according to Byun and Joung (2018). Regarding gender aspects, according to findings by Mclaren and colleagues in 2022, girls may learn more mathematics from digital learning games than boys. In their study, even reporting greater behavioural and cognitive engagement, boys did not learn more with the game than girls.

• 

  Co-design of an Interactive Wellness Park: Exploring Design Requirements for a Multimodal Outdoor Physical Web Installation with Older Adults

  Fatima Badmos

  The global demographic landscape is experiencing a notable shift, characterised by a growing proportion of adults over 60. According to projections, the proportion of individuals aged 60 and above is expected to reach one-sixth of the global population by 2030. Furthermore, by 2050, this demographic is projected to exceed a staggering two billion people. Amidst this shift, there is an urgent need to develop interactive and innovative solutions to address older adults' unique challenges, particularly in outdoor physical activity.

  A co-design methodology involving older adults’ participation from the idea generation to the application development process will be adopted to address these challenges. For the initial study, twenty-four older adults were recruited to participate in four focus groups to explore their perceptions and experiences about their physical activity routines. The focus groups’ workshop data will be used to ideate initial design ideas for the proposed physical web interface.

  The outcomes seek to harness the synergy of web technology and physical installations to create an engaging and accessible outdoor environment for older adults. This study aspires to offer a scalable and replicable model for other communities, contributing to the broader academic and social discourse on age-friendly urban design and technology-enabled active ageing.

• 

  TechMate: A Research-Driven Toolkit to Enhance Gender Balance in Computing Education

  Alina Berry and Sarah Jane Delany

  This poster presents a toolkit of practical initiatives and guidance on how to enhance gender balance in computing higher education. The suggested initiatives are designed in the way that could be adapted for a use in a local context, especially in universities in the UK or in Ireland. The initiatives are categorised under four main areas: Policy, Pedagogy, Influence & Support and Promotion & Engagement. Additionally, guidance is given on mechanisms to evaluate the impact of these initiatives. This work will be of interest to champions looking to enhance gender balance in their computing courses.

• 

  The Proof Is In The Pudding – Using Perceived Stress To Measure Short-Term Impact in Initiatives to Enhance Gender Balance in Computing Education

  Alina Berry and Sarah Jane Delany

  The problem of gender imbalance in computing higher education has forced academics and professionals to implement a wide range of initiatives. Many initiatives use recruitment or retention numbers as their most obvious evidence of impact. This type of evidence of impact is, however, more resource heavy to obtain, as well as often requires a longitudinal approach. There are many shorter term initiatives that use other ways to measure their success.

  First, this poster presents with a review of existing evaluation measures in interventions to recruit and retain women in computing education across the board. Three main groups of evaluation come out of this review: statistical data, feedback and instruments. Second, this work reveals what type of evaluation is typically present in what types of initiatives. Finally, it recommends Perceived Stress Scale instrument with data collected in a retrospective pre- and then post survey as a lightweight evaluation method for short-term impact.

  This research aims to assist creators of initiatives in demonstrating quick wins of their efforts to enhance gender balance in STEM disciplines.

• 

  Application of the Two-Variable Model to Simulate a Multisensory Reaction-Time Task

  Rebecca Brady and John Butler

  To navigate the world in an efficient manner, the brain seamlessly integrates signals received across multiple sensory modalities. Behavioral studies have suggested that multisensory processing is a winner-take-all sensory response mechanism to some optimal combination of sensory signals. In addition, multiple sensory cues are not always beneficial with some studies showing maladaptive multisensory processing as an identifier of older adults prone to falls from age matched healthy controls.

  A stalwart of modelling sensory decision-making is the work by (Wong &Wang, 2006) but to date almost all of this research has been focused on unisensory tasks. We extend the reduced two-variable model developed to model an audio-visual speeded reaction-time task. This model consists of a system of ordinary differential equations motivated by biological data to simulate both unisensory and multisensory processing and behavioral responses.

  The model includes noise to account for trial-to-trial variability and by sampling from a set of model parameters we simulated different ‘participants’ sensory processing. With these straightforward manipulations the model replicates data from multisensory behavioral studies. The results show how biologically motivated mathematical models can be used to probe the potential mechanisms responsible for the observed differences in unisensory and multisensory responses.

