Data-driven discovery of coordinates and governing equations

DOI10.1073/PNAS.1906995116zbMATH Open1433.68396arXiv1904.02107OpenAlexW2981906852WikidataQ90856756 ScholiaQ90856756MaRDI QIDQ5218490FDOQ5218490

Authors: Kathleen Champion, Bethany Lusch, Nathan Kutz, S. L. Brunton 

Publication date: 4 March 2020

Published in: Proceedings of the National Academy of Sciences (Search for Journal in Brave)

Abstract: The discovery of governing equations from scientific data has the potential to transform data-rich fields that lack well-characterized quantitative descriptions. Advances in sparse regression are currently enabling the tractable identification of both the structure and parameters of a nonlinear dynamical system from data. The resulting models have the fewest terms necessary to describe the dynamics, balancing model complexity with descriptive ability, and thus promoting interpretability and generalizability. This provides an algorithmic approach to Occam's razor for model discovery. However, this approach fundamentally relies on an effective coordinate system in which the dynamics have a simple representation. In this work, we design a custom autoencoder to discover a coordinate transformation into a reduced space where the dynamics may be sparsely represented. Thus, we simultaneously learn the governing equations and the associated coordinate system. We demonstrate this approach on several example high-dimensional dynamical systems with low-dimensional behavior. The resulting modeling framework combines the strengths of deep neural networks for flexible representation and sparse identification of nonlinear dynamics (SINDy) for parsimonious models. It is the first method of its kind to place the discovery of coordinates and models on an equal footing.

Full work available at URL: https://arxiv.org/abs/1904.02107

Recommendations

• Discovering governing equations from data by sparse identification of nonlinear dynamical systems
• Weak SINDy: Galerkin-based data-driven model selection
• Data-driven discovery of PDEs in complex datasets
• Data-driven sparse identification of nonlinear dynamical systems using linear multistep methods
• Deep hidden physics models: deep learning of nonlinear partial differential equations

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Computational aspects of data analysis and big data (68T09)

Cited In (only showing first 100 items - show all)

