Skip to main content

Institute for Data Science and Artificial Intelligence

Uncertainty Quantification

Theme Leads: Prof Danny Williamson and Prof Peter Challenor

In order to understand the world we use models. These may be derived from the laws of physics and involve the use of partial differential equations or they may be empirical and based on the analysis of data. Computationally such models can be very expensive to run. Increasingly numerical models are used as part of the process of decision making in the real world, for example in the design of large engineering projects or in our response to climate change. But these models can never be perfect and there is always uncertainty on their predictions. This needs to be taken into account when they are used in decision making. Researchers at Exeter are concerned with the development of methods to quantify the uncertainty in such circumstances and to incorporate this uncertainty into the decision making process. We work in a number of application areas involving decision making using complex models in the environment, climate, engineering, epidemiology and healthcare.