02 / RESEARCH DIRECTIONS
Six ways into
the same curiosity.
Different data. Connected questions.
What can we learn—and what should we trust?
Clinical NLP
Extracting social and behavioural information from clinical notes, comparing biomedical text classifiers across corpora, and looking beyond a single performance score.
Computer Vision
Skin-lesion referral and breast MRI with clinical data: examining whether image-based predictions are reliable and useful beyond the development dataset.
Trustworthy AI
How well do uncertainty estimates travel across hospitals, and which conclusions can incomplete or selectively recorded evidence support?
Prediction Modelling
Dynamic acute kidney injury prediction, with attention to when information becomes available and how a model is evaluated.
Health Informatics
Short-term ICU laboratory prediction and eating-disorder dynamics in everyday life: connecting the timing of observations with the question being studied.
AI for Health
Learning from sleep and movement signals, with participant-level evaluation and careful comparisons of adaptive and fixed measurement strategies.