Deep importance sampling based on regression for model inversion and emulation

e-Archivo Repository

Show simple item record Martino, Luca Delgado Gómez, David Llorente Fernández, Fernando Camps-Valls, Gustavo 2022-02-21T19:50:41Z 2022-02-21T19:50:41Z 2021-09-01
dc.identifier.bibliographicCitation Llorente, F., Martino, L., Delgado-Gómez, D., & Camps-Valls, G. (2021). Deep importance sampling based on regression for model inversion and emulation. En Digital Signal Processing, 116, p. 103104
dc.identifier.issn 1051-2004
dc.description.abstract Understanding systems by forward and inverse modeling is a recurrent topic of research in many domains of science and engineering. In this context, Monte Carlo methods have been widely used as powerful tools for numerical inference and optimization. They require the choice of a suitable proposal density that is crucial for their performance. For this reason, several adaptive importance sampling (AIS) schemes have been proposed in the literature. We here present an AIS framework called Regression-based Adaptive Deep Importance Sampling (RADIS). In RADIS, the key idea is the adaptive construction via regression of a non-parametric proposal density (i.e., an emulator), which mimics the posterior distribution and hence minimizes the mismatch between proposal and target densities. RADIS is based on a deep architecture of two (or more) nested IS schemes, in order to draw samples from the constructed emulator. The algorithm is highly efficient since employs the posterior approximation as proposal density, which can be improved adding more support points. As a consequence, RADIS asymptotically converges to an exact sampler under mild conditions. Additionally, the emulator produced by RADIS can be in turn used as a cheap surrogate model for further studies. We introduce two specific RADIS implementations that use Gaussian Processes (GPs) and Nearest Neighbors (NN) for constructing the emulator. Several numerical experiments and comparisons show the benefits of the proposed schemes. A real-world application in remote sensing model inversion and emulation confirms the validity of the approach.
dc.description.sponsorship This work has been supported by Spanish government via grant FPU19/00815, by Agencia Estatal de Investigación AEI (project SPGRAPH, ref. num. PID2019-105032GB-I00), by the Found action by the Community of Madrid in the framework of the Multiannual Agreement with the Rey Juan Carlos University in line of action 1, “Encouragement of Young Phd students investigation”, Project Ref. F661 Acronym Mapping-UCI, and by the European Research Council (ERC) under the ERC Consolidator Grant 2014 project SEDAL (647423).
dc.language.iso eng
dc.publisher Elsevier
dc.rights © 2021 The Authors.
dc.rights Atribución-NoComercial-SinDerivadas 3.0 España
dc.subject.other Adaptive regression
dc.subject.other Bayesian inference
dc.subject.other Emulation
dc.subject.other Importance sampling
dc.subject.other Model inversion
dc.subject.other Remote sensing
dc.title Deep importance sampling based on regression for model inversion and emulation
dc.type article
dc.subject.eciencia Estadística
dc.rights.accessRights openAccess
dc.relation.projectID Gobierno de España. PID2019-105032GB-I00
dc.relation.projectID AT-2021
dc.type.version publishedVersion
dc.identifier.publicationfirstpage 1
dc.identifier.publicationlastpage 22
dc.identifier.publicationtitle DIGITAL SIGNAL PROCESSING
dc.identifier.publicationvolume 116
dc.identifier.uxxi AR/0000028894
dc.contributor.funder Agencia Estatal de Investigación (España)
dc.affiliation.dpto UC3M. Departamento de Estadística
 Find Full text

Files in this item

*Click on file's image for preview. (Embargoed files's preview is not supported)

The following license files are associated with this item:

This item appears in the following Collection(s)

Show simple item record