Using Deep Learning to Improve Ensemble Smoother: Applications to Subsurface Characterization

by   Jiangjiang Zhang, et al.

Ensemble smoother (ES) has been widely used in various research fields to reduce the uncertainty of the system-of-interest. However, the commonly-adopted ES method that employs the Kalman formula, that is, ES_(K), does not perform well when the probability distributions involved are non-Gaussian. To address this issue, we suggest to use deep learning (DL) to derive an alternative update scheme for ES in complex data assimilation applications. Here we show that the DL-based ES method, that is, ES_(DL), is more general and flexible. In this new update scheme, a high volume of training data are generated from a relatively small-sized ensemble of model parameters and simulation outputs, and possible non-Gaussian features can be preserved in the training data and captured by an adequate DL model. This new variant of ES is tested in two subsurface characterization problems with or without Gaussian assumptions. Results indicate that ES_(DL) can produce similar (in the Gaussian case) or even better (in the non-Gaussian case) results compared to those from ES_(K). The success of ES_(DL) comes from the power of DL in extracting complex (including non-Gaussian) features and learning nonlinear relationships from massive amounts of training data. Although in this work we only apply the ES_(DL) method in parameter estimation problems, the proposed idea can be conveniently extended to analysis of model structural uncertainty and state estimation in real-time forecasting studies.


page 6

page 8

page 11

page 12

page 13


SWAT Watershed Model Calibration using Deep Learning

Watershed models such as the Soil and Water Assessment Tool (SWAT) consi...

A comparative study of non-deep learning, deep learning, and ensemble learning methods for sunspot number prediction

Solar activity has significant impacts on human activities and health. O...

From feature selection to continues optimization

Metaheuristic algorithms (MAs) have seen unprecedented growth thanks to ...

From feature selection to continuous optimization

Metaheuristic algorithms (MAs) have seen unprecedented growth thanks to ...

Efficient Training Data Generation for Phase-Based DOA Estimation

Deep learning (DL) based direction of arrival (DOA) estimation is an act...

Deep-learning-driven Reliable Single-pixel Imaging with Uncertainty Approximation

Single-pixel imaging (SPI) has the advantages of high-speed acquisition ...

Please sign up or login with your details

Forgot password? Click here to reset