Abstract:Uncertainty propagation in complex engineering systems often poses significant computational challenges related to modeling and quantifying probability distributions of model outputs, as those emerge as the result of various sources of uncertainty that are inherent in the system under investigation. Gaussian Processes regression (GPs) is a robust meta-modeling technique that allows for fast model prediction and exploration of response surfaces. Multi-fidelity variations of GPs further leverage information from cheap and low fidelity model simulations in order to improve their predictive performance on the high fidelity model. In order to cope with the high volume of data required to train GPs in high dimensional design spaces, a common practice is to introduce latent design variables that are typically projections of the original input space to a lower dimensional subspace, and therefore substitute the problem of learning the initial high dimensional mapping, with that of training a GP on a low dimensional space. In this paper, we present a Bayesian approach to identify optimal transformations that map the input points to low dimensional latent variables. The “projection” mapping consists of an orthonormal matrix that is considered a priori unknown and needs to be inferred jointly with the GP parameters, conditioned on the available training data. The proposed Bayesian inference scheme relies on a two-step iterative algorithm that samples from the marginal posteriors of the GP parameters and the projection matrix respectively, both using Markov Chain Monte Carlo (MCMC) sampling. In order to take into account the orthogonality constraints imposed on the orthonormal projection matrix, a Geodesic Monte Carlo sampling algorithm is employed, that is suitable for exploiting probability measures on manifolds. We extend the proposed framework to multi-fidelity models using GPs including the scenarios of training multiple outputs together. We validate our framework on three synthetic problems with a known lower-dimensional subspace. The benefits of our proposed framework, are illustrated on the computationally challenging aerodynamic optimization of a last-stage blade for an industrial gas turbine, where we study the effect of an 85-dimensional shape parameterization of a three-dimensional airfoil on two output quantities of interest, specifically on the aerodynamic efficiency and the degree of reaction.

Intrinsic Gaussian Process on Unknown Manifolds with Probabilistic Metrics

Manifold Gaussian Processes for Regression

Residual Deep Gaussian Processes on Manifolds

Implicit Gaussian process representation of vector fields over arbitrary latent manifolds

Implicit Manifold Gaussian Process Regression

On Probabilistic Pullback Metrics on Latent Hyperbolic Manifolds

Intrinsic Gaussian Vector Fields on Manifolds

High-Dimensional Intrinsic Interpolation Using Gaussian Process Regression and Diffusion Maps

Generalized Moving Least-Squares for Solving Vector-valued PDEs on Unknown Manifolds

Thin and Deep Gaussian Processes

Vecchia Gaussian Processes: Probabilistic Properties, Minimax Rates and Methodological Developments

Physics Inspired Approaches To Understanding Gaussian Processes

Manifold Learning for the Emulation of Spatial Fields from Computational Models.

Gaussian Process on the Product of Directional Manifolds

Inference for Gaussian Processes with Matérn Covariogram on Compact Riemannian Manifolds

Mixture Probabilistic Principal Geodesic Analysis

Posterior Contraction Rates for Matérn Gaussian Processes on Riemannian Manifolds

Sparse Gaussian Processes with Manifold-Preserving Graph Reduction

Bayesian Learning of Orthogonal Embeddings for Multi-Fidelity Gaussian Processes

Latent Gaussian Process Regression