The discovery of dynamics via linear multistep methods and deep learning: error estimation

DOI10.1137/21M140691XzbMATH Open1506.65105arXiv2103.11488MaRDI QIDQ5096451FDOQ5096451

Authors: Qiang Du, Yiqi Gu, Haizhao Yang, Chao Zhou 

Publication date: 17 August 2022

Published in: SIAM Journal on Numerical Analysis (Search for Journal in Brave)

Abstract: Identifying hidden dynamics from observed data is a significant and challenging task in a wide range of applications. Recently, the combination of linear multistep methods (LMMs) and deep learning has been successfully employed to discover dynamics, whereas a complete convergence analysis of this approach is still under development. In this work, we consider the deep network-based LMMs for the discovery of dynamics. We put forward error estimates for these methods using the approximation property of deep networks. It indicates, for certain families of LMMs, that the $el{l}^2$ grid error is bounded by the sum of $O(h^p)$ and the network approximation error, where $h$ is the time step size and $p$ is the local truncation error order. Numerical results of several physically relevant examples are provided to demonstrate our theory.

Full work available at URL: https://arxiv.org/abs/2103.11488

Recommendations

• Discovery of dynamics using linear multistep methods
• Data-driven sparse identification of nonlinear dynamical systems using linear multistep methods
• PDE-Net 2.0: learning PDEs from data with a numeric-symbolic hybrid deep network
• Deep hidden physics models: deep learning of nonlinear partial differential equations
• Error estimates for deep learning methods in fluid dynamics

zbMATH Keywords

deep learningconvergence analysislinear multistep methodsLorenz systemdata-driven modelingdiscovery of dynamics

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Multistep, Runge-Kutta and extrapolation methods for ordinary differential equations (65L06) Stability and convergence of numerical methods for ordinary differential equations (65L20) Numerical solution of inverse problems involving ordinary differential equations (65L09)

Cites Work

• PDE-Net 2.0: learning PDEs from data with a numeric-symbolic hybrid deep network
• Discovering governing equations from data by sparse identification of nonlinear dynamical systems
• Automated reverse engineering of nonlinear dynamical systems
• Title not available (Why is that?)
• Volume estimate of submanifolds in compact Riemannian manifolds
• Random projections of smooth manifolds
• An Introduction to Numerical Analysis
• Title not available (Why is that?)
• A special stability problem for linear multistep methods
• Convergence and stability in the numerical integration of ordinary differential equations
• The conditioning of Toeplitz band matrices
• Numerical solution of ordinary differential equations.
• Dynamic systems identification with Gaussian processes
• Volumes of compact manifolds
• Inferring solutions of differential equations using noisy multi-fidelity data
• Hidden physics models: machine learning of nonlinear partial differential equations
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Machine learning of linear differential equations using Gaussian processes
• Numerical Gaussian processes for time-dependent and nonlinear partial differential equations
• Deep hidden physics models: deep learning of nonlinear partial differential equations
• A mean field view of the landscape of two-layer neural networks
• Deep network with approximation error being reciprocal of width to power of square root of depth
• Numerical aspects for approximating governing equations using data
• Adaptive backstepping control, synchronization and circuit simulation of a 3-D novel jerk chaotic system with two hyperbolic sinusoidal nonlinearities
• Nonparametric inference of interaction laws in systems of agents from trajectory data
• IDENT: identifying differential equations with numerical time evolution
• Robust data-driven discovery of governing physical laws with error bars
• Error bounds for deep ReLU networks using the Kolmogorov-Arnold superposition theorem
• Data-driven deep learning of partial differential equations in modal space
• Data driven governing equations approximation using deep neural networks
• Discovery of dynamics using linear multistep methods
• Deep Network Approximation for Smooth Functions
• Deep ReLU Networks Overcome the Curse of Dimensionality for Generalized Bandlimited Functions
• Deep network approximation characterized by number of neurons
• Machine learning for prediction with missing dynamics
• Deep learning of dynamics and signal-noise decomposition with time-stepping constraints
• The Barron space and the flow-induced function spaces for neural network models
• Data-driven discovery of emergent behaviors in collective dynamics
• SubTSBR to tackle high noise and outliers for data-driven discovery of differential equations

Cited In (12)

• Towards the ultimate understanding of MUSCL: Pitfalls in achieving third-order accuracy
• Data-driven sparse identification of nonlinear dynamical systems using linear multistep methods
• Quadrature rule based discovery of dynamics by data-driven denoising
• Error analysis based on inverse modified differential equations for discovery of dynamics using linear multistep methods and deep learning
• Time-series machine-learning error models for approximate solutions to parameterized dynamical systems
• A framework for machine learning of model error in dynamical systems
• On the identifiability of nonlocal interaction kernels in first-order systems of interacting particles on Riemannian manifolds
• Implementation and (inverse modified) error analysis for implicitly templated ODE-nets
• Discovery of dynamics using linear multistep methods
• Solving PDEs on unknown manifolds with machine learning
• Bridging the gap: machine learning to resolve improperly modeled dynamics
• Structure-preserving recurrent neural networks for a class of Birkhoffian systems

This page was built for publication: The discovery of dynamics via linear multistep methods and deep learning: error estimation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5096451)