Conservative estimate for the excursion set of a deterministic function under a Gaussian random field model
Speaker: Dario Azzimonti (UniBern)
Abstract
The problem of estimating the excursion set of a deterministic function under a limited evaluation budget can be approached with Gaussian random field (GRF) modelling. Here we review two recent techniques based on Gaussian random field priors and we propose a fast algorithm to compute conservative estimates from the joint distribution of the posterior field.
Outline:
- Introduction to GRF modelling in the context of estimating excursion sets of deterministic functions;
- Definition of conservative estimates of excursion sets;
- Conservative excursion set estimate using a novel MCQMC algorithm.