• 

  Framework for Trustworthy AI in the Health Sector

  Mykhailo Danilevskyi

  The European Commission defines that Trustworthy AI should be lawful, ethical and robust. The ethical component and its technical methods are the main focus of the research. According to this, the initial research goal is to create a methodology for evaluating datasets for ML modeling using ethical principles in the healthcare domain. Ethical risk assessment will help to ensure compliance with principles such as privacy, fairness, safety and transparency which are especially important for the Health Care sector. At the same time, risks must be evaluated with respect to AI model performance and possible scenarios of risk mitigation. Ethical risk mitigation techniques involve data modification, elimination of private information from datasets that directly influence AI modelling. Therefore ethical risk mitigation techniques should be carefully selected depending on domain and context. In this research work, we present an analysis of these techniques.

• 

  Detecting Road Intersections from Satellite Images using Convolutinal Neural Networks

  Fatmaelzahraa Eltaher, Luis Miralles-Pechuán, Jane Courtney, and Susan McKeever

  The location of intersections is an important consideration for vulnerable road users such as People with Blindness or Visually Impairment (PBVI) or children. Route planning applications, however, do not give information about the location of intersections as this information is not available at scale. In this paper, we propose a deep learning framework to automatically detect the location of intersections from satellite images using convolutional neural networks. For this purpose, we labelled 7,342 Google maps images from Washington, DC, USA to create a dataset. This dataset covers a region of 58.98 km$^{2}$ and has 7,548 intersections. We then applied a recent object detection model (EfficientDet) to detect the location of intersections. Experiments based on the road network in Washington, DC, show that the accuracy of our model is within 5 meters for 88.6\% of the predicted intersections. Most of our predicted centres of the intersections (approx 80\%) are within 2 metres of the ground truth centre. Using hybrid images, we obtained an average recall and an average precision of 76.5\% and 82.8\% respectively, computed for values of Intersection Over Union (IOU) from 0.5 to 0.95, step 0.05. We have published an automation script to enable the reproduction of our dataset for other researchers.

• 

  Enhancing Health Data Representation for Older Adults: Unlocking Opportunities

  Peterson Jean

  The prevalence of off-the-shelf wearable devices increases the monitoring and measurement of critical physiological parameters like activity, sleep, heart rate, and blood pressure. However, the accessibility of health data representations poses challenges for older adults, who often struggle to understand the criticality of their own health data without assistance. This poster highlights the challenges older adults face in accessing their health data from wearable technologies, specifically focusing on data representations.

  To address these challenges, it proposes a methodology that involves a heuristic evaluation of existing data representations with experts and accessibility studies with older adults using a mixed methods approach and objective measures. The heuristic evaluation aims to gain deeper insights into the usability of health data representations and identify underlying accessibility challenges to create a more tailored group of tasks and potential accessibility scenarios to be evaluated in health data representations by older adults.

  The outcomes seek to improve accessibility by developing a multimodal prototype with visual, auditory, and tactile feedback and promoting better health conditions and support for older adults through personalized and inclusive health data representation.

• 

  A Framework for Confusion Mitigation in Task-Oriented Interactions

  Na Li and Robert J. Ross

  Confusion is a mental state that can be triggered in task-oriented interactions and which can if left unattended lead to boredom, frustration, or disengagement from the task at hand. Previous work has demonstrated that confusion can be detected in situated human-robot interactions from visual and auditory cues. Therefore, in the next step, we propose appropriate interaction structures in this study, which should be used to mitigate confusion. We motivate and describe this dialogue mechanism through an information state-style dialogue framework and policies, and also outline the approach we are taking to integrate such a meta-conversational goal alongside core task-oriented considerations in modern data-driven conversational techniques.

• 

  Spoken Language to Irish Sign Language Machine Translation: A Linguistically Informed Approach

  Jesus Aguilar Lopez, Irene Murtagh, and Sheila Castilho

  Many deaf and hard-of-hearing individuals rely on sign language (SL) on a daily basis as a preferred language. Although nowadays there are significant advances on spoken language research, current approaches are often neither linguistically motivated nor tailored to the unique features of SLs. Further research and development are necessary to enhance Sign Language Machine Translation (SLMT) and bring it to a similar level as spoken language MT. This research will endeavour to improve the accuracy and efficiency of SLMT systems, making them more accessible to the Deaf community and empowering deaf and hard of hearing individuals to communicate more effectively with the rest of the world.

• 

  From Overlay to Interplay - Subverting the Message and Creating the Surreal with Augmented Reality

  Nina Lyons

  This proof of concept utilises content creation tools that create diegetic presentation which is not commonly utilised in AR systems. In this study, the virtual overlay that becomes visible through AR displays diegetic content that disrupts the meaning of the poster highlighting the potential for AR as a visual communication medium and the opportunities that AR has, as a medium, for creating narrative.

