Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 217 - High-Fidelity Gaussian Process Surrogate Modeling: Deep and Shallow
Type: Invited
Date/Time: Wednesday, August 11, 2021 : 10:00 AM to 11:50 AM
Sponsor: Section on Physical and Engineering Sciences
Abstract #316774
Title: Computer Model Emulation and Uncertainty Quantification Using a Deep Gaussian Process
Author(s): Derek Bingham* and Ilya Mandel and Daniel Williamson and Faezeh Yazdi
Companies: Simon Fraser University and School of Physics and Astronomy, Monash University and University of Exeter and Simon Fraser University
Keywords: Astrostatistics; computer experiments; hierarchical model

Computer models, are often used to explore physical systems. Increasingly, there are cases where the model is fast and the code is not readily available to scientists, but a large suite of model evaluations is available. In these cases, an emulator is used to stand in for the computer model. This work was motivated by a simulator for the chirp mass of binary black hole mergers where no output is observed for large portions of the input space and more that 10^6 simulator evaluations are available. This poses two problems: (i) the need to address the discontinuity when observing no chirp mass; and (ii) performing statistical inference with a large number of simulator evaluations. The traditional approach for emulation is to use a stationary Gaussian process (GP). Unfortunately, when the simulation design is large, evaluation of the GP likelihood is computationally intractable. In this talk, we propose to use a deep GP for computer model emulation. We explore the impact of the choices of when setting up the deep GP on posterior inference and apply the proposed approach to the real application.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2021 program