Model reduction of dynamical systems on nonlinear manifolds using deep convolutional autoencoders
From MaRDI portal
Abstract: Nearly all model-reduction techniques project the governing equations onto a linear subspace of the original state space. Such subspaces are typically computed using methods such as balanced truncation, rational interpolation, the reduced-basis method, and (balanced) POD. Unfortunately, restricting the state to evolve in a linear subspace imposes a fundamental limitation to the accuracy of the resulting reduced-order model (ROM). In particular, linear-subspace ROMs can be expected to produce low-dimensional models with high accuracy only if the problem admits a fast decaying Kolmogorov -width (e.g., diffusion-dominated problems). Unfortunately, many problems of interest exhibit a slowly decaying Kolmogorov -width (e.g., advection-dominated problems). To address this, we propose a novel framework for projecting dynamical systems onto nonlinear manifolds using minimum-residual formulations at the time-continuous and time-discrete levels; the former leads to manifold Galerkin projection, while the latter leads to manifold least-squares Petrov--Galerkin (LSPG) projection. We perform analyses that provide insight into the relationship between these proposed approaches and classical linear-subspace reduced-order models; we also derive a posteriori discrete-time error bounds for the proposed approaches. In addition, we propose a computationally practical approach for computing the nonlinear manifold, which is based on convolutional autoencoders from deep learning. Finally, we demonstrate the ability of the method to significantly outperform even the optimal linear-subspace ROM on benchmark advection-dominated problems, thereby demonstrating the method's ability to overcome the intrinsic -width limitations of linear subspaces.
Recommendations
- Non-intrusive model reduction of large-scale, nonlinear dynamical systems using deep learning
- The neural network shifted-proper orthogonal decomposition: a machine learning approach for non-linear reduction of hyperbolic equations
- A tailored convolutional neural network for nonlinear manifold learning of computational physics data using unstructured spatial discretizations
- A comprehensive deep learning-based approach to reduced order modeling of nonlinear time-dependent parametrized PDEs
- Neural network closures for nonlinear model order reduction
Cites work
- $\mathcal{H}_2$ Model Reduction for Large-Scale Linear Dynamical Systems
- A low-cost, goal-oriented `compact proper orthogonal decomposition' basis for model reduction of static systems
- A subspace approach to balanced truncation for model reduction of nonlinear control systems
- Adaptive h-refinement for reduced-order models
- Adaptive reduced basis methods for nonlinear convection-diffusion equations
- Adaptive subgradient methods for online learning and stochastic optimization
- Approximated Lax pairs for the reduced order integration of nonlinear evolution equations
- Conservative model reduction for finite-volume models
- Convergence rates for greedy algorithms in reduced basis methods
- Coupling of substructures for dynamic analyses.
- Decreasing the temporal complexity for nonlinear, implicit reduced-order models by forecasting
- Deep learning
- Efficient non linear model reduction via a least-squares Petrov-Galerkin projection and compressive tensor approximations
- Galerkin v. least-squares Petrov-Galerkin projection in nonlinear model reduction
- Hessian eigenmaps: Locally linear embedding techniques for high-dimensional data
- Interpolation of functions with parameter dependent jumps by transformed snapshots
- Interpolatory projection methods for parameterized model reduction
- Kolmogorov widths and low-rank approximations of parametric elliptic PDEs
- Laplacian Eigenmaps for Dimensionality Reduction and Data Representation
- Learning representations by back-propagating errors
- Linearly recurrent autoencoder networks for learning dynamics
- Model order reduction for parametrized nonlinear hyperbolic problems as an application to uncertainty quantification
- Model order reduction for problems with large convection effects
- MODEL REDUCTION FOR FLUIDS, USING BALANCED PROPER ORTHOGONAL DECOMPOSITION
- Neural network modeling for near wall turbulent flow.
- Nonlinear model order reduction based on local reduced-order bases
- Nonlinear reduced basis approximation of parameterized evolution equations via the method of freezing
- Online adaptive model reduction for nonlinear systems via low-rank updates
- Optimization methods for large-scale machine learning
- Practical quasi-Newton methods for solving nonlinear systems
- Principal component analysis in linear systems: Controllability, observability, and model reduction
- Progressive construction of a parametric reduced-order model for PDE-constrained optimization
- Reduced basis techniques for nonlinear conservation laws
- Reducing the Dimensionality of Data with Neural Networks
- Self-organized formation of topologically correct feature maps
- The GNAT method for nonlinear model reduction: effective implementation and application to computational fluid dynamics and turbulent flows
- The shifted proper orthogonal decomposition: a mode decomposition for multiple transport phenomena
- Turbulence, Coherent Structures, Dynamical Systems and Symmetry
- Turbulence, coherent structures, dynamical systems and symmetry.
