Modeling the dynamics of PDE systems with physics-constrained deep auto-regressive networks
From MaRDI portal
Publication:2222972
DOI10.1016/J.JCP.2019.109056zbMATH Open1454.65130arXiv1906.05747OpenAlexW2951392159WikidataQ126978599 ScholiaQ126978599MaRDI QIDQ2222972FDOQ2222972
Authors: Nicholas Geneva, Nicholas Zabaras 
Publication date: 28 January 2021
Published in: Journal of Computational Physics (Search for Journal in Brave)
Abstract: In recent years, deep learning has proven to be a viable methodology for surrogate modeling and uncertainty quantification for a vast number of physical systems. However, in their traditional form, such models can require a large amount of training data. This is of particular importance for various engineering and scientific applications where data may be extremely expensive to obtain. To overcome this shortcoming, physics-constrained deep learning provides a promising methodology as it only utilizes the governing equations. In this work, we propose a novel auto-regressive dense encoder-decoder convolutional neural network to solve and model non-linear dynamical systems without training data at a computational cost that is potentially magnitudes lower than standard numerical solvers. This model includes a Bayesian framework that allows for uncertainty quantification of the predicted quantities of interest at each time-step. We rigorously test this model on several non-linear transient partial differential equation systems including the turbulence of the Kuramoto-Sivashinsky equation, multi-shock formation and interaction with 1D Burgers' equation and 2D wave dynamics with coupled Burgers' equations. For each system, the predictive results and uncertainty are presented and discussed together with comparisons to the results obtained from traditional numerical analysis methods.
Full work available at URL: https://arxiv.org/abs/1906.05747
Recommendations
- Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data
- Adversarial uncertainty quantification in physics-informed neural networks
- Deep hidden physics models: deep learning of nonlinear partial differential equations
- PhyCRNet: physics-informed convolutional-recurrent network for solving spatiotemporal PDEs
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
uncertainty quantificationautoregressive modeldeep neural networksdynamic partial differential equationsphysics-informed machine learningconvolutional encoder-decoder
Artificial neural networks and deep learning (68T07) PDEs in connection with fluid mechanics (35Q35) Numerical methods for partial differential equations, initial value and time-dependent initial-boundary value problems (65M99)
Cites Work
- DGM: a deep learning algorithm for solving partial differential equations
- Pattern recognition and machine learning.
- Acceleration of Stochastic Approximation by Averaging
- 10.1162/15324430152748236
- Deep learning
- GMRES: A Generalized Minimal Residual Algorithm for Solving Nonsymmetric Linear Systems
- The Kuramoto-Sivashinsky equation: a bridge between PDE's and dynamical systems
- Title not available (Why is that?)
- Exponential time differencing for stiff systems
- Variational principles for nonpotential operators
- The numerical solution of linear ordinary differential equations by feedforward neural networks
- Solution of nonlinear ordinary differential equations by feedforward neural networks
- Limit set of trajectories of the coupled viscous Burgers' equations
- A computational meshfree technique for the numerical solution of the two-dimensional coupled Burgers' equations
- A first course in numerical analysis.
- Nonlinear analysis of hydrodynamic instability in laminar flames—II. Numerical experiments
- A search for the simplest chaotic partial differential equation
- Order and complexity in the Kuramoto-Sivashinsky model of weakly turbulent interfaces
- On the state space geometry of the Kuramoto-Sivashinsky flow in a periodic domain
- Quantifying model form uncertainty in Reynolds-averaged turbulence models with Bayesian deep neural networks
- Simulator-free solution of high-dimensional stochastic elliptic partial differential equations using deep neural networks
- Bayesian deep convolutional encoder-decoder networks for surrogate modeling and uncertainty quantification
- The deep Ritz method: a deep learning-based numerical algorithm for solving variational problems
- Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data
- Adversarial uncertainty quantification in physics-informed neural networks
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Solving for high-dimensional committor functions using artificial neural networks
- Reynolds averaged turbulence modelling using deep neural networks with embedded invariance
- Deep UQ: learning deep neural network surrogate models for high dimensional uncertainty quantification
- Deep hidden physics models: deep learning of nonlinear partial differential equations
- Using machine learning to replicate chaotic attractors and calculate Lyapunov exponents from data
- Title not available (Why is that?)
- A Proof that Artificial Neural Networks Overcome the Curse of Dimensionality in the Numerical Approximation of Black–Scholes Partial Differential Equations
Cited In (62)
- Pseudo grid-based physics-informed convolutional-recurrent network solving the integrable nonlinear lattice equations
- Deep learning in computational mechanics: a review
- \(\Phi\)-DVAE: physics-informed dynamical variational autoencoders for unstructured data assimilation
- f-PICNN: a physics-informed convolutional neural network for partial differential equations with space-time domain
- KAN-ODEs: Kolmogorov-Arnold network ordinary differential equations for learning dynamical systems and hidden physics
- Physics-informed multi-grid neural operator: theory and an application to porous flow simulation
- Learning on predictions: fusing training and autoregressive inference for long-term spatiotemporal forecasts
- Prediction of spatiotemporal dynamics using deep learning: coupled neural networks of long short-terms memory, auto-encoder and physics-informed neural networks
- Differentiability in unrolled training of neural physics simulators on transient dynamics
- CEENs: causality-enforced evolutional networks for solving time-dependent partial differential equations
- Adversarial uncertainty quantification in physics-informed neural networks
- Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data
- VI-DGP: a variational inference method with deep generative prior for solving high-dimensional inverse problems
- SeismicNET: physics-informed neural networks for seismic wave modeling in semi-infinite domain
- Graph network surrogate model for subsurface flow optimization
- Deep autoencoders for physics-constrained data-driven nonlinear materials modeling
- NeuralPDE: modelling dynamical systems from data
- Finite basis physics-informed neural networks (FBPINNs): a scalable domain decomposition approach for solving differential equations
- Bayesian physics informed neural networks for real-world nonlinear dynamical systems
- Deep learning of dynamics and signal-noise decomposition with time-stepping constraints
- A nonlocal physics-informed deep learning framework using the peridynamic differential operator
- Label-free learning of elliptic partial differential equation solvers with generalizability across boundary value problems
- Multi-fidelity physics constrained neural networks for dynamical systems
- Data-driven discovery of coarse-grained equations
- Bayesian reduced-order deep learning surrogate model for dynamic systems described by partial differential equations
- Advanced physics-informed neural networks for numerical approximation of the coupled Schrödinger-KdV equation
- Learning to Discretize: Solving 1D Scalar Conservation Laws via Deep Reinforcement Learning
- On obtaining sparse semantic solutions for inverse problems, control, and neural network training
- Understanding and Mitigating Gradient Flow Pathologies in Physics-Informed Neural Networks
- Multi-level convolutional autoencoder networks for parametric prediction of spatio-temporal dynamics
- Bi-fidelity modeling of uncertain and partially unknown systems using DeepONets
- A neural network multigrid solver for the Navier-Stokes equations
- PhySR: physics-informed deep super-resolution for spatiotemporal data
- Physics-incorporated convolutional recurrent neural networks for source identification and forecasting of dynamical systems
- Transformers for modeling physical systems
- CAN-PINN: a fast physics-informed neural network based on coupled-automatic-numerical differentiation method
- Deep physics corrector: a physics enhanced deep learning architecture for solving stochastic differential equations
- Numerical Analysis for Convergence of a Sample-Wise Backpropagation Method for Training Stochastic Neural Networks
- Variational inference formulation for a model-free simulation of a dynamical system with unknown parameters by a recurrent neural network
- Physics-informed ConvNet: learning physical field from a shallow neural network
- Bi-fidelity variational auto-encoder for uncertainty quantification
- Learning in Modal Space: Solving Time-Dependent Stochastic PDEs Using Physics-Informed Neural Networks
- A physically constrained variational autoencoder for geochemical pattern recognition
- Long-time integration of parametric evolution equations with physics-informed DeepONets
- Bounded nonlinear forecasts of partially observed geophysical systems with physics-constrained deep learning
- A hybrid partitioned deep learning methodology for moving interface and fluid-structure interaction
- 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
- Physics-informed graph neural Galerkin networks: a unified framework for solving PDE-governed forward and inverse problems
- Learning time-dependent PDEs with a linear and nonlinear separate convolutional neural network
- Adaptive learning of effective dynamics for online modeling of complex systems
- On the use of neural networks for full waveform inversion
- Solving inverse problems using conditional invertible neural networks
- Scientific machine learning through physics-informed neural networks: where we are and what's next
- Physics-constrained deep learning forecasting: an application with capacitance resistive model
- Identification of the flux function of nonlinear conservation laws with variable parameters
- Extended dynamic mode decomposition for inhomogeneous problems
- PhyGeoNet: physics-informed geometry-adaptive convolutional neural networks for solving parameterized steady-state PDEs on irregular domain
- An unsupervised latent/output physics-informed convolutional-LSTM network for solving partial differential equations using peridynamic differential operator
- On the use of graph neural networks and shape‐function‐based gradient computation in the deep energy method
- Deep capsule encoder–decoder network for surrogate modeling and uncertainty quantification
Uses Software
This page was built for publication: Modeling the dynamics of PDE systems with physics-constrained deep auto-regressive networks
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2222972)
