Rachel Prudden: Probabilistic modelling for atmospheric science: beyond the noise

In both atmospheric science and machine learning, it is important to capture the uncertainty of predictions. This information can avoid the dangers of relying on over-confident predictions which may be incorrect, and help to understand the potential for high-impact rare events. Nonetheless, to focus only on capturing uncertainty risks giving an incomplete picture of the strengths of probabilistic modelling. At their heart, probabilistic models are about information. How does a new observation influence our mental model of the world? In this talk, I will discuss the use of probabilistic models in atmospheric science and how they can aid scientists in interpreting incomplete observations. I will also present some early work in this direction for spatial reasoning and time-series analysis.

Rachel is a senior scientist in the Informatics Lab, working on ways to combine physics-based numerical simulation with machine learning. Her research spans several projects with collaborators in the Met Office and beyond - active projects include emulation of gravity wave parameterisations with ML models; applications of causal inference to climate indices; using spatial statistical models to understand convective-scale variability; and new methods for spatial verification. She is currently undertaking a PhD at the University of Exeter, with a focus on using Gaussian random fields for stochastic super-resolution of convective-scale fields.

Видео Rachel Prudden: Probabilistic modelling for atmospheric science: beyond the noise канала Oxford ML and Physics Seminars