Parameter inference with deep jointly informed neural networks
Abstract
Abstract A common challenge in modeling inertial confinement fusion (ICF) experiments with computer simulations is that many of the simulation inputs are unknown and cannot be directly measured. Often, parameters that are measured in the experiment are used to infer the unknown inputs by solving the inverse problem: finding the set of simulation inputs that result in outputs consistent with the experimental observations. In ICF, this process is often referred to as a “post‐shot analysis.” Post‐shot analyses are challenging as the inverse problem is often highly degenerate, the input parameter space is vast, and simulations are computationally expensive. In this work, deep neural network models equipped with model uncertainty estimates are used to train inverse models, which map directly from output to input space, to find the distribution of post‐shot simulations that are consistent with experimental observations. The inverse model approach is compared to Markov chain Monte Carlo (MCMC) sampling of the forward model, which maps from input to output space, for parameter inference tasks of varying complexity. The inverse models perform best when searching vast parameter spaces for post‐shot simulations that are consistent with a large number of observables, where MCMC sampling can be prohibitively expensive. We demonstrate howmore »
- Authors:
-
[1];
[2];
- Lawrence Livermore National Laboratory Livermore California, Department of Nuclear Engineering Texas A&, University College Station Texas
- Lawrence Livermore National Laboratory Livermore California
- Department of Aerospace and Mechanical Engineering University of Notre Dame Notre Dame Indiana
- Publication Date:
- Sponsoring Org.:
- USDOE
- OSTI Identifier:
- 1546092
- Grant/Contract Number:
- DE‐AC52‐07NA27344
- Resource Type:
- Publisher's Accepted Manuscript
- Journal Name:
- Statistical Analysis and Data Mining
- Additional Journal Information:
- Journal Name: Statistical Analysis and Data Mining Journal Volume: 12 Journal Issue: 6; Journal ID: ISSN 1932-1864
- Publisher:
- Wiley Blackwell (John Wiley & Sons)
- Country of Publication:
- United States
- Language:
- English
Citation Formats
Humbird, Kelli D., Peterson, J. Luc, and McClarren, Ryan G.. Parameter inference with deep jointly informed neural networks. United States: N. p., 2019.
Web. doi:10.1002/sam.11435.
Humbird, Kelli D., Peterson, J. Luc, & McClarren, Ryan G.. Parameter inference with deep jointly informed neural networks. United States. https://doi.org/10.1002/sam.11435
Humbird, Kelli D., Peterson, J. Luc, and McClarren, Ryan G.. Thu .
"Parameter inference with deep jointly informed neural networks". United States. https://doi.org/10.1002/sam.11435.
@article{osti_1546092,
title = {Parameter inference with deep jointly informed neural networks},
author = {Humbird, Kelli D. and Peterson, J. Luc and McClarren, Ryan G.},
abstractNote = {Abstract A common challenge in modeling inertial confinement fusion (ICF) experiments with computer simulations is that many of the simulation inputs are unknown and cannot be directly measured. Often, parameters that are measured in the experiment are used to infer the unknown inputs by solving the inverse problem: finding the set of simulation inputs that result in outputs consistent with the experimental observations. In ICF, this process is often referred to as a “post‐shot analysis.” Post‐shot analyses are challenging as the inverse problem is often highly degenerate, the input parameter space is vast, and simulations are computationally expensive. In this work, deep neural network models equipped with model uncertainty estimates are used to train inverse models, which map directly from output to input space, to find the distribution of post‐shot simulations that are consistent with experimental observations. The inverse model approach is compared to Markov chain Monte Carlo (MCMC) sampling of the forward model, which maps from input to output space, for parameter inference tasks of varying complexity. The inverse models perform best when searching vast parameter spaces for post‐shot simulations that are consistent with a large number of observables, where MCMC sampling can be prohibitively expensive. We demonstrate how augmenting inverse models with autoencoders enable the inclusion of several dozen observables in the inverse mapping, reducing the degeneracy of the model and improving the accuracy of the post‐shot analysis.},
doi = {10.1002/sam.11435},
journal = {Statistical Analysis and Data Mining},
number = 6,
volume = 12,
place = {United States},
year = {2019},
month = {8}
}
https://doi.org/10.1002/sam.11435
Search WorldCat to find libraries that may hold this journalWeb of Science
Works referenced in this record:
The high velocity, high adiabat, “Bigfoot” campaign and tests of indirect-drive implosion scaling
journal, May 2018
- Casey, D. T.; Thomas, C. A.; Baker, K. L.
- Physics of Plasmas, Vol. 25, Issue 5
On the limited memory BFGS method for large scale optimization
journal, August 1989
- Liu, Dong C.; Nocedal, Jorge
- Mathematical Programming, Vol. 45, Issue 1-3
Deep Neural Network Initialization With Decision Trees
journal, May 2019
- Humbird, Kelli D.; Peterson, J. Luc; Mcclarren, Ryan G.
- IEEE Transactions on Neural Networks and Learning Systems, Vol. 30, Issue 5
Least angle regression
journal, April 2004
- Tibshirani, Robert; Johnstone, Iain; Hastie, Trevor
- The Annals of Statistics, Vol. 32, Issue 2
Interpreting inertial fusion neutron spectra
journal, February 2016
- Munro, David H.
- Nuclear Fusion, Vol. 56, Issue 3
Analysis of the neutron time-of-flight spectra from inertial confinement fusion experiments
journal, November 2015
- Hatarik, R.; Sayre, D. B.; Caggiano, J. A.
- Journal of Applied Physics, Vol. 118, Issue 18
Hedonic housing prices and the demand for clean air
journal, March 1978
- Harrison, David; Rubinfeld, Daniel L.
- Journal of Environmental Economics and Management, Vol. 5, Issue 1
Monte Carlo sampling methods using Markov chains and their applications
journal, April 1970
- Hastings, W. K.
- Biometrika, Vol. 57, Issue 1
BART: Bayesian additive regression trees
journal, March 2010
- Chipman, Hugh A.; George, Edward I.; McCulloch, Robert E.
- The Annals of Applied Statistics, Vol. 4, Issue 1
X-ray shadow imprint of hydrodynamic instabilities on the surface of inertial confinement fusion capsules by the fuel fill tube
journal, March 2017
- MacPhee, A. G.; Casey, D. T.; Clark, D. S.
- Physical Review E, Vol. 95, Issue 3
Three-dimensional HYDRA simulations of National Ignition Facility targets
journal, May 2001
- Marinak, M. M.; Kerbel, G. D.; Gentile, N. A.
- Physics of Plasmas, Vol. 8, Issue 5