- Visualizing data using t-SNE
Cited in
(only showing first 100 items - show all)- Data-driven nonintrusive reduced order modeling for dynamical systems with moving boundaries using Gaussian process regression
- Deep learning of thermodynamics-aware reduced-order models from data
- Reduced-order modeling of deep neural networks
- Numerical solution of the parametric diffusion equation by deep neural networks
- Model reduction and neural networks for parametric PDEs
- Registration-based model reduction in complex two-dimensional geometries
- POD-DL-ROM: enhancing deep learning-based reduced order models for nonlinear parametrized PDEs by proper orthogonal decomposition
- Reduced order models for Lagrangian hydrodynamics
- Adaptive POD-Galerkin technique for reservoir simulation and optimization
- Projection-based model reduction of dynamical systems using space-time subspace and machine learning
- Multi-fidelity regression using artificial neural networks: efficient approximation of parameter-dependent output quantities
- Deep learning of conjugate mappings
- Mapping of coherent structures in parameterized flows by learning optimal transportation with Gaussian models
- A learning-based projection method for model order reduction of transport problems
- Modal decomposition of flow data via gradient-based transport optimization
- Adaptive machine learning-based surrogate modeling to accelerate PDE-constrained optimization in enhanced oil recovery
- Local Lagrangian reduced-order modeling for the Rayleigh-Taylor instability by solution manifold decomposition
- Deep-HyROMnet: a deep learning-based operator approximation for hyper-reduction of nonlinear parametrized PDEs
- Model reduction of convection-dominated partial differential equations via optimization-based implicit feature tracking
- Latent-space time evolution of non-intrusive reduced-order models using Gaussian process emulation
- A theoretical analysis of deep neural networks and parametric PDEs
- A probabilistic generative model for semi-supervised training of coarse-grained surrogates and enforcing physical constraints through virtual observables
- On the stability of projection-based model order reduction for convection-dominated laminar and turbulent flows
- A long short-term memory embedding for hybrid uplifted reduced order models
- Non-intrusive model reduction of large-scale, nonlinear dynamical systems using deep learning
- Physics-informed machine learning for reduced-order modeling of nonlinear problems
- Meta-mgnet: meta multigrid networks for solving parameterized partial differential equations
- Model reduction for multi-scale transport problems using model-form preserving least-squares projections with variable transformation
- A fast and accurate physics-informed neural network reduced order model with shallow masked autoencoder
- Machine learning moment closure models for the radiative transfer equation. I: Directly learning a gradient based closure
- Registration-based model reduction of parameterized two-dimensional conservation laws
- Projection-tree reduced-order modeling for fast \(N\)-body computations
- The neural network shifted-proper orthogonal decomposition: a machine learning approach for non-linear reduction of hyperbolic equations
- A comparison of neural network architectures for data-driven reduced-order modeling
- Physics-informed cluster analysis and a priori efficiency criterion for the construction of local reduced-order bases
- A non-intrusive neural network model order reduction algorithm for parameterized parabolic PDEs
- Proper orthogonal decomposition and physical field reconstruction with artificial neural networks (ANN) for supercritical flow problems
- Inadequacy of linear methods for minimal sensor placement and feature selection in nonlinear systems: a new approach using secants
- Model order reduction for compressible flows solved using the discontinuous Galerkin methods
- Time-series machine-learning error models for approximate solutions to parameterized dynamical systems
- Machine learning for fast and reliable solution of time-dependent differential equations
- Domain-decomposition least-squares Petrov-Galerkin (DD-LSPG) nonlinear model reduction
- Data-driven reduced bond graph for nonlinear multiphysics dynamic systems
- Windowed space-time least-squares Petrov-Galerkin model order reduction for nonlinear dynamical systems
- Data-driven closure of projection-based reduced order models for unsteady compressible flows
- Data-driven model order reduction for problems with parameter-dependent jump-discontinuities
- A FOM/ROM hybrid approach for accelerating numerical simulations
- A pruning algorithm preserving modeling capabilities for polycrystalline data
- Synthetic nonlinear second-order oscillators on Riemannian manifolds and their numerical simulation
