His previous roles include Visiting Researcher at Microsoft Research in the AI4Science and AI teams, Co-Director of the UKRI Centre for Doctoral Training in the Application of Artificial Intelligence to the study of Environmental Risks (AI4ER CDT), and Course Director of the Machine Learning and Machine Intelligence MPhil programme. Rich has received over £30 million of funding as a Principal or Co-Investigator and has been awarded the Cambridge Students' Union Teaching Award for Lecturing.
Rich's current research interests include:
Probabilistic Machine Learning Fundamentals (including generative models and uncertainty-aware approaches)
Environmental Prediction (especially for weather, earth systems, and climate prediction)
Spatio-temporal Modelling (especially using a fusion of large-scale deep learning and probabilistic modelling for scientific applications)
Weather forecasting is critical for a range of human activities including transportation, agriculture, industry, as well as the safety of the general public. Machine learning models have the potential to transform the complex weather prediction pipeline, but current approaches still rely on numerical weather prediction (NWP) systems, limiting forecast speed and accuracy. In this talk I will introduce a machine learning model that replaces the entire operational NWP pipeline. Aardvark Weather, an end-to-end data-driven weather prediction system, ingests raw observations and outputs global gridded forecasts and local station forecasts. Further, it can be optimised end-to-end to maximise performance over quantities of interest. Global forecasts outperform an operational NWP baseline for multiple variables and lead times. Local station forecasts are skillful up to ten days lead time and achieve comparable and often lower errors than a post-processed global NWP baseline and a state-of-the-art end-to-end forecasting system with input from human forecasters. These forecasts are produced with a remarkably simple neural process model using just 8% of the input data and three orders of magnitude less compute than existing NWP and hybrid AI-NWP methods. We think that Aardvark Weather will be the starting point for a new generation of end-to-end machine learning models for medium-range forecasting that will reduce computational costs by orders of magnitude and enable the rapid and cheap creation of bespoke models for users in a variety of fields, including for the developing world where state-of-the-art local models are not currently available. Along the way, I will briefly discuss another project I have been involved with at Microsoft in which we built a foundation model of the atmosphere called Aurora, applying it to meteorological tasks, atmospheric chemistry, and ocean wave forecasting.