• 

  They Followed the Cheese - Visual Web-Based Experiments Showcasing the Opportunities Available for the Presentation of Information in AR

  Nina Lyons

  The desktop paradigm bridged a sizeable usability gap at a time before the term “usability” became a conscious objective goal of developers. Mapping the capabilities of the computer to how a user understands their work desk, helped users understand how to interact with a computer. The paradigm has evolved and continues to evolve and reshape itself in different platforms. However, the UIs still rely heavily on established laws and principles that suit the conditions of screens. AR has the opportunity to leverage Semiotic Extension to create context relevant UIs that rely on the focus of the real-time view to inform the user, rather than the device. This is a departure from established design patterns that may cause frustration for both the designers and the users. Nevertheless, it is an area that needs experimentation, to find levels of communication that work for AR. This poster presents the findings of a web-based visual experiment that posed the hypothesis: Can simple visual cues change the user’s interaction behaviour with a system? If so, is there a threshold between the communication value of the cues, where no behavioural change can be identified.

• 

  Classifying Recaptured Identity Documents Using the Biomedical Meijering and Sato Algorithms

  John Magee, Stephen Sheridan, and Christina Thorpe

  Recaptured identity documents are a low-cost, high-risk threat to modern eKYC systems. Bad actors can easily manipulate images and print them. Existing solutions typically demand manual review of remotely captured identity documents, this is expensive and does not scale. In 2022, the UK National Crime Agency estimated fraud cost business hundreds of billion pounds per year and document forgery is an area of investigation by Europol.

• 

  Using Machine Learning for Web Accessibility

  Tlamelo Makati

  This research will explore the potential of machine learning to enhance web accessibility. Web accessibility is typically defined in terms of Web Accessibility Guidelines (WCAG), which states that everyone should be able to perceive, operate, understand and interpret the web regardless of disability or use of assistive technology. We would like to consult digital accessibility experts through interviews and focus groups to understand the web accessibility auditing and remediation processes in detail, with a focus on web navigation. An important goal of this work is to establish development processes where all stakeholders can leverage machine-learning tools to produce more accessible websites.

• 

  Improve Engagement with Full Labs and Motivated Students: Interactive labs via low stakes assessment

  Susan McKeever, Patricia O'Byrne, and Amanda O'Farrell

  Poor engagement and attendance is an endemic problem at third level, particularly post covid. Our approach shows how the use of regular in-lab assessment and challenges can dramatically increase student participation and learning. Using three case studies, we demonstrate how we have successfully used this low-stakes assessment approach to improve student outcomes, across a range of modules.

• 

  Explaining Deep Learning Time Series Classification Models using a Decision Tree

  Ephrem T. Mekonnen, Pierpaolo Dondio, and Luca Longo

  This preliminary study proposes a new post hoc method to explain deep learning-based time series classification models using a decision tree. Our approach generates a decision tree graph or rulesets as an explanation, improving interpretability compared to saliency map-based methods. The method involves two phases: training and evaluating the deep learning-based time series classification model and extracting prototypical events from the evaluation set to train the decision tree classifier. We conducted experiments on artificial and real datasets, evaluating the explanations based on accuracy, fidelity, number of nodes, and depth. Our preliminary findings suggest that our post-hoc method improves the interpretability and trust of complex time series classification models.

• 

  Towards Accommodating Gerunds within the Sign Language Lexicon

  Zaid Mohammed and Irene Murtagh

  This work is part of ongoing research work that focuses on the linguistic analysis and computational description of five different Sign Languages (SLs) namely Irish Sign Language (ISL), Flemish Sign Language (VGT), Dutch Sign Language (NGT), Spanish Sign Language (LSE), and British Sign Language (BSL) as part of the SignON project. This work will be leveraged to inform the development of SL lexicon entries for a Sign Language Machine Translation (SLMT) system. In particular, this research focuses on ISL. We investigate the existence of constructions similar to or equivalent in functionality to gerunds in spoken language, in particular, English. The initial findings indicate that such constructions do indeed exist and that they can take many forms.

