Theory-guided auto-encoder for surrogate construction and inverse modeling

DOI10.1016/j.cma.2021.114037zbMath1502.65291arXiv2011.08618OpenAlexW3104994177MaRDI QIDQ2237777

Haibin Chang, Nanzhe Wang, Dongxiao Zhang

Publication date: 28 October 2021

Published in: Computer Methods in Applied Mechanics and Engineering (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2011.08618

zbMATH Keywords

uncertainty quantificationinverse modelingconvolutional neural network (CNN)auto-encodersurrogate constructiontheory-guided auto-encoder (TgAE)

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Computer aspects of numerical algorithms (65Y99)

Related Items

Surrogate and inverse modeling for two-phase flow in porous media via theory-guided convolutional neural network ⋮ Scientific machine learning through physics-informed neural networks: where we are and what's next ⋮ Deep reinforcement learning for optimal well control in subsurface systems with uncertain geology ⋮ HRW: Hybrid Residual and Weak Form Loss for Solving Elliptic Interface Problems with Neural Network ⋮ Deep-learning-based upscaling method for geologic models via theory-guided convolutional neural network ⋮ Use of multifidelity training data and transfer learning for efficient construction of subsurface flow surrogate models

Uses Software

• MODFLOW
• MODFLOW-2000
• ImageNet
• Adam
• AlexNet

Cites Work

• Unnamed Item
• A deep-learning-based surrogate model for data assimilation in dynamic subsurface flow problems
• An efficient, high-order perturbation approach for flow in random porous media via Karhunen-Loève and polynomial expansions.
• Levenberg-Marquardt forms of the iterative ensemble smoother for efficient history matching and uncertainty quantification
• Efficient uncertainty quantification for dynamic subsurface flow with surrogate by theory-guided neural network
• Physics-informed semantic inpainting: application to geostatistical modeling
• Surrogate modeling for fluid flows based on physics-constrained deep learning without simulation data
• Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Large-Scale Machine Learning with Stochastic Gradient Descent