We investigate the use of learning approaches to handle Bayesian inverse problems in a computationally efficient way when signals be inverted present moderately high number dimensions and are large number. propose tractable regression approach which has advantage produce full probability distributions as approximations target posterior distributions. In addition provide confidence indices on pr...

We propose a novel adaptive importance sampling scheme for Bayesian inversion problems where the inference of variables interest and power data noise are carried out using distinct (but interacting) methods. More specifically, we consider analysis (i.e., parameters model to invert), whereas employ maximum likelihood approach estimation power. The whole technique is implemented by means an itera...

In recent years several approaches have been proposed for global optimization in matched-eld inversion. Understanding the mechanics of these optimization methods is important in order to construct fast inversion schemes suitable for underwater propagation problems. It is the purpose of this paper to investigate the comparative performances of two global optimization techniques, genetic algorith...

In this paper we discuss general regularization estimators. This class includes Tikhonov type and spectral cut-off estimators as well as iterative methods, such as ν-methods and the Landweber iteration. The latter estimators achieve the same (optimal) convergence rates as spectral cut-off, but do not require explicit spectral information on the operator and are often much faster to compute than...

We present two new techniques for the inversion of first-arrival times to estimate velocity structure. These travel-time inversion techniques are unique in that they do not require the calculation of ray paths. First-arrival times are calculated using a finite-difference scheme that iteratively solves the eikonal equations for the position of the wavefront. The first inversion technique is a di...

In this paper, modified solutions were compared through utilizing three different approximate methods for bar structures. The modifications considered various changes in the initial design. To authors' best of knowledge, the studies have carried out on this matter so far are not broad enough and have considerred the simeltaneous variations of size, geometry and topology on the bar structures. I...

We develop a new adaptive inversion procedure: Dataadaptive Resolution Inversion (DRI) method, which eliminates the need of selecting a parameterization prior to inversion. Instead, one performs a hierarchical search for the correct parameterization while solving a sequence of inverse problems with an increasing dimension of parameterization. A parsimonious approach to inverse problems usually ...