Abstract:PreviousNext No AccessSEG Technical Program Expanded Abstracts 2019Regularized supervised descent method for 2-D magnetotelluric data inversionAuthors: Rui GuoMaokun LiFan YangShenheng XuAria AbubakarRui GuoTsinghua UniversitySearch for more papers by this author, Maokun LiTsinghua UniversitySearch for more papers by this author, Fan YangTsinghua UniversitySearch for more papers by this author, Shenheng XuTsinghua UniversitySearch for more papers by this author, and Aria AbubakarSchlumbergerSearch for more papers by this authorhttps://doi.org/10.1190/segam2019-3215540.1 SectionsAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractWe study a 2-D inversion algorithm for magnetotelluric (MT) data using regularized supervised descent method (SDM). The inversion contains two stages: offline training and online prediction. In the training stage, the descent direction to minimize an objective function is learned from training samples. In the prediction stage, model reconstruction is achieved by solving a regularized prediction equation with the learned directions. This method enables us to incorporate various prior information into deterministic inversion, and also allows the physical modeling an opportunity to guide the machine learning prediction. Furthermore, the regularized SDM can accelerate the online inversion by avoiding the computation of the Jacobian matrix. Numerical examples show that by using the regularized SDM, the inversion converges rapidly and the reconstructed models can be more accurate.Presentation Date: Tuesday, September 17, 2019Session Start Time: 1:50 PMPresentation Time: 3:05 PMLocation: Poster Station 2Presentation Type: PosterKeywords: electromagnetics, magnetotelluric, inversion, 2D, frequency-domainPermalink: https://doi.org/10.1190/segam2019-3215540.1FiguresReferencesRelatedDetailsCited byPhysics-Embedded Machine Learning for Electromagnetic Data Imaging: Examining three types of data-driven imaging methodsIEEE Signal Processing Magazine, Vol. 40, No. 21-D Inversion of GREATEM Data by Supervised Descent LearningIEEE Geoscience and Remote Sensing Letters, Vol. 19 SEG Technical Program Expanded Abstracts 2019ISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2019 Pages: 5407 publication data© 2019 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 10 Aug 2019 CITATION INFORMATION Rui Guo, Maokun Li, Fan Yang, Shenheng Xu, and Aria Abubakar, (2019), "Regularized supervised descent method for 2-D magnetotelluric data inversion," SEG Technical Program Expanded Abstracts : 2508-2512. https://doi.org/10.1190/segam2019-3215540.1 Plain-Language Summary Keywordselectromagneticsmagnetotelluricinversion2Dfrequency-domainPDF DownloadLoading ...

Effects of different amplitudes of noise and missing data on regular inversion of magnetotelluric response

A FEASIBILITY STUDY OF RAPID TWO-DIMENSIONAL MT CURVE COMPARISON INVERSION

ONE-DIMENSIONAL MAGNETOTELLURIC CURVE COMPARISON INVERSION WITH PHASE INFORMATION

Error effects on magnetotelluric inversion

Two-dimensional magnetotelluric inversion method based on deep learning

Sensitivity study on two-dimensional and three-dimensional magnetotelluric inversions

Inversion of marine magnetotelluric and controlled source electromagnetic data with site gap and low signal-to-noise ratio

3-D Inversion of the Data-Normalized Time-Lapse Controlled-Source Audio-Frequency Magnetotelluric Method

One-dimensional magnetotelluric parallel inversion using a ResNet1D-8 residual neural network

A two-step three dimensional marine magnetotelluric inversion method with considering rugged seafloor topography: Synthetic studies

Physics-Informed Deep Learning Inversion with Application to Noisy Magnetotelluric Measurements

An Intelligent MT Data Inversion Method with Seismic Attribute Enhancement.

High-resolution Magnetotelluric Data Inversion Constrained with Seismic Texture

Magnetotelluric Closed-Loop Inversion

3-D Magnetotelluric Deep Learning Inversion Guided by Pseudo-Physical Information

Appraisal of the Magnetotelluric and Magnetovariational Transfer Functions' Selection in a 3-D Inversion

Regularized supervised descent method for 2-D magnetotelluric data inversion

Gradient filtering regularization for 3D MT inversion based on unstructured tetrahedral discretization

Two-dimensional sharp boundary magnetotelluric inversion using Bayesian theory

An MT Data Inversion Method Constrained by Seismic Texture

Power-Type Structural Self-Constrained Inversion Methods of Gravity and Magnetic Data