Mark Girolami

We develop methods for refining models using noisy and indirect observations.

New models simplify handling of function-based data in various fields.

Combining physics models and deep learning to improve predictions and manage uncertainty.