• 

  Attention-based Gender-Stereotype Detection

  Manuela Nayantara and Sarah Jane Delany

  Gender stereotypes are perceptions about the typical physical, emotional, and social characteristics of individuals. Unlike gender bias which can result in the systematic and unfair treatment of individuals based on their gender, gender stereotypes do not always perpetuate a negative impact. Hence, there is no solid definition that frames what is considered a gender-stereotype in text. In addition, there is also a lack of labelled gender-stereotype datasets. This has led to most of the work in literature being about gender bias and not gender stereotypes. Therefore, in our research, we frame a clear definition of what constitutes a gender-stereotype in text. And using fairness benchmark datasets that were not originally intended for gender-stereotype studies, we propose an end-to-end approach to automatically identify language features that influence a gender-stereotypical sentence or text. We do this using the learned attention weights of a model. Our evaluations on the features that the model was using to predict gender-stereotypes aligned with our definition of gender-stereotypes.

• 

  Computer Science Outreach to Inform Secondary School Students’ Perceptions of Computer Science: Preliminary Findings

  Karen Nolan, Roisin Faherty, Keith Quille, Keith Nolan, Amanda O'Farrell, and Brett A. Becker

  This poster describes a longitudinal K-12 outreach programme to promote Computer Science in Ireland, which ran over a three-year period from 2017- 2020. A pilot phase was conducted in the first year from 2017-2018 with 2900 students participating. The implementation phase began in 2018, when 7320 students participated across the 2018-2019 and 2019-2020 academic years. The programme consisted of a free onsite school delivery of a two-hour camp that introduced students to a range of Computing topics: addressing computing perceptions, introduction to coding, and exploration of computational thinking. Schools self-selected, and the programme reached a large number of schools with varied socio-economic and gender diversity, along with schools across every county in Ireland. The student ages ranged from third class (7 - 8 years old) in primary school to sixth year in second level (17 - 18 years old). This poster focuses specifically on the research data collected during the implementation phase (N=1211) from secondary school students (12 - 18 years old) not enrolled in the formal Leaving Certificate Computer Science subject. Looking at student perceptions of Computer Science and if the outreach positively impacted those perceptions and built student interest in pursuing further study in Computer Science.

• 

  CSLINC: A Nationwide CS MOOC for Second-level Students

  Karen Nolan, Keith Quille, and Brett A. Becker

  This poster introduces CSLINC, a free scaffolded MOOC framework tailored to second-level students in Ireland that consists of: an online platform built for accessibility; a suite of modules developed upon international best practices with varying co-creators; and automated assessment and certificates of completion. Its aim is to provide content to promote national CS curricula to all second-level students in Ireland. In September 2021, CSLINC launched to 10,000 students across 100 schools. Future work will include collecting and collating research to validate CSLINC’s goals, scaffolding that will build foundations for national curriculum learning outcomes, and measure its impact on students, their perceptions and follow on CS uptake at second-level in Ireland.

• 

  The European Commission and AI: Guidelines, Acts and Plans Impacting the Teaching Of AI and Teaching With AI

  Keith Quille, Brett A. Becker, and Lidia Vidal-Meliá

  Recent developments, guidelines, and acts by the European Commission have started to frame policy for AI and related areas such as ML and data, not only for the broader community, but in the context of education specifically. This poster presents a succinct overview of these developments. Specifically, we look to bring together all publications that might impact the teaching of AI (for example, teacher expectations in the coming years around AI competencies) and publications that affect the use of AI in the classroom. We mean using tools and systems that incorporate both ‘Good Old Fashioned’ AI and those that can directly impact students. This poster is of value to both the European and the wider CER communities and practitioners, as it brings together several guidelines, acts, and plans that are not easily searchable or linked. The publications presented in this poster will impact the teaching of AI and teaching with AI in Europe, and insights can be drawn and compared for other jurisdictions as the educational world adapts to and with AI.

• 

  A Mode Sum Regularization Prescription in Quantum Field Theory in Curved Spacetimes in Higher Dimensions and for Charged Scalar Fields

  Eoin Scanlon

  Semi-classical gravity combines classical treatment of the gravitational field with quantum mechanical treatment of matter fields. A significant challenge however is the divergence contained within the stress-energy tensor when solving the semi-classical Einstein equations. This work extends to higher dimensions an extremely efficient method for renormalizing the stress-energy tensor of a quantum scalar field in spherically-symmetric black hole spacetimes, thereby removing the divergences. The method applies to a scalar field with arbitrary field parameters. The utility of the method is demonstrated by computing the renormalized stress-energy tensor for a scalar field in the Schwarzschild black hole spacetime for odd dimensions.