Deep multiscale model learning
From MaRDI portal
Artificial neural networks and deep learning (68T07) Multigrid methods; domain decomposition for initial value and initial-boundary value problems involving PDEs (65M55) Flows in porous media; filtration; seepage (76S05) Mathematical modeling or simulation for problems pertaining to fluid mechanics (76-10)
Abstract: The objective of this paper is to design novel multi-layer neural network architectures for multiscale simulations of flows taking into account the observed data and physical modeling concepts. Our approaches use deep learning concepts combined with local multiscale model reduction methodologies to predict flow dynamics. Using reduced-order model concepts is important for constructing robust deep learning architectures since the reduced-order models provide fewer degrees of freedom. Flow dynamics can be thought of as multi-layer networks. More precisely, the solution (e.g., pressures and saturations) at the time instant depends on the solution at the time instant and input parameters, such as permeability fields, forcing terms, and initial conditions. One can regard the solution as a multi-layer network, where each layer, in general, is a nonlinear forward map and the number of layers relates to the internal time steps. We will rely on rigorous model reduction concepts to define unknowns and connections for each layer. In each layer, our reduced-order models will provide a forward map, which will be modified ("trained") using available data. It is critical to use reduced-order models for this purpose, which will identify the regions of influence and the appropriate number of variables. Because of the lack of available data, the training will be supplemented with computational data as needed and the interpolation between data-rich and data-deficient models. We will also use deep learning algorithms to train the elements of the reduced model discrete system. We will present main ingredients of our approach and numerical results. Numerical results show that using deep learning and multiscale models, we can improve the forward models, which are conditioned to the available data.
Recommendations
- Model-Based Deep Learning
- A multi-stage deep learning based algorithm for multiscale model reduction
- Deep Gaussian mixture models
- Efficient hybrid explicit-implicit learning for multiscale problems
- Learning with deep cascades
- Multimodal learning with deep Boltzmann machines
- Multiresolution convolutional autoencoders
Cites work
- A Multiscale Finite Element Method for Numerical Homogenization
- A Sub-Grid Structure Enhanced Discontinuous Galerkin Method for Multiscale Diffusion and Convection-Diffusion Problems
- A conservative local multiscale model reduction technique for Stokes flows in heterogeneous perforated domains
- A multiscale method for porous microstructures
- Adaptive reduced basis finite element heterogeneous multiscale method
- An adaptive GMsFEM for high-contrast flow problems
- Approximation by superpositions of a sigmoidal function
- Bayesian deep convolutional encoder-decoder networks for surrogate modeling and uncertainty quantification
- Constraint energy minimizing generalized multiscale finite element method in the mixed formulation
- Generalized Multiscale Finite Element Methods for Wave Propagation in Heterogeneous Media
- Generalized multiscale finite element method for elasticity equations
- Generalized multiscale finite element methods (GMsFEM)
- Global-local nonlinear model reduction for flows in heterogeneous porous media
- Implementation of a locally conservative numerical subgrid upscaling scheme for two-phase Darcy flow
- Inferring solutions of differential equations using noisy multi-fidelity data
- Machine learning of linear differential equations using Gaussian processes
- Mathematical homogenization of nonperiodic heterogeneous media subjected to large deformation transient loading
- Metric-based upscaling
- Multi-output separable Gaussian process: towards an efficient, fully Bayesian paradigm for uncertainty quantification
- Multiscale empirical interpolation for solving nonlinear PDEs
- Multiscale finite element methods for high-contrast problems using local spectral basis functions
- Multiscale finite element methods for nonlinear problems and their applications
- Non-local multi-continua upscaling for flows in heterogeneous fractured media
- Online adaptive local multiscale model reduction for heterogeneous problems in perforated domains
- Reduced basis approximation for nonlinear parametrized evolution equations based on empirical operator interpolation
- Solution of inverse problems with limited forward solver evaluations: a Bayesian perspective
- Solving for high-dimensional committor functions using artificial neural networks
- Space-time multiscale model for wave propagation in heterogeneous media
- The deep Ritz method: a deep learning-based numerical algorithm for solving variational problems
- The heterogeneous multiscale finite element method for elliptic homogenization problems in perforated domains
- The heterogeneous multiscale methods
- The variational multiscale method -- a paradigm for computational mechanics
- Two-scale FEM for homogenization problems
Cited in
(40)- Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data
- Learning computational upscaling models for a class of convection-diffusion equations
- A multiscale neural network based on hierarchical nested bases
- Weak SINDy for partial differential equations
- <scp>A</scp> second‐order reduced multiscale method for nonlinear shell structures with orthogonal periodic configurations
- Multicontinuum homogenization for coupled flow and transport equations
- A Variational Neural Network Approach for Glacier Modelling with Nonlinear Rheology
- A deep domain decomposition method based on Fourier features
- Solver-free classical computational homogenization for nonlinear periodic heterogeneous media
- MIM: a deep mixed residual method for solving high-order partial differential equations
- AMS-Net: Adaptive Multiscale Sparse Neural Network with Interpretable Basis Expansion for Multiphase Flow Problems
- Deep learning for PDE-based inverse problems. Abstracts from the workshop held October 27 -- November 1, 2024
- Automated construction of effective potential via algorithmic implicit bias
- Bayesian sparse learning with preconditioned stochastic gradient MCMC and its applications
- Learning-based multi-continuum model for multiscale flow problems
- Prediction of discretization of online GMsFEM using deep learning for Richards equation
- Prediction of numerical homogenization using deep learning for the Richards equation
- Higher-order multi-scale deep Ritz method (HOMS-DRM) and its convergence analysis for solving thermal transfer problems of composite materials
- Convolutional neural network based reduced order modeling for multiscale problems
- A multi-stage deep learning based algorithm for multiscale model reduction
- A Multiscale Neural Network Based on Hierarchical Matrices
- Deep global model reduction learning in porous media flow simulation
- Reduced-order deep learning for flow dynamics. The interplay between deep learning and model reduction
- Learning rays via deep neural network in a ray-based IPDG method for high-frequency Helmholtz equations in inhomogeneous media
- Efficient hybrid explicit-implicit learning for multiscale problems
- Higher-order multi-scale physics-informed neural network (HOMS-PINN) method and its convergence analysis for solving elastic problems of authentic composite materials
- Multicontinuum splitting schemes for multiscale wave problems
- A deep learning based nonlinear upscaling method for transport equations
- Solving nonconvex energy minimization problems in martensitic phase transitions with a mesh-free deep learning approach
- DeepParticle: learning invariant measure by a deep neural network minimizing Wasserstein distance on data generated from an interacting particle method
- Physics-informed data-driven model for fluid flow in porous media
- A data-driven surrogate to image-based flow simulations in porous media
- Multilevel minimization for deep residual networks
- Globally Convergent Multilevel Training of Deep Residual Networks
- Solver-free reduced order homogenization for nonlinear periodic heterogeneous media
- A Bayesian multiscale deep learning framework for flows in random media
- Adaptive deep density approximation for stochastic dynamical systems
- Multiscale data-driven modeling for nonlinear dynamical system with external inputs in porous media
- MFPC-Net: multi-fidelity physics-constrained neural process
- A seamless multiscale operator neural network for inferring bubble dynamics
This page was built for publication: Deep multiscale model learning
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2223279)
