Bayesian deep convolutional encoder-decoder networks for surrogate modeling and uncertainty quantification
From MaRDI portal
Abstract: We are interested in the development of surrogate models for uncertainty quantification and propagation in problems governed by stochastic PDEs using a deep convolutional encoder-decoder network in a similar fashion to approaches considered in deep learning for image-to-image regression tasks. Since normal neural networks are data intensive and cannot provide predictive uncertainty, we propose a Bayesian approach to convolutional neural nets. A recently introduced variational gradient descent algorithm based on Stein's method is scaled to deep convolutional networks to perform approximate Bayesian inference on millions of uncertain network parameters. This approach achieves state of the art performance in terms of predictive accuracy and uncertainty quantification in comparison to other approaches in Bayesian neural networks as well as techniques that include Gaussian processes and ensemble methods even when the training data size is relatively small. To evaluate the performance of this approach, we consider standard uncertainty quantification benchmark problems including flow in heterogeneous media defined in terms of limited data-driven permeability realizations. The performance of the surrogate model developed is very good even though there is no underlying structure shared between the input (permeability) and output (flow/pressure) fields as is often the case in the image-to-image regression models used in computer vision problems. Studies are performed with an underlying stochastic input dimensionality up to where most other uncertainty quantification methods fail. Uncertainty propagation tasks are considered and the predictive output Bayesian statistics are compared to those obtained with Monte Carlo estimates.
Recommendations
- Deep UQ: learning deep neural network surrogate models for high dimensional uncertainty quantification
- Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data
- Conditional deep surrogate models for stochastic, high-dimensional, and multi-fidelity systems
- Scaling Up Bayesian Uncertainty Quantification for Inverse Problems Using Deep Neural Networks
- An adaptive surrogate modeling based on deep neural networks for large-scale Bayesian inverse problems
Cites work
- A Fast Learning Algorithm for Deep Belief Nets
- A machine learning approach for efficient uncertainty quantification using multiscale methods
- Bayesian model and dimension reduction for uncertainty propagation: applications in random media
- Cases for the nugget in modeling computer experiments
- Deep learning in fluid dynamics
- Gaussian processes for machine learning.
- Hyperband: a novel bandit-based approach to hyperparameter optimization
- Learning deep architectures for AI
- Multi-output local Gaussian process regression: applications to uncertainty quantification
- Multi-output separable Gaussian process: towards an efficient, fully Bayesian paradigm for uncertainty quantification
- Reducing the Dimensionality of Data with Neural Networks
- The Wiener--Askey Polynomial Chaos for Stochastic Differential Equations
- Visualizing data using t-SNE
Cited in
(only showing first 100 items - show all)- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Stochastic modeling of inhomogeneities in the aortic wall and uncertainty quantification using a Bayesian encoder-decoder surrogate
- Deep multiscale model learning
- Deep learning of vortex-induced vibrations
- Reduced-order deep learning for flow dynamics. The interplay between deep learning and model reduction
- GINNs: graph-informed neural networks for multiscale physics
- A probabilistic generative model for semi-supervised training of coarse-grained surrogates and enforcing physical constraints through virtual observables
- On generalized residual network for deep learning of unknown dynamical systems
- Simulator-free solution of high-dimensional stochastic elliptic partial differential equations using deep neural networks
- Accelerated reactive transport simulations in heterogeneous porous media using Reaktoro and Firedrake
- Discovery of dynamics using linear multistep methods
- An adaptive surrogate modeling based on deep neural networks for large-scale Bayesian inverse problems
- 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
- A deep-learning-based surrogate model for data assimilation in dynamic subsurface flow problems
- Data driven governing equations approximation using deep neural networks
- Use of multifidelity training data and transfer learning for efficient construction of subsurface flow surrogate models
- An intelligent multi-fidelity surrogate-assisted multi-objective reservoir production optimization method based on transfer stacking
- Quantifying total uncertainty in physics-informed neural networks for solving forward and inverse stochastic problems
- Active learning Bayesian support vector regression model for global approximation
- GAN-based priors for quantifying uncertainty in supervised learning
- Data-driven deep learning of partial differential equations in modal space
- Transfer learning based multi-fidelity physics informed deep neural network
- Scalable uncertainty quantification for deep operator networks using randomized priors
- Deep UQ: learning deep neural network surrogate models for high dimensional uncertainty quantification
- LaSDI: parametric latent space dynamics identification
- Simulation of the 3D hyperelastic behavior of ventricular myocardium using a finite-element based neural-network approach
- Machine learning in cardiovascular flows modeling: predicting arterial blood pressure from non-invasive 4D flow MRI data using physics-informed neural networks
- Quantifying model form uncertainty in Reynolds-averaged turbulence models with Bayesian deep neural networks
- Efficient uncertainty quantification for dynamic subsurface flow with surrogate by theory-guided neural network
- Deep-learning-based surrogate flow modeling and geological parameterization for data assimilation in 3D subsurface flow
- What machine learning can do for computational solid mechanics
- ConvPDE-UQ: convolutional neural networks with quantified uncertainty for heterogeneous elliptic partial differential equations on varied domains
