Emulator Generation Gadget

RESOURCE

Abstract

This software automates the production of standalone emulators for complex computer models. The end result of running the software is a piece of code called an emulator, a statistical approximation which predicts the output of complex computer simulation, but at a much faster speed. The emulator is built using a training set of inputs and outputs from the simulations. The input-output pairs are used in a Bayesian scheme to estimate the parameters of a Gaussian process based regression model that learns the input-output relationship. The estimated parameters are combined with a generic C skeleton to produce a compile-able instantiation of the emulator.
Developers:
Sweeney, Christine [1] Biswas, Ayan [1] Moran, Kelly [1] Lawrence, Earl [1] Gattiker, James [1]
  1. Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
Release Date:
2018-04-04
Project Type:
Open Source, Publicly Available Repository
Software Type:
Scientific
Licenses:
BSD 3-clause "New" or "Revised" License
Sponsoring Org.:
Code ID:
10487
Site Accession Number:
C18002
Research Org.:
Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
Country of Origin:
United States

RESOURCE

Citation Formats

Sweeney, Christine, Biswas, Ayan, Moran, Kelly, Lawrence, Earl, and Gattiker, James. Emulator Generation Gadget. Computer Software. https://github.com/lanl/EGG. USDOE. 04 Apr. 2018. Web. doi:10.11578/dc.20180419.1.
Sweeney, Christine, Biswas, Ayan, Moran, Kelly, Lawrence, Earl, & Gattiker, James. (2018, April 04). Emulator Generation Gadget. [Computer software]. https://github.com/lanl/EGG. https://doi.org/10.11578/dc.20180419.1.
Sweeney, Christine, Biswas, Ayan, Moran, Kelly, Lawrence, Earl, and Gattiker, James. "Emulator Generation Gadget." Computer software. April 04, 2018. https://github.com/lanl/EGG. https://doi.org/10.11578/dc.20180419.1.
@misc{ doecode_10487,
title = {Emulator Generation Gadget},
author = {Sweeney, Christine and Biswas, Ayan and Moran, Kelly and Lawrence, Earl and Gattiker, James},
abstractNote = {This software automates the production of standalone emulators for complex computer models. The end result of running the software is a piece of code called an emulator, a statistical approximation which predicts the output of complex computer simulation, but at a much faster speed. The emulator is built using a training set of inputs and outputs from the simulations. The input-output pairs are used in a Bayesian scheme to estimate the parameters of a Gaussian process based regression model that learns the input-output relationship. The estimated parameters are combined with a generic C skeleton to produce a compile-able instantiation of the emulator.},
doi = {10.11578/dc.20180419.1},
url = {https://doi.org/10.11578/dc.20180419.1},
howpublished = {[Computer Software] \url{https://doi.org/10.11578/dc.20180419.1}},
year = {2018},
month = {apr}
}