- A reduced order modeling method based on GNAT-embedded hybrid snapshot simulation
- Quadratic approximation manifold for mitigating the Kolmogorov barrier in nonlinear projection-based model order reduction
- LaSDI: parametric latent space dynamics identification
- Minimally-invasive parametric model-order reduction for sweep-based radiation transport
- Kolmogorov n-width and Lagrangian physics-informed neural networks: a causality-conforming manifold for convection-dominated PDEs
- Multi-fidelity surrogate modeling using long short-term memory networks
- Bayesian operator inference for data-driven reduced-order modeling
- Operator inference for non-intrusive model reduction with quadratic manifolds
- Parametric dynamic mode decomposition for reduced order modeling
- Adaptive sparse interpolation for accelerating nonlinear stochastic reduced-order modeling with time-dependent bases
- Model reduction for the material point method via an implicit neural representation of the deformation map
- Reduced order modeling for parametrized generalized Newtonian fluid flows
- Model order reduction of positive real systems based on mixed Gramian balanced truncation with error bounds
- An online manifold learning approach for model reduction of dynamical systems
- The Random Feature Model for Input-Output Maps between Banach Spaces
- Modeling synchronization in globally coupled oscillatory systems using model order reduction
- Projection-based model reduction with dynamically transformed modes
- Optimization-Based Modal Decomposition for Systems with Multiple Transports
- A tailored convolutional neural network for nonlinear manifold learning of computational physics data using unstructured spatial discretizations
- Space-time registration-based model reduction of parameterized one-dimensional hyperbolic PDEs
- Fourier collocation and reduced basis methods for fast modeling of compressible flows
- Fast \(L^2\) optimal mass transport via reduced basis methods for the Monge-Ampère equation
- A deep learning approach to Reduced Order Modelling of parameter dependent partial differential equations
- Rank-adaptive structure-preserving model order reduction of Hamiltonian systems
- Modern Koopman theory for dynamical systems
- Optimizing oblique projections for nonlinear systems using trajectories
- Piecewise-global nonlinear model order reduction for PDE-constrained optimization in high-dimensional parameter spaces
- Deep neural network surrogates for nonsmooth quantities of interest in shape uncertainty quantification
- A Registration Method for Model Order Reduction: Data Compression and Geometry Reduction
- Leveraging reduced-order models for state estimation using deep learning
- On the structure of time-delay embedding in linear models of non-linear dynamical systems
- Model order reduction of combustion processes with complex front dynamics
- Multilevel Fine-Tuning: Closing Generalization Gaps in Approximation of Solution Maps under a Limited Budget for Training Data
- On the role of nonlinear correlations in reduced-order modelling
- From snapshots to manifolds – a tale of shear flows
- Reinforcement-learning-based actuator selection method for active flow control
- Manifold Approximations via Transported Subspaces: Model Reduction for Transport-Dominated Problems
- Symplectic Model Reduction of Hamiltonian Systems on Nonlinear Manifolds and Approximation with Weakly Symplectic Autoencoder
- Reduced basis methods for time-dependent problems
- DRIPS: a framework for dimension reduction and interpolation in parameter space
- A two-stage deep learning architecture for model reduction of parametric time-dependent problems
- Neural-network-augmented projection-based model order reduction for mitigating the Kolmogorov barrier to reducibility
- Model Reduction for Nonlinear Systems by Balanced Truncation of State and Gradient Covariance
- Nonlinear reduced-order modeling for three-dimensional turbulent flow by large-scale machine learning
- A non‐linear non‐intrusive reduced order model of fluid flow by auto‐encoder and self‐attention deep learning methods
- Canonical and noncanonical Hamiltonian operator inference
- Hybrid analysis and modeling, eclecticism, and multifidelity computing toward digital twin revolution
- A neural network‐enhanced reproducing kernel particle method for modeling strain localization
- Deep learning‐based reduced order models for the real‐time simulation of the nonlinear dynamics of microstructures
- Preconditioned least‐squares Petrov–Galerkin reduced order models
- Optimization problems for PDEs in weak space-time form. Abstracts from the workshop held March 5--10, 2023
This page was built for publication: Model reduction of dynamical systems on nonlinear manifolds using deep convolutional autoencoders
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2223001)
