Koopman operator learning using invertible neural networks
From MaRDI portal
Abstract: In Koopman operator theory, a finite-dimensional nonlinear system is transformed into an infinite but linear system using a set of observable functions. However, manually selecting observable functions that span the invariant subspace of the Koopman operator based on prior knowledge is inefficient and challenging, particularly when little or no information is available about the underlying systems. Furthermore, current methodologies tend to disregard the importance of the invertibility of observable functions, which leads to inaccurate results. To address these challenges, we propose the so-called FlowDMD, a Flow-based Dynamic Mode Decomposition that utilizes the Coupling Flow Invertible Neural Network (CF-INN) framework. FlowDMD leverages the intrinsically invertible characteristics of the CF-INN to learn the invariant subspaces of the Koopman operator and accurately reconstruct state variables. Numerical experiments demonstrate the superior performance of our algorithm compared to state-of-the-art methodologies.
Recommendations
- Linearly recurrent autoencoder networks for learning dynamics
- Koopman dynamic-oriented deep learning for invariant subspace identification and full-state prediction of complex systems
- Generalizing dynamic mode decomposition: balancing accuracy and expressiveness in Koopman approximations
- Physics-Informed Probabilistic Learning of Linear Embeddings of Nonlinear Dynamics with Guaranteed Stability
- A study on data-driven identification and representation of nonlinear dynamical systems with a physics-integrated deep learning approach: Koopman operators and nonlinear normal modes
Cites work
- A data-driven approximation of the koopman operator: extending dynamic mode decomposition
- A kernel-based method for data-driven Koopman spectral analysis
- Automated solution of differential equations by the finite element method. The FEniCS book
- Deep hidden physics models: deep learning of nonlinear partial differential equations
- Deep learning nonlinear multiscale dynamic problems using Koopman operator
- Discovering governing equations from data by sparse identification of nonlinear dynamical systems
- Dynamic mode decomposition and its variants
- Ergodic theory, dynamic mode decomposition, and computation of spectral properties of the Koopman operator
- Extended dynamic mode decomposition with dictionary learning: A data-driven adaptive spectral decomposition of the Koopman operator
- Higher order dynamic mode decomposition
- scientific article; zbMATH DE number 6678650 (Why is no real title available?)
- Linearly recurrent autoencoder networks for learning dynamics
- Normalizing flows for probabilistic modeling and inference
- On dynamic mode decomposition: theory and applications
- Online Dynamic Mode Decomposition for Time-Varying Systems
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Physics-Informed Probabilistic Learning of Linear Embeddings of Nonlinear Dynamics with Guaranteed Stability
- Prediction accuracy of dynamic mode decomposition
- Randomized Dynamic Mode Decomposition
- Residual dynamic mode decomposition: robust and verified Koopmanism
- When and why PINNs fail to train: a neural tangent kernel perspective
Cited in
(7)- Solving nonlinear ordinary differential equations using the invariant manifolds and Koopman eigenfunctions
- Koopman dynamic-oriented deep learning for invariant subspace identification and full-state prediction of complex systems
- Approximation of translation invariant Koopman operators for coupled non-linear systems
- Latent diffeomorphic dynamic mode decomposition
- Sparse discovery of differential equations based on multi-fidelity Gaussian process
- Resolution invariant deep operator network for PDEs with complex geometries
- Extraction of nonlinearity in neural networks with Koopman operator
This page was built for publication: Koopman operator learning using invertible neural networks
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6126575)