• Energy determines multiple stability in time-delayed systems
• Airfoil-based convolutional autoencoder and long short-term memory neural network for predicting coherent structures evolution around an airfoil
• Weak SINDy: Galerkin-based data-driven model selection
• Parsimony as the ultimate regularizer for physics-informed machine learning
• On the role of nonlinear correlations in reduced-order modelling
• Multiresolution convolutional autoencoders
• Spiral-wave dynamics in excitable media: insights from dynamic mode decomposition
• Data-driven sparse identification of nonlinear dynamical systems using linear multistep methods
• A nonlocal physics-informed deep learning framework using the peridynamic differential operator
• Multiscale model reduction for incompressible flows
• Latent-space time evolution of non-intrusive reduced-order models using Gaussian process emulation
• Incorporating physical constraints in a deep probabilistic machine learning framework for coarse-graining dynamical systems
• LaSDI: parametric latent space dynamics identification
• Learning nonlinear state-space models using autoencoders
• Implicit Regularization and Momentum Algorithms in Nonlinearly Parameterized Adaptive Control and Prediction
• Bayesian operator inference for data-driven reduced-order modeling
• Data-driven discovery of coarse-grained equations
• Reduced operator inference for nonlinear partial differential equations
• Time-series learning of latent-space dynamics for reduced-order model closure
• System identification through Lipschitz regularized deep neural networks
• Model reduction for multi-scale transport problems using model-form preserving least-squares projections with variable transformation
• Hierarchical deep learning neural network (HiDeNN): an artificial intelligence (AI) framework for computational science and engineering
• Low-dimensional models of single neurons: a review
• Discovering transition phenomena from data of stochastic dynamical systems with Lévy noise
• A framework for machine learning of model error in dynamical systems
• Discovery of slow variables in a class of multiscale stochastic systems via neural networks
• The structure of reconstructed flows in latent spaces
• Kernel-based parameter estimation of dynamical systems with unknown observation functions
• Data-driven discovery of governing equations for transient heat transfer analysis
• Physics-informed neural networks for data-driven simulation: advantages, limitations, and opportunities
• Sparse identification of nonlinear dynamics with low-dimensionalized flow representations
• Generative Adversarial Network for Probabilistic Forecast of Random Dynamical Systems
• Robust data-driven discovery of governing physical laws with error bars
• SINDy-PI: a robust algorithm for parallel implicit sparse identification of nonlinear dynamics
• Equation discovery from data: promise and pitfalls, from rabbits to Mars
• Output-weighted and relative entropy loss functions for deep learning precursors of extreme events
• Predicting shallow water dynamics using echo-state networks with transfer learning
• Deep learning of conjugate mappings
• Extracting non-Gaussian governing laws from data on mean exit time
• Deep learning‐based reduced order models for the real‐time simulation of the nonlinear dynamics of microstructures
• Deep learning models for global coordinate transformations that linearise PDEs
• Multi-fidelity surrogate modeling using long short-term memory networks
• Bayesian differential programming for robust systems identification under uncertainty
• gLaSDI: parametric physics-informed greedy latent space dynamics identification
• Reduced-order autodifferentiable ensemble Kalman filters
• DL-PDE: Deep-Learning Based Data-Driven Discovery of Partial Differential Equations from Discrete and Noisy Data
• Physics-constrained, data-driven discovery of coarse-grained dynamics
• Learning sparse nonlinear dynamics via mixed-integer optimization
• Adaptive learning of effective dynamics for online modeling of complex systems
• Mini-workshop: Analysis of data-driven optimal control. Abstracts from the mini-workshop held May 9--15, 2021 (hybrid meeting)
• Unsupervised discovery of interpretable hyperelastic constitutive laws
• Discovering governing equations from data by sparse identification of nonlinear dynamical systems
• Title not available (Why is that?)
• Towards out of distribution generalization for problems in mechanics
• Reconstruction of normal forms by learning informed observation geometries from data
• Data-driven discovery of multiscale chemical reactions governed by the law of mass action
• Data-Driven Discovery of Governing Equations for Coarse-Grained Heterogeneous Network Dynamics
• Adaptive integral alternating minimization method for robust learning of nonlinear dynamical systems from highly corrupted data
• Generalized quadratic embeddings for nonlinear dynamics using deep learning
• Calibrating multi-dimensional complex ODE from noisy data via deep neural networks
• Hausdorff metric based training of kernels to learn attractors with application to 133 chaotic dynamical systems
• Data-driven reduced-order modeling of spatiotemporal chaos with neural ordinary differential equations
• Regularized least absolute deviation-based sparse identification of dynamical systems
• Learning the temporal evolution of multivariate densities via normalizing flows
• Data-driven identification of dynamical models using adaptive parameter sets
• Constructing differential equations using only a scalar time-series about continuous time chaotic dynamics
• On sparse regression, \(L_p\)-regularization, and automated model discovery
• Koopman neural operator as a mesh-free solver of non-linear partial differential equations
• Identifying causality drivers and deriving governing equations of nonlinear complex systems
• Data augmentation-based statistical inference of diffusion processes
• Quantum Mechanics for Closure of Dynamical Systems
• Optimal reconstruction of vector fields from data for prediction and uncertainty quantification
• tLaSDI: thermodynamics-informed latent space dynamics identification
• Phase autoencoder for limit-cycle oscillators
• Deep learning in computational mechanics: a review
• Data-driven reduced order models using invariant foliations, manifolds and autoencoders
• Dynamical system identification, model selection, and model uncertainty quantification by Bayesian inference
• Discovering efficient periodic behaviors in mechanical systems via neural approximators
• EKF-SINDy: empowering the extended Kalman filter with sparse identification of nonlinear dynamics
• \(\Phi\)-DVAE: physics-informed dynamical variational autoencoders for unstructured data assimilation
• Machine learning in viscoelastic fluids via energy-based kernel embedding
• Discrepancy Modeling Framework: Learning Missing Physics, Modeling Systematic Residuals, and Disambiguating between Deterministic and Random Effects
• Machine learning methods for reduced order modeling
• Forecasting and predicting stochastic agent-based model data with biologically-informed neural networks
• Discovering interpretable Lagrangian of dynamical systems from data
• Approximation identification for the stochastic time-delayed dynamical system
• SDYN-GANs: adversarial learning methods for multistep generative models for general order stochastic dynamics
• Non parametric maximin aggregation for data with inhomogeneity
• Double diffusion maps and their latent harmonics for scientific computations in latent space
• Learning the flux and diffusion function for degenerate convection-diffusion equations using different types of observations
• Reduced order modeling of parametrized systems through autoencoders and SINDy approach: continuation of periodic solutions
• Data driven discovery of systems of ordinary differential equations using nonconvex multitask learning
• Testing topological conjugacy of time series
• Model free data assimilation with Takens embedding
• IGA-Graph-Net: isogeometric analysis-reuse method based on graph neural networks for topology-consistent models
• Reinforcement learning in non-Markovian environments
• GPLaSDI: Gaussian process-based interpretable latent space dynamics identification through deep autoencoder
• Kernel functions embed into the autoencoder to identify the sparse models of nonlinear dynamics
• A data-driven framework for learning hybrid dynamical systems
• Machine learning enhanced Hankel dynamic-mode decomposition

This page was built for publication: Data-driven discovery of coordinates and governing equations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5218490)