- A deep domain decomposition method based on Fourier features
- Data-driven soliton solutions and model parameters of nonlinear wave models via the conservation-law constrained neural network method
- Error estimates and physics informed augmentation of neural networks for thermally coupled incompressible Navier Stokes equations
- An encoder-decoder deep surrogate for reverse time migration in seismic imaging under uncertainty
- A physics-informed variational DeepONet for predicting crack path in quasi-brittle materials
- Surrogate and inverse modeling for two-phase flow in porous media via theory-guided convolutional neural network
- A bi-fidelity surrogate modeling approach for uncertainty propagation in three-dimensional hemodynamic simulations
- Uncertainty propagation using infinite mixture of gaussian processes and variational Bayesian inference
- Theory-guided auto-encoder for surrogate construction and inverse modeling
- A Bayesian multiscale deep learning framework for flows in random media
- Mosaic flows: a transferable deep learning framework for solving PDEs on unseen domains
- PhyCRNet: physics-informed convolutional-recurrent network for solving spatiotemporal PDEs
- A representative volume element network (RVE-net) for accelerating RVE analysis, microscale material identification, and defect characterization
- The deep learning Galerkin method for the general Stokes equations
- Bayesian model and dimension reduction for uncertainty propagation: applications in random media
- A mesh-free method for interface problems using the deep learning approach
- Surrogate modeling of high-dimensional problems via data-driven polynomial chaos expansions and sparse partial least square
- Extended dynamic mode decomposition for inhomogeneous problems
- Solving and learning nonlinear PDEs with Gaussian processes
- Adaptive deep density approximation for Fokker-Planck equations
- Physics-Informed Probabilistic Learning of Linear Embeddings of Nonlinear Dynamics with Guaranteed Stability
- Multifidelity approaches for uncertainty quantification
- Solving inverse problems using conditional invertible neural networks
- Scaling Up Bayesian Uncertainty Quantification for Inverse Problems Using Deep Neural Networks
- Physics-informed generative adversarial networks for stochastic differential equations
- Model reduction and neural networks for parametric PDEs
- Bayesian distillation of deep learning models
- The Random Feature Model for Input-Output Maps between Banach Spaces
- Multifidelity data fusion in convolutional encoder/decoder networks
- Parametric generation of conditional geological realizations using generative neural networks
- Deep learning methods for partial differential equations and related parameter identification problems
- Conditional deep surrogate models for stochastic, high-dimensional, and multi-fidelity systems
- A physics-aware, probabilistic machine learning framework for coarse-graining high-dimensional systems in the small data regime
- An efficient neural network method with plane wave activation functions for solving Helmholtz equation
- Modeling the dynamics of PDE systems with physics-constrained deep auto-regressive networks
- Learning phase field mean curvature flows with neural networks
- A deep learning approach to Reduced Order Modelling of parameter dependent partial differential equations
- Uncertainty quantification of graph convolution neural network models of evolving processes
- Energy-informed graph transformer model for solid mechanical analyses
- Label-free learning of elliptic partial differential equation solvers with generalizability across boundary value problems
- Adaptive deep density approximation for stochastic dynamical systems
- Wasserstein generative adversarial uncertainty quantification in physics-informed neural networks
- Quantum Mechanics for Closure of Dynamical Systems
- Surrogate modeling for Bayesian inverse problems based on physics-informed neural networks
- A sample-efficient deep learning method for multivariate uncertainty qualification of acoustic-vibration interaction problems
- A heteroencoder architecture for prediction of failure locations in porous metals using variational inference
- Learning deep implicit Fourier neural operators (IFNOs) with applications to heterogeneous material modeling
- Epistemic uncertainty quantification in deep learning classification by the delta method
- A nonlinear-manifold reduced-order model and operator learning for partial differential equations with sharp solution gradients
- A framework for strategic discovery of credible neural network surrogate models under uncertainty
- Probabilistic partition of unity networks for high‐dimensional regression problems
- Deep learning of biological models from data: applications to ODE models
- Bi-Orthogonal fPINN: A Physics-Informed Neural Network Method for Solving Time-Dependent Stochastic Fractional PDEs
- Physics-informed data-driven model for fluid flow in porous media
- Fully probabilistic deep models for forward and inverse problems in parametric PDEs
- Stein's method meets computational statistics: a review of some recent developments
- opPINN: physics-informed neural network with operator learning to approximate solutions to the Fokker-Planck-Landau equation
- A data-driven surrogate to image-based flow simulations in porous media
- Kernel methods are competitive for operator learning
- Bi-fidelity variational auto-encoder for uncertainty quantification
- Operator learning using random features: a tool for scientific computing
- Deep capsule encoder–decoder network for surrogate modeling and uncertainty quantification
- A Data-Assisted Two-Stage Method for the Inverse Random Source Problem
- Output-based adaptive aerodynamic simulations using convolutional neural networks
- Stein variational gradient descent with local approximations
- f-PICNN: a physics-informed convolutional neural network for partial differential equations with space-time domain
- Mesh-Conv: convolution operator with mesh resolution independence for flow field modeling
Describes a project that uses
Uses Software
This page was built for publication: Bayesian deep convolutional encoder-decoder networks for surrogate modeling and uncertainty quantification
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1721865)
