Researchers led by Tomohisa Okazaki, Research Scientist, and Naonori Ueda, Team Leader from the Disaster Resilience Science Team have developed an innovative analysis method for crustal deformation by a physics-informed deep learning approach.
The findings could lead to new analysis methods by deep learning for earthquake processes and future earthquake potentials.
In conventional numerical methods, model regions are divided into simple meshes on a computer. However, numerous meshes are required to represent complex underground structures, which results in high computational costs.
In this study, the research team proposed a physics-informed deep learning approach to model crustal deformation due to earthquakes. This approach can represent continuous deformation flexibly and thus may be a powerful tool for realizing a wide variety of modeling applications in earthquake analyses.
This study was published in Nature Communications* on November 19, 2022.
*Nature Communications is an open-access journal that publishes high-quality research from all areas of the natural sciences.
Title:
Physics-informed deep learning approach for modeling crustal deformation
Authors:
Tomohisa Okazaki, Takeo Ito, Kazuro Hirahara & Naonori Ueda
For more information, please see the following website.
Nature Communications
