Research
Current Research
PhD in Statistics and Machine Learning (2025-2029) Imperial College London | Mathematics for Future Climate CDT
My research focuses on generative modelling, extreme value theory and uncertainty quantification with applications in machine learning weather prediction systems. My work includes developing methods to improve the reliability and interpretability of generative models, particularly when sampling outside their training distribution.
Recent Work
MSc Statistics Thesis “Approximate Bayesian Computation with Proper Scoring Rules: An Inference Framework for Calibrated Probabilistic Forecasts”
This report develops a framework combining ABC methods with proper scoring rules to improve probabilistic forecast calibration.
Research Interests
- Unsupervised learning
- Generative modelling
- Extreme value theory, extrapolation and out-of-distribution generalisation
- Probabilistic forecasting, calibration and uncertainty quantification in machine learning systems
- Bayesian methods and approximate inference
- Statistical computing and computational methods