Stability of Gibbs Posteriors from the Wasserstein Loss for Bayesian Full Waveform Inversion
arXiv:2004.03730
Abstract
Recently, the Wasserstein loss function has been proven to be effective when applied to deterministic full-waveform inversion (FWI) problems. We consider the application of this loss function in Bayesian FWI so that the uncertainty can be captured in the solution. Other loss functions that are commonly used in practice are also considered for comparison. Existence and stability of the resulting Gibbs posteriors are shown on function space under weak assumptions on the prior and model. In particular, the distribution arising from the Wasserstein loss is shown to be quite stable with respect to high-frequency noise in the data. We then illustrate the difference between the resulting distributions numerically, using Laplace approximations to estimate the unknown velocity field and uncertainty associated with the estimates.
arXiv comment32 pages, 8 figures
References in corpus (9)
- Bayesian Inference with Optimal Maps
- Scalable and efficient algorithms for the propagation of uncertainty from data through inference to prediction for large-scale problems, with application to flow of the Antarctic ice sheet
- Geometric MCMC for Infinite-Dimensional Inverse Problems
- On the well-posedness of Bayesian inverse problems
- Bayesian inversion in resin transfer molding
- Well-posed Bayesian inverse problems and heavy-tailed stable quasi-Banach space priors
- On parameter estimation with the Wasserstein distance
- Multiplicative noise in Bayesian inverse problems: Well-posedness and consistency of MAP estimators
- Analysis and Application of Optimal Transport For Challenging Seismic Inverse Problems