Mary Murphy, (PhD Researcher)

I am a PhD researcher in the School of Computing at Ulster University (commenced September 2025), developing machine learning models for sex-stratified diabetes prediction. My research addresses a significant gap in current diabetes screening: most risk prediction tools are not transparent and risk unfairly predicting risk across genders and ethnicities.

My research seeks to validate these models across US, UK, and Northern Ireland populations, culminating in a mobile app for community level diabetes screening deployed. My MSc research (Ulster, 2025) demonstrated that tree-based machine learning methods outperform deep learning for diabetes prediction, however, with variable fairness using unified, non sex stratified models.

With over 20 years' professional experience spanning manufacturing engineering and entrepreneurship (founding and running my own business for 13 years), I bring both technical expertise and lived experience to diabetes research. I am an active diabetes advocate with grassroots peer support groups NIDoc and DiabetesChat, volunteer with The NI Diabetes Network (board and digital subgroup levels), and sit on PPI Advisory panels with CURÁM Medical Devices and NI SPPG Pharmacy as well as being an IMPACT Associate with INNODIA for Type 1 Diabetes research.

Through integrating machine learning expertise within a robust, trustworthy and transparent framework with community understanding and involvement, I aim to develop screening tools that are clinically accurate, genuinely useful, and equitable across demographic groups.