A near-stationary subspace for ridge approximation
From MaRDI portal
Publication:2310068
Abstract: Response surfaces are common surrogates for expensive computer simulations in engineering analysis. However, the cost of fitting an accurate response surface increases exponentially as the number of model inputs increases, which leaves response surface construction intractable for high-dimensional, nonlinear models. We describe ridge approximation for fitting response surfaces in several variables. A ridge function is constant along several directions in its domain, so fitting occurs on the coordinates of a low-dimensional subspace of the input space. We review essential theory for ridge approximation---e.g., the best mean-squared approximation and an optimal low-dimensional subspace---and we prove that the gradient-based active subspace is near-stationary for the least-squares problem that defines an optimal subspace. Motivated by the theory, we propose a computational heuristic that uses an estimated active subspace as an initial guess for a ridge approximation fitting problem. We show a simple example where the heuristic fails, which reveals a type of function for which the proposed approach is inappropriate. We then propose a simple alternating heuristic for fitting a ridge function, and we demonstrate the effectiveness of the active subspace initial guess applied to an airfoil model of drag as a function of its 18 shape parameters.
Recommendations
- Data-driven polynomial ridge approximation using variable projection
- Dimension reduction via Gaussian ridge functions
- Active subspace methods in theory and practice: applications to kriging surfaces
- Gradient free active subspace construction using Morris screening elementary effects
- Embedded ridge approximations
Cites work
- A Comparison of Three Methods for Selecting Values of Input Variables in the Analysis of Output from a Computer Code
- A note on the Poincaré inequality for convex domains
- A taxonomy of global optimization methods based on response surfaces
- Active subspace methods in theory and practice: applications to kriging surfaces
- Active subspaces. Emerging ideas for dimension reduction in parameter studies
- Algorithms for Separable Nonlinear Least Squares Problems
- An inequality for the multivariate normal distribution
- Applied Linear Regression
- Capturing ridge functions in high dimensions from point queries
- Design and analysis of computer experiments. With comments and a rejoinder by the authors
- Error and Perturbation Bounds for Subspaces Associated with Certain Eigenvalue Problems
- Exploiting active subspaces to quantify uncertainty in the numerical simulation of the hyshot II scramjet
- Exploring Regression Structure Using Nonparametric Functional Estimation
- Gaussian processes for machine learning.
- Gaussian processes with built-in dimensionality reduction: applications to high-dimensional uncertainty propagation
- Global sensitivity analysis: The primer
- scientific article; zbMATH DE number 3168214 (Why is no real title available?)
- scientific article; zbMATH DE number 1220060 (Why is no real title available?)
- scientific article; zbMATH DE number 1246413 (Why is no real title available?)
- scientific article; zbMATH DE number 6159604 (Why is no real title available?)
- scientific article; zbMATH DE number 1456803 (Why is no real title available?)
- scientific article; zbMATH DE number 5060482 (Why is no real title available?)
- Introduction to uncertainty quantification
- Learning functions of few arbitrary linear parameters in high dimensions
- Learning non-parametric basis independent models from point queries via low-rank methods
- Nested Latin hypercube designs
- On Nonlinear Functions of Linear Combinations
- Orthogonal polynomials of several variables
- Predicting the output from a complex computer code when fast approximations are available
- Projection pursuit
- Pymanopt: a Python toolbox for optimization on manifolds using automatic differentiation
- Ridge functions
- Survey of modeling and optimization strategies to solve high-dimensional design problems with computationally-expensive black-box functions
- The design and analysis of computer experiments.
- The Elements of Statistical Learning
- The Geometry of Algorithms with Orthogonality Constraints
- Tractability and strong tractability of linear multivariate problems
- Weighted Poincaré and Friedrichs Inequalities
Cited in
(17)- Embedded ridge approximations
- Sparse mixture models inspired by ANOVA decompositions
- Generalization bounds for sparse random feature expansions
- Linear/ridge expansions: enhancing linear approximations by ridge functions
- Generalized bounds for active subspaces
- Inverse regression for ridge recovery: a data-driven approach for parameter reduction in computer experiments
- The matrix ridge approximation: algorithms and applications
- Modified Active Subspaces Using the Average of Gradients
- Dimension reduction via Gaussian ridge functions
- A Lipschitz Matrix for Parameter Reduction in Computational Science
- Interpretable approximation of high-dimensional data
- Sequential Learning of Active Subspaces
- Gaussian quadrature and polynomial approximation for one-dimensional ridge functions
- Data-driven polynomial ridge approximation using variable projection
- Nonlinear level set learning for function approximation on sparse data with applications to parametric differential equations
- Applications of kernel ridge estimation to the problem of computing the aerodynamical characteristics of a passenger plane (in comparison with results obtained with artificial neural networks)
- A Riemannian View on Active Subspaces
This page was built for publication: A near-stationary subspace for ridge approximation
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2310068)
