TIA Centre: News

We are delighted to announce that one of the TIA Centre’s PhD students, Gozde Gunesli, has officially defended their PhD viva this week - congratulations Gozde!

Gozde’s research, titled “Machine Learning Methods for Multi-Resolution Problems in Computational Pathology”, presents novel artificial intelligence methods for robust, efficient, and interpretable analysis of routine histology slides. Her work introduces a federated learning approach for collaborative model training across multiple hospitals, designed to improve robustness to data heterogeneity, enabling models to generalise more effectively in real-world settings. She also develops a faster segmentation-free method, based on deep learning, for obtaining nuclear morphology measurements from tissue images directly, delivering accurate results with reduced computational overhead. Finally, she presents an interpretable machine learning framework to uncover cellular organisation patterns in the tumour microenvironment, revealing spatial arrangements linked to immune response, patient survival, and clinically relevant biomarkers. Together, these contributions pave the way for more generalisable, efficient, and biologically meaningful computational pathology applications for large-scale cancer research and clinical translation."

We were also pleased and grateful to welcome Prof Stephen McKenna from the University of Dundee as the external examiner.

We are grateful to GlaxoSmithKline (GSK) for contributing funding to Gozde’s PhD.