Emulation of CPU-demanding reactive transport models: a comparison of Gaussian processes, polynomial chaos expansion, and deep neural networks
From MaRDI portal
Publication:2009855
Abstract: This paper presents a detailed comparison between 3 methods for emulating CPU-intensive reactive transport models (RTMs): Gaussian processes (GPs), polynomial chaos expansion (PCE) and deep neural networks (DNNs). Besides direct emulation of the simulated uranium concentration time series, replacing the original RTM by its emulator is also investigated for global sensitivity analysis (GSA), uncertainty propagation (UP) and probabilistic calibration using Markov chain Monte Carlo (MCMC) sampling. The selected DNN is found to be superior to both GPs and PCE in reproducing the input - output behavior of the considered 8-dimensional and 13-dimensional CPU-intensive RTMs. Furthermore, the two used PCE variants: standard PCE and sparse PCE (sPCE) appear to always provide the least accuracy while not differing much in performance. As a consequence of its better emulation capabilities, the used DNN outperforms the two other methods for UP. In addition, DNNs and GPs offer equally good approximations to the true first-order and total-order Sobol sensitivity indices while PCE does somewhat less well. Most surprisingly, despite its superior emulation skills the DNN approach leads to the worst solution of the considered synthetic inverse problem which involves 1224 measurement data with low noise. This apparently contradicting behavior is at least partially due to the small but complicated deterministic noise that affects the DNN-based predictions. Indeed, this complex error structure can drive the emulated solutions far away from the true posterior distribution. Overall, our findings indicate that when the available training set is relatively small (75 - to 500 input - output examples) and fixed beforehand, DNNs can well emulate RTMs but are not suited to emulation-based inversion. In contrast, GPs perform fairly well across all considered tasks: direct emulation, GSA, UP, and inversion.
Recommendations
- Computer emulation with nonstationary Gaussian processes
- Emulating dynamic non-linear simulators using Gaussian processes
- An efficient methodology for modeling complex computer codes with Gaussian processes
- Deep UQ: learning deep neural network surrogate models for high dimensional uncertainty quantification
- A mechanistic dynamic emulator
Cites work
- A Comparison of Three Methods for Selecting Values of Input Variables in the Analysis of Output from a Computer Code
- Analysis of variance designs for model output
- Approximation by superpositions of a sigmoidal function
- Bayesian sparse polynomial chaos expansion for global sensitivity analysis
- Deep learning
- Gaussian processes for machine learning.
- Global sensitivity indices for nonlinear mathematical models and their Monte Carlo estimates
- GPflow: a Gaussian process library using TensorFlow
- scientific article; zbMATH DE number 6016067 (Why is no real title available?)
- scientific article; zbMATH DE number 49187 (Why is no real title available?)
- scientific article; zbMATH DE number 2117879 (Why is no real title available?)
- scientific article; zbMATH DE number 1425054 (Why is no real title available?)
- Inference from iterative simulation using multiple sequences
- Principal component analysis.
- Reactive transport codes for subsurface environmental simulation
- Spectral Methods for Uncertainty Quantification
- Uncertainty management in simulation-optimization of complex systems. Algorithms and applications
- Variance based sensitivity analysis of model output. Design and estimator for the total sensitivity index
Cited in
(4)- Gaussian active learning on multi-resolution arbitrary polynomial chaos emulator: concept for bias correction, assessment of surrogate reliability and its application to the carbon dioxide benchmark
- Surrogate modeling in irreversible electroporation towards real-time treatment planning
- A data-driven approach for estimating functions in a multivariate nonparametric regression model based on B-splines with an application to geoscience
- A novel variable selection method in nonlinear multivariate models using B-splines with an application to geoscience
This page was built for publication: Emulation of CPU-demanding reactive transport models: a comparison of Gaussian processes, polynomial chaos expansion, and deep neural networks
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2009855)
