Structured Bayesian Gaussian process latent variable model: applications to data-driven dimensionality reduction and high-dimensional inversion

Abstract

We introduce a methodology for nonlinear inverse problems using a variational Bayesian approach where the unknown quantity is a spatial field. A structured Bayesian Gaussian process latent variable model is used both to construct a low-dimensional generative model of the sample-based stochastic prior as well as a surrogate for the forward evaluation. Its Bayesian formulation captures epistemic uncertainty introduced by the limited number of input and output examples, automatically selects an appropriate dimensionality for the learned latent representation of the data, and rigorously propagates the uncertainty of the data-driven dimensionality reduction of the stochastic space through the forward model surrogate. The structured Gaussian process model explicitly leverages spatial information for an informative generative prior to improve sample efficiency while achieving computational tractability through Kronecker product decompositions of the relevant kernel matrices. Importantly, the Bayesian inversion is carried out by solving a variational optimization problem, replacing traditional computationally-expensive Monte Carlo sampling. The methodology is demonstrated on an elliptic PDE and is shown to return well-calibrated posteriors and is tractable with latent spaces with over 100 dimensions.

References in corpus (7)

• Weight Uncertainty in Neural Networks
• Deep UQ: Learning deep neural network surrogate models for high dimensional uncertainty quantification
• Deep Gaussian Processes
• Kernel Interpolation for Scalable Structured Gaussian Processes (KISS-GP)
• Data-Driven Model Reduction for the Bayesian Solution of Inverse Problems
• Gaussian Process Models with Parallelization and GPU acceleration
• Structured Bayesian Gaussian process latent variable model

Cited by in corpus (15)

• Surrogate Modeling for Fluid Flows Based on Physics-Constrained Deep Learning Without Simulation Data
• Physics-Constrained Deep Learning for High-dimensional Surrogate Modeling and Uncertainty Quantification without Labeled Data
• A Physics Informed Neural Network for Time-Dependent Nonlinear and Higher Order Partial Differential Equations
• Transformers for Modeling Physical Systems
• Calibrate, Emulate, Sample
• Attention U-Net as a surrogate model for groundwater prediction
• Surrogate assisted active subspace and active subspace assisted surrogate -- A new paradigm for high dimensional structural reliability analysis
• A physics-aware, probabilistic machine learning framework for coarse-graining high-dimensional systems in the Small Data regime
• Data-driven discovery of free-form governing differential equations
• Bayesian multiscale deep generative model for the solution of high-dimensional inverse problems
• A Bi-fidelity Surrogate Modeling Approach for Uncertainty Propagation in Three-Dimensional Hemodynamic Simulations
• A Survey of Bayesian Statistical Approaches for Big Data
• Bayesian learning of orthogonal embeddings for multi-fidelity Gaussian Processes
• Bayesian Modelling of Pattern Formation from One Snapshot of Pattern
• Bayesian task embedding for few-shot Bayesian optimization