11:15 AM - 11:35 AM
[U12-02] Driving Earthquake Seismology Science through Machine Learning
★Invited Papers
Keywords:Machine learning, Earthquake seismology, Earthquake catalog development, Seismicity analysis, Ground-motion prediction, Geodetic data
Recently, machine learning (ML) technology, including deep learning, has made remarkable progress in various scientific fields, including earthquake seismology, producing vast research findings. Here, we review the applications of ML in several fields of earthquake seismology and discuss the strengths and difficulties of using ML.
First, we explore studies on the development of earthquake catalogs, including their elemental processes such as event detection/classification, arrival time picking, similar waveform searching, focal mechanism analysis, and paleoseismic record analysis. We then introduce studies related to earthquake risk evaluation and seismicity analysis. Additionally, we review studies on ground-motion prediction, which are classified into four groups depending on whether the output is ground-motion intensity or ground-motion time series and the input is features (individual measurable properties) or time series. We discuss the effect of imbalanced ground-motion data on ML models and the approaches taken to address the problem. Finally, we summarize the analysis of geodetic data related to crustal deformation, focusing on clustering analysis and detection of geodetic signals caused by seismic/aseismic phenomena.
ML technologies have significantly advanced these fields; however, unique challenges persist. For example, the imbalance in natural datasets is problematic in many cases, possibly causing misevaluation or misinterpretation. Effective approaches to address this problem include data augmentation, simultaneous use of domain knowledge, and transfer learning. Although challenges arise from the black-box nature of DL, the latest techniques, such as PINNs, BNN, and XAI, can address them. The efficiency, accuracy, and flexibility of ML are the driving forces behind the establishment of its usage for various tasks in earthquake seismology. There remain many problems where ML can effectively solve, and its application will further expand and advance our knowledge of earthquake seismology.
First, we explore studies on the development of earthquake catalogs, including their elemental processes such as event detection/classification, arrival time picking, similar waveform searching, focal mechanism analysis, and paleoseismic record analysis. We then introduce studies related to earthquake risk evaluation and seismicity analysis. Additionally, we review studies on ground-motion prediction, which are classified into four groups depending on whether the output is ground-motion intensity or ground-motion time series and the input is features (individual measurable properties) or time series. We discuss the effect of imbalanced ground-motion data on ML models and the approaches taken to address the problem. Finally, we summarize the analysis of geodetic data related to crustal deformation, focusing on clustering analysis and detection of geodetic signals caused by seismic/aseismic phenomena.
ML technologies have significantly advanced these fields; however, unique challenges persist. For example, the imbalance in natural datasets is problematic in many cases, possibly causing misevaluation or misinterpretation. Effective approaches to address this problem include data augmentation, simultaneous use of domain knowledge, and transfer learning. Although challenges arise from the black-box nature of DL, the latest techniques, such as PINNs, BNN, and XAI, can address them. The efficiency, accuracy, and flexibility of ML are the driving forces behind the establishment of its usage for various tasks in earthquake seismology. There remain many problems where ML can effectively solve, and its application will further expand and advance our knowledge of earthquake seismology.