On the convergence of the SINDy algorithm
From MaRDI portal
Abstract: One way to understand time-series data is to identify the underlying dynamical system which generates it. This task can be done by selecting an appropriate model and a set of parameters which best fits the dynamics while providing the simplest representation (i.e. the smallest amount of terms). One such approach is the sparse identification of nonlinear dynamics framework [6] which uses a sparsity-promoting algorithm that iterates between a partial least-squares fit and a thresholding (sparsity-promoting) step. In this work, we provide some theoretical results on the behavior and convergence of the algorithm proposed in [6]. In particular, we prove that the algorithm approximates local minimizers of an unconstrained -penalized least-squares problem. From this, we provide sufficient conditions for general convergence, rate of convergence, and conditions for one-step recovery. Examples illustrate that the rates of convergence are sharp. In addition, our results extend to other algorithms related to the algorithm in [6], and provide theoretical verification to several observed phenomena.
Recommendations
- SINDy-PI: a robust algorithm for parallel implicit sparse identification of nonlinear dynamics
- Sparse identification of nonlinear dynamical systems via reweighted \(\ell_1\)-regularized least squares
- Discovering governing equations from data by sparse identification of nonlinear dynamical systems
- Weak SINDy: Galerkin-based data-driven model selection
- Simultaneous Identification and Denoising of Dynamical Systems
Cites work
- scientific article; zbMATH DE number 845714 (Why is no real title available?)
- Automated reverse engineering of nonlinear dynamical systems
- Compressed Sensing With Cross Validation
- Constrained sparse Galerkin regression
- Discovering governing equations from data by sparse identification of nonlinear dynamical systems
- Extracting Sparse High-Dimensional Dynamics from Limited Data
- Hidden physics models: machine learning of nonlinear partial differential equations
- Iterative hard thresholding for compressed sensing
- Iterative thresholding for sparse approximations
- Just relax: convex programming methods for identifying sparse signals in noise
- Learning partial differential equations via data discovery and sparse optimization
- Model selection for dynamical systems via sparse regression and information criteria
- Model selection for hybrid dynamical systems via sparse regression
- Sparse identification of nonlinear dynamics for model predictive control in the low-data limit
Cited in
(47)- Recovery guarantees for polynomial coefficients from weakly dependent data with outliers
- Enhancing sparse identification of nonlinear dynamics with Earth-mover distance and group similarity
- Learning dynamical systems from noisy data with Weak-form Kernel Ridge Regression
- A Review of Data‐Driven Discovery for Dynamic Systems
- Adaptive integral alternating minimization method for robust learning of nonlinear dynamical systems from highly corrupted data
- Regularized least absolute deviation-based sparse identification of dynamical systems
- Weak SINDy for partial differential equations
- On sparse regression, \(L_p\)-regularization, and automated model discovery
- An efficient first-order conditional gradient algorithm in data-driven sparse identification of nonlinear dynamics to solve sparse recovery problems under noise
- Structured model selection via ℓ1−ℓ2 optimization
- Data-driven sparse identification of nonlinear dynamical systems using linear multistep methods
- A dynamical system-based framework for dimension reduction
- Sparsity-promoting algorithms for the discovery of informative Koopman-invariant subspaces
- Bayesian autoencoders for data-driven discovery of coordinates, governing equations and fundamental constants
- SubTSBR to tackle high noise and outliers for data-driven discovery of differential equations
- On higher order drift and diffusion estimates for stochastic SINDy
- Some practical regards on the application of the physics-informed sparse identification for strongly NESs
- The \(l_1\)-based sparsification of energy interactions in unsteady lid-driven cavity flow
- Sparse identification of nonlinear dynamical systems via reweighted \(\ell_1\)-regularized least squares
- Dynamical system identification, model selection, and model uncertainty quantification by Bayesian inference
- Model selection via reweighted partial sparse recovery
- MODNO: multi-operator learning with distributed neural operators
- Data-driven delay identification with SINDy
- Asymptotic consistency of the WSINDy algorithm in the limit of continuum data
- The rate of convergence of Sinkhorn balancing
- Sparse identification of nonlinear dynamics with low-dimensionalized flow representations
- Identifying Nonlinear Dynamics with High Confidence from Sparse Data
- Sparse dynamical system identification with simultaneous structural parameters and initial condition estimation
- A Nonautonomous Equation Discovery Method for Time Signal Classification
- Sparse methods for automatic relevance determination
- Identifying stochastic dynamics via finite expression methods
- Group projected subspace pursuit for identification of variable coefficient differential equations (GP-IDENT)
- SINDy-PI: a robust algorithm for parallel implicit sparse identification of nonlinear dynamics
- Sequential thresholded quantile estimator for sparse regression
- Poincaré maps for multiscale physics discovery and nonlinear Floquet theory
- Convergence of weak-SINDy surrogate models
- Data discovery of lower dimensional equations of turbulent flows
- Data driven discovery of systems of ordinary differential equations using nonconvex multitask learning
- Reconstruction of delay differential equations via learning parameterized dictionary
- Sparse regression for low-dimensional time-dynamic varying coefficient models with application to air quality data
- Sparse identification of nonlinear economic dynamical model
- Convergence of a threshold-type algorithm using the signed distance function
- Statistical machine learning tools for probabilistic closures of turbulence models
- WeakIdent: weak formulation for identifying differential equation using narrow-fit and trimming
- Learning mean-field equations from particle data using WSINDy
- Sparse identification of dynamical systems by reweighted \(l_1\)-regularized least absolute deviation regression
- Implications of convergence rates in Sinkhorn balancing
This page was built for publication: On the convergence of the SINDy algorithm
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5197636)
