Presenter: Nicolas Durrande

Level: Professional

CPD: 6 hours

Kriging (or Gaussian process regression) has proven to be of great interest when trying to approximate a costly function to evaluate, in a closed form. The aim of the workshop is to show how useful surrogate models can be; to detail how to build such models in R; and to make clear the assumptions these models rely on. In the morning we introduce the concepts of Kriging through lectures and lab sessions. During the afternoon we see in more detail how to tune these models and how they can be used to solve optimization problems. The aim of the afternoon lab is to quickly find the settings of a catapult numerical simulator that give the longest shot.

This course precedes the RSS 2017 International Conference.