Data Science

The world is continuously impacted by data, transforming science, society, industries, policies, and even entire economies. It is then no surprise that analysis of such tremendous amounts of data needs to become an integral part of all aspects of life. And it is because of this we have seen Data Science (DS) gaining an impressive momentum in recent decade, with many organisational entities adopting data-driven approaches to decision making.

Despite this, many areas can investigate and improve further the application of DS techniques, while others have still to explore their potential at all. To this end, we strive to support and elaborate on the use of such techniques to gain insights, discover patterns, and generate models that have the potential to contribute to ambitious initiatives, initiatives that aim to fundamentally improve various aspects of life. Such endeavour is based on interdisciplinary collaboration, attempting to reach and connect various research domains, to provide meaningful and unified solutions, driven by data.

These areas include:

Multi-Modal Age Assurance in Mixed Reality Environments for Online Child Safety

Ensuring online child safety has become a critical concern in recent years due to the potential exposure to inappropriate content and grooming by online predators. The Online Safety and Media Regulation Bill was signed into Irish law in December 2022, beginning a continuous legislative effort in this field, with more laws expected to follow in the coming years. Age assurance can become a vital tool for mitigating online risks, especially in mixed reality environments, where avatars may not accurately reflect users' real-world ages, making it easier for online predators to groom children in highly social and immersive environments.

While there have been studies exploring the use of AI for age verification in general, the specific application of these technologies in mixed reality environments is still relatively new. Despite the progress made so far, there is still a significant amount of work needed to create and verify solutions that strike a balance between accuracy and privacy while minimising the impact on user experience.

To address these challenges, this project aims to develop age assurance technologies that are both effective and accurate while preserving privacy and integrating seamlessly into the environment to avoid disrupting user experience. These will be integrated in a multi-modal interface for age in mixed reality environments.

Although the solution requires a multitude of cross-discipline technologies, the implementation of the main component draws extensively from DS. Specifically, it attempts to combine the latest developments in the areas of machine/deep learning, computer vision, NLP, and biometric analysis to analyse physical attributes, contextual information, and biometric data.

Contributors: Dr Christina Thorpe, Malik Awais Khan, and Dr Aurelia Power.

 

Enhancing Learning for Children with Autism: The OpenEarl Project

With the advances and availability of diagnosis, more children than ever are being diagnosed with an Autistic Spectrum Disorder (ASD) at an early age (2-5), leading to early interventions becoming a main stay in the ongoing effort to provide effective life changing skills that are built on year by year by educators. Two of the main issues commonly encountered by learners and relevant others are the cost barriers for software for augmentative and alternative communication (AAC) tools and education software solutions that replicate real-world scenarios. Although ASD is an issue for many children, little effort has been placed on modular reusability and extensible software solutions. 

The Open Early Intervention Platform (OpenEarl) project aims to provide a set of web-based mobile first platform independent solutions for the complete development cycle of resources for autism spectrum disorder children to aid education.  These include:

• An open ASD specific early intervention digital resource platform for use by children, educators, and parents to aid the creation, use and monitoring/reporting of daily learning activities.
• A text based parsing tool for automatic education resource generation and rendering using open data sources.
• Real-time monitoring and reporting of child progression and comprehension mapped to predefined learning outcomes set by educators and behaviour analysts.
• Resource progression identification and suggestion during education activities to encourage natural progression fitted to a child's learning abilities and learner specific requirements.
• Education model alignment monitor, e.g., dynamic resource difficulty monitors and difficulty adjustment tool.

While the objectives of this project are multifaceted, involving a variety of technologies and methodologies across several domains, DS plays a major role in designing and developing the final solution, by harnessing real-time analytics, machine learning, artificial intelligence, and natural language processing advancement to further tailor the learning experience of the child with respect to education ability, stimulus requirements and sensory issues that are unique to each child. For instance,

- we aim to identify what artificial intelligence and data analytics techniques can be employed to aid the progression of a learner with autism spectrum disorder though monitoring and dynamic resource modification. 

- we also aim to identify whether linguistics and natural language processing can aid the process of automatic generation of learning resources. 

Contributors: Dr Kyle Goslin, Dr Markus Hofmann, and Dr Aurelia Power.

 

Check out our other Research Areas