Workshop on "Reducing dimensions and cost for UQ in complex systems"
Isaac Newton Institute, Cambridge
5 - 9 March 2018
https://www.newton.ac.uk/event/unqw03
Deadline for applications: 3rd January 2017
Uncertainty quantification (UQ) in complex mathematical models is a huge computational challenge for many reasons. Simple UQ tasks such as the estimation of statistical properties of system outputs often require multiple calls to a deterministic solver. A single solver call is already very expensive for complex mathematical models. Advanced UQ tasks such as sensitivity and reliability analysis, parameter identification, or optimal control and design often involve several layers of increasing complexity where each layer requires the performance of a specific UQ task. This workshop will address efficient numerical and statistical methods for reducing the overall cost of solving discrete problems that arise in UQ studies, focusing on methodologies that reduce the dimension of the problems to be solved.
Talks will be organised around topics such as: multilevel and multifidelity methods; reduced basis methods; dimension reduction strategies; low rank and tensor methods; challenges in Gaussian process emulation, and active subspaces. Workshop speakers and participants will be encouraged to explore connections between these topics. There will also be two contributed poster presentations. Please indicate on the registration form if you are interested in presenting a poster. A limited amount of funding may be available to support PhD students.
This workshop is the third event in a six-month programme on Uncertainty Quantification at the Isaac Newton Institute.