Solid Earth Sciences (S)
Session Sub-category Complex & General(CG)
Session ID S-CG71
Title Driving Solid Earth Science through Machine Learning
Short Title Machine Learning in Solid Earth Sciences
Main Convener Name Hisahiko Kubo
Affiliation National Research Institute for Earth Science and Disaster Resilience
Co-Convener 1 Name Makoto Naoi
Affiliation Hokkaido University
Co-Convener 2 Name Keisuke Yano
Affiliation The Institute of Statistical Mathematics
Co-Convener 3 Name Yusuke Tanaka
Affiliation Geospatial Information Authority of Japan
Session Language 
J
Scope 
Machine learning (ML) has brought innovations and remarkable results in various science fields including solid earth science. This session provides an opportunity to inspire each other for future developments by bringing studies of ML applications in various fields including solid earth science. We invite a wide range of presentations on ML and related research, from the budding to the more advanced.
Because this session aims to promote the integration of information science, including machine learning, with solid earth sciences, the participants' areas of expertise and technical terminology vary widely; therefore, the session language will be designated as "J." Verbal communication (oral presentation, Q&A etc.) can be given in either English or Japanese, while text in slides/posters must include English (Japanese text may be accompanied).
Presentation Format Oral and Poster presentation
Invited Authors Daisuke Matsuoka (Japan Agency for Marine-Earth Science and Technology)
Yuya Sasaki (Osaka University)
Time Presentation No Title Presenter
Oral Presentation May 27 PM1
13:45 - 14:00 SCG71-01 Reconstructing mantle thermal convection with an inverse PINN Atsushi Nakao
14:00 - 14:15 SCG71-02 Improvement of Generalization in Seismic Ground-Motion Surrogate Models using Gradient-informed Spectral Feature Distillation Yuto Kuroki
14:15 - 14:30 SCG71-03 Rapid wavefield forecasting for earthquake early warning via deep sequence to sequence learning Rie Nakata
14:30