Structured model selection via ℓ1−ℓ2 optimization
From MaRDI portal
Publication:6141559
Abstract: Automated model selection is an important application in science and engineering. In this work, we develop a learning approach for identifying structured dynamical systems from undersampled and noisy spatiotemporal data. The learning is performed by a sparse least-squares fitting over a large set of candidate functions via a nonconvex sparse optimization solved by the alternating direction method of multipliers. Using a Bernstein-like inequality with a coherence condition, we show that if the set of candidate functions forms a structured random sampling matrix of a bounded orthogonal system, the recovery is stable and the error is bounded. The learning approach is validated on synthetic data generated by the viscous Burgers' equation and two reaction-diffusion equations. The computational results demonstrate the theoretical guarantees of success and the efficiency with respect to the ambient dimension and the number of candidate functions.
Recommendations
- Extracting Structured Dynamical Systems Using Sparse Optimization With Very Few Samples
- Sparse learning of partial differential equations with structured dictionary matrix
- Extracting Sparse High-Dimensional Dynamics from Limited Data
- Sparse identification of nonlinear dynamical systems via reweighted \(\ell_1\)-regularized least squares
- Robust and optimal sparse regression for nonlinear PDE models
Cites work
- scientific article; zbMATH DE number 3605240 (Why is no real title available?)
- scientific article; zbMATH DE number 845714 (Why is no real title available?)
- A Probabilistic and RIPless Theory of Compressed Sensing
- A dual algorithm for the solution of nonlinear variational problems via finite element approximation
- A mathematical introduction to compressive sensing
- A method for finding structured sparse solutions to nonnegative least squares problems with applications
- Adjustment of an Inverse Matrix Corresponding to a Change in One Element of a Given Matrix
- Automated reverse engineering of nonlinear dynamical systems
- Computing sparse representation in a highly coherent dictionary based on difference of L₁ and L₂
- Data-driven identification of parametric partial differential equations
- Discovering governing equations from data by sparse identification of nonlinear dynamical systems
- Distributed optimization and statistical learning via the alternating direction method of multipliers
- Enhancing sparsity by reweighted \(\ell _{1}\) minimization
- Extracting Sparse High-Dimensional Dynamics from Limited Data
- Extracting Structured Dynamical Systems Using Sparse Optimization With Very Few Samples
- Fast L1-L2 minimization via a proximal operator
- High-Dimensional Probability: An Introduction with Applications in Data Science
- Large deviations for sums of partly dependent random variables
- Learning partial differential equations via data discovery and sparse optimization
- Minimization of \(\ell_{1-2}\) for compressed sensing
- New restricted isometry property analysis for \(\ell_1-\ell_2\) minimization methods
- On the convergence of the SINDy algorithm
- Proximal splitting methods in signal processing
- SINDy-PI: a robust algorithm for parallel implicit sparse identification of nonlinear dynamics
- Sparse identification of nonlinear dynamics for model predictive control in the low-data limit
- Sparse learning of partial differential equations with structured dictionary matrix
- Splitting Algorithms for the Sum of Two Nonlinear Operators
- The DC (Difference of convex functions) programming and DCA revisited with DC models of real world nonconvex optimization problems
- Weak SINDy for partial differential equations
- Weak SINDy: Galerkin-based data-driven model selection
Cited in
(3)
This page was built for publication: Structured model selection via ℓ1−ℓ2 optimization
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6141559)
