Landmark-Warped Emulators for Models with Misaligned Functional Response
- Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
- Univ. of California, Santa Cruz, CA (United States)
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
Many computer models output functional data, and in some cases, these functional data have similar, but misaligned, shape characteristics. In this paper, we introduce a general approach for building emulators for computer models that output misaligned functional data when key values in the functional response (landmarks) can be easily identified. This approach has two main parts: modeling the aligned (using the landmarks) functional data, and modeling the functions that map the misaligned data to the aligned space (warping functions). As the warping functions are required to be monotonic, we give special attention to modeling monotonic functional response data. We discuss how our approach can be easily applied for a variety of typical emulators, such as Gaussian processes, Bayesian multivariate adaptive regression splines, and Bayesian additive regression trees, and how sensitivity analysis can be performed. We demonstrate our approach by building emulators for two applications: (1) a high-energy-density physics computer model used to simulate inertial confinement fusion ignition experiments, where model outputs are highly misaligned, and (2) a multiphysics continuum hydrocode used to simulate high-velocity impact experiments, where model outputs are only slightly misaligned. In case (1) traditional methods cannot be applied, while in (2) they can be applied, but the proposed method performs significantly better.
- Research Organization:
- Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE
- Grant/Contract Number:
- 89233218CNA000001
- OSTI ID:
- 1844165
- Report Number(s):
- LA-UR-21-30482
- Journal Information:
- SIAM/ASA Journal on Uncertainty Quantification, Vol. 10, Issue 1; ISSN 2166-2525
- Publisher:
- Society for Industrial and Applied Mathematics (SIAM)Copyright Statement
- Country of Publication:
- United States
- Language:
- English
Similar Records
Linking big models to big data: efficient ecosystem model calibration through Bayesian model emulation
A 5 {mu}m IMAGE OF {beta} PICTORIS b AT A SUB-JUPITER PROJECTED SEPARATION: EVIDENCE FOR A MISALIGNMENT BETWEEN THE PLANET AND THE INNER, WARPED DISK