Computing Seminars – Winter 2026

The Winter 2026 Computing Seminar Series brought together researchers, students, academics, and industry experts to showcase innovative research in Computing at University of Buckingham and explore emerging developments across the computing discipline.

Presentation by Isuri Gatamanna Arachchige, a PhD student

The series began with a presentation by Isuri Gatamanna Arachchige, a PhD student, titled “Early Alzheimer’s Disease Identification via Multi-Omics Integration and Functional Biomarker Discovery”. Isuri’s work focuses on innovative approaches for the early detection of Alzheimer’s disease through multi-omics data integration and functional biomarker discovery. The project aims to improve early diagnosis by integrating genomic and transcriptomic data using interpretable deep learning models, thereby reducing reliance on black-box predictions and improving transparency in clinical applications.

Presentation by Rashmi Perera, a PhD student

Rashmi Perera, also a PhD student, delivered a thought-provoking talk based on her ongoing research titled “Artificial Intelligence in Viva Voce Examinations”. Her work explores the potential role of AI in supporting viva voce examinations in higher education, with a particular focus on the PhD level. The study investigates whether large language models can generate viva questions comparable in quality to those produced by human examiners, as well as assess student responses effectively.

Presentation by Madara Dassanayake, a PhD student

Madara Dassanayake, a final-year PhD student, delivered a captivating presentation titled “A Digital Twin Framework for Short-Term Solar Energy Forecasting Using Game Engines”. She proposed using digital twins of photovoltaic systems within Unreal Engine 5 to forecast energy output. Rather than relying on machine-learning models trained on historical energy data, her framework adopts physics-based modelling alongside live weather data to forecast solar power generation in real time.

Presentation by Dr Tom Longshaw, Director of Research & Development at Zizo

Dr Tom Longshaw, Director of Research & Development at Zizo, delivered a talk titled “Classification in the Presence of Sparse Matches”. He explained that classification seeks to identify rules that can predict characteristics of data based on available information. However, when datasets are highly imbalanced, models may appear to perform well by correctly identifying the majority class while failing to detect smaller but more important classes. This can result in misleadingly high accuracy, sometimes as high as 99%, while still limiting the discovery of meaningful classification rules. He also discussed how subjective human data, weak labels, aggregation issues, and black-box models can further complicate classification tasks. His presentation highlighted the use of genetic algorithms, which are often more transparent, repeatable, and easier to justify than neural networks, which can behave as black boxes.

The final seminar of Winter 2026 was delivered by Ethan Caldeira, another PhD student, titled “Classification of Elevated Trait Anxiety from MRI Data”. His research focuses on using deep learning and neuroimaging to identify functional and structural brain patterns associated with mental health conditions, particularly anxiety. He outlined two research pipelines based on MRI and fMRI data, covering preprocessing, region-of-interest selection, model development, classification, and experimental evaluation.

The Winter 2026 Computing Seminar Series showcased the breadth and quality of research within the School of Computing at the University of Buckingham. By bringing together PhD researchers and industry professionals, the series fostered knowledge exchange, constructive feedback, and valuable discussion, helping to strengthen ongoing research and encourage future collaboration.