Dimension Free Nonasymptotic Bounds on the Accuracy of High-Dimensional Laplace Approximation
From MaRDI portal
Publication:6062239
Abstract: This note attempts to revisit the classical results on Laplace approximation in a modern non-asymptotic and dimension free form. Such an extension is motivated by applications to high dimensional statistical and optimization problems. The established results provide explicit non-asymptotic bounds on the quality of a Gaussian approximation of the posterior distribution in total variation distance in terms of the so called emph{effective dimension} ( p_G ). This value is defined as interplay between information contained in the data and in the prior distribution. In the contrary to prominent Bernstein - von Mises results, the impact of the prior is not negligible and it allows to keep the effective dimension small or moderate even if the true parameter dimension is huge or infinite. We also address the issue of using a Gaussian approximation with inexact parameters with the focus on replacing the Maximum a Posteriori (MAP) value by the posterior mean and design the algorithm of Bayesian optimization based on Laplace iterations. The results are specified to the case of nonlinear inverse problem.
Recommendations
- scientific article; zbMATH DE number 795286
- Simple upper and lower bounds for the multivariate Laplace approximation
- \(N\)-widths and \(\varepsilon \)-dimensions for high-dimensional approximations
- Multivariate Laplace's approximation with estimated error and application to limit theorems
- New error bounds for Laplace approximationviaStein’s method
- Non-intrusive low-rank separated approximation of high-dimensional stochastic models
- Error Bounds for Lanczos-Based Matrix Function Approximation
- A least-squares approximation of partial differential equations with high-dimensional random inputs
- Precise Error Analysis of Regularized <inline-formula> <tex-math notation="LaTeX">$M$ </tex-math> </inline-formula>-Estimators in High Dimensions
Cites work
- Analysis of the ensemble Kalman filter for inverse problems
- Asymptotic approximations of integrals
- Asymptotic methods in statistical decision theory
- Asymptotic Statistics
- Bernstein-von Mises theorems and uncertainty quantification for linear inverse problems
- Bernstein-von Mises theorems for statistical inverse problems. I: Schrödinger equation
- Efficient nonparametric Bayesian inference for X-ray transforms
- Error bounds for multidimensional Laplace approximation
- Error bounds for the Laplace approximation for definite integrals
- High-dimensional Bayesian inference via the unadjusted Langevin algorithm
- scientific article; zbMATH DE number 3474264 (Why is no real title available?)
- Inverse problems: a Bayesian perspective
- Large ball probabilities, Gaussian comparison and anti-concentration
- Multivariate Laplace's approximation with estimated error and application to limit theorems
- Non-asymptotic error estimates for the Laplace approximation in Bayesian inverse problems
- On the convergence of the Laplace approximation and noise-level-robustness of Laplace-based Monte Carlo methods for Bayesian inverse problems
- Pattern recognition and machine learning.
- Random gradient-free minimization of convex functions
- Simple upper and lower bounds for the multivariate Laplace approximation
- The Bernstein-von Mises theorem and nonregular models
- Theoretical Guarantees for Approximate Sampling from Smooth and Log-Concave Densities
Cited in
(6)- Optimal experimental design: formulations and computations
- On the approximation accuracy of Gaussian variational inference
- Statistics and learning theory in the era of artificial intelligence. Abstracts from the workshop held June 23--28, 2024
- Improved dimension dependence in the Bernstein-von Mises theorem via a new Laplace approximation bound
- Laplace and saddlepoint approximations in high dimensions
- Scalability of the second-order reliability method for stochastic differential equations with multiplicative noise
This page was built for publication: Dimension Free Nonasymptotic Bounds on the Accuracy of High-Dimensional Laplace Approximation
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6062239)
