Gaussian Process Subspace Prediction for Model Reduction
From MaRDI portal
Abstract: Subspace-valued functions arise in a wide range of problems, including parametric reduced order modeling (PROM). In PROM, each parameter point can be associated with a subspace, which is used for Petrov-Galerkin projections of large system matrices. Previous efforts to approximate such functions use interpolations on manifolds, which can be inaccurate and slow. To tackle this, we propose a novel Bayesian nonparametric model for subspace prediction: the Gaussian Process Subspace regression (GPS) model. This method is extrinsic and intrinsic at the same time: with multivariate Gaussian distributions on the Euclidean space, it induces a joint probability model on the Grassmann manifold, the set of fixed-dimensional subspaces. The GPS adopts a simple yet general correlation structure, and a principled approach for model selection. Its predictive distribution admits an analytical form, which allows for efficient subspace prediction over the parameter space. For PROM, the GPS provides a probabilistic prediction at a new parameter point that retains the accuracy of local reduced models, at a computational complexity that does not depend on system dimension, and thus is suitable for online computation. We give four numerical examples to compare our method to subspace interpolation, as well as two methods that interpolate local reduced models. Overall, GPS is the most data efficient, more computationally efficient than subspace interpolation, and gives smooth predictions with uncertainty quantification.
Recommendations
- Gaussian Process Prediction using Design-Based Subsampling
- Gaussian processes with built-in dimensionality reduction: applications to high-dimensional uncertainty propagation
- Variable Selection for Gaussian Process Models using Experimental Design-Based Subagging
- Hilbert space methods for reduced-rank Gaussian process regression
- Switching and Learning in Feedback Systems
- Gaussian Process Regression on Nested Spaces
- Reducing subspace models for large‐scale covariance regression
- Prediction With Approximated Gaussian Process Dynamical Models
- scientific article; zbMATH DE number 6747921
Cites work
- $\mathcal{H}_2$ Model Reduction for Large-Scale Linear Dynamical Systems
- A real time procedure for affinely dependent parametric model order reduction using interpolation on Grassmann manifolds
- A survey of projection-based model reduction methods for parametric dynamical systems
- Active subspaces. Emerging ideas for dimension reduction in parameter studies
- An ``\(hp\) certified reduced basis method for parametrized elliptic partial differential equations
- An efficient surrogate model for emulation and physics extraction of large eddy simulations
- An observation-driven time-dependent basis for a reduced description of transient stochastic systems
- An Online Method for Interpolating Linear Parametric Reduced-Order Models
- Bayesian manifold regression
- Compressed Gaussian process for manifold regression
- Dynamic mode decomposition of numerical and experimental data
- Efficient emulators of computer experiments using compactly supported correlation functions, with an application to cosmology
- Extrinsic Gaussian processes for regression and classification on manifolds
- Finding structure with randomness: probabilistic algorithms for constructing approximate matrix decompositions
- Fréchet regression for random objects with Euclidean predictors
- Gaussian processes for machine learning.
- Geodesic finite elements of higher order
- Hermite interpolation and data processing errors on Riemannian matrix manifolds
- scientific article; zbMATH DE number 6159604 (Why is no real title available?)
- Interpolatory weighted-\(\mathcal{H}_2\) model reduction
- Intrinsic Gaussian processes on complex constrained domains
- Nonlinear model reduction via discrete empirical interpolation
- Normal-bundle bootstrap
- Principal component analysis in linear systems: Controllability, observability, and model reduction
- Quasi-interpolation in Riemannian manifolds
- Riemannian \(L^{p}\) center of mass: existence, uniqueness, and convexity
- Schubert varieties and distances between subspaces of different dimensions
- Statistics on special manifolds
- The design and analysis of computer experiments
- The Geometry of Algorithms with Orthogonality Constraints
- The Scaling and Squaring Method for the Matrix Exponential Revisited
Cited in
(15)- Reduced-load equivalence for Gaussian processes
- Variable Selection for Gaussian Process Models using Experimental Design-Based Subagging
- Geometric subspace updates with applications to online adaptive nonlinear model reduction
- Dimension Reduction for Gaussian Process Emulation: An Application to the Influence of Bathymetry on Tsunami Heights
- scientific article; zbMATH DE number 6747921 (Why is no real title available?)
- DRIPS: a framework for dimension reduction and interpolation in parameter space
- Sliced Gradient-Enhanced Kriging for High-Dimensional Function Approximation
- A Hierarchical Expected Improvement Method for Bayesian Optimization
- Statistical Learning for Nonlinear Dynamical Systems with Applications to Aircraft-UAV Collisions
- ProSpar-GP: Scalable Gaussian Process Modeling with Massive Nonstationary Datasets
- Additive Multi-Index Gaussian Process Modeling, with Application to Multi-Physics Surrogate Modeling of the Quark-Gluon Plasma
- Gaussian processes with built-in dimensionality reduction: applications to high-dimensional uncertainty propagation
- Stochastic subspace via probabilistic principal component analysis for characterizing model error
- Adapting projection-based reduced-order models using projected Gaussian process
- Hilbert space methods for reduced-rank Gaussian process regression
This page was built for publication: Gaussian Process Subspace Prediction for Model Reduction
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5864693)
