Fast Methods for Creating Accurate Emulators via Sparse Grid Designs

• Activity Number: 6
• Type: Invited
• Date/Time: Sunday, August 9, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Matthew Plumlee*
• Companies: Georgia Institute of Technology
• Keywords: Computer expirements ; Gaussian process ; High dimensional expirements

Abstract

Gaussian processes are widely employed to develop a predictor of an expensive function based on observations. In high dimensional settings, the traditional framework for developing a predictor is often too computationally expensive to be practical. We demonstrate that when sparse grid experimental designs are employed the resulting optimal predictor is quick to compute and has reasonable accuracy.