Including physics in deep learning - An example from 4D seismic pressure saturation inversion

J. S. Dramsch, G. Corte, H. Amini, C. MacBeth, M. Lüthje

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Geoscience data often have to rely on strong priors in the face of uncertainty. Additionally, we often try to detect or model anomalous sparse data that can appear as an outlier in machine learning models. These are classic examples of imbalanced learning. Approaching these problems can benefit from including prior information from physics models or transforming data to a beneficial domain. We show an example of including physical information in the architecture of a neural network as prior information. We go on to present noise injection at training time to successfully transfer the network from synthetic data to field data.

Original languageEnglish
Title of host publication81st EAGE Conference and Exhibition 2019 Workshop Programme
PublisherEAGE Publishing BV
ISBN (Electronic)9789462822924
DOIs
Publication statusPublished - 3 Jun 2019
Event81st EAGE Conference and Exhibition 2019 - London, United Kingdom
Duration: 3 Jun 20196 Jun 2019

Conference

Conference81st EAGE Conference and Exhibition 2019
CountryUnited Kingdom
CityLondon
Period3/06/196/06/19

ASJC Scopus subject areas

  • Geochemistry and Petrology
  • Geophysics

Fingerprint Dive into the research topics of 'Including physics in deep learning - An example from 4D seismic pressure saturation inversion'. Together they form a unique fingerprint.

  • Cite this

    Dramsch, J. S., Corte, G., Amini, H., MacBeth, C., & Lüthje, M. (2019). Including physics in deep learning - An example from 4D seismic pressure saturation inversion. In 81st EAGE Conference and Exhibition 2019 Workshop Programme EAGE Publishing BV. https://doi.org/10.3997/2214-4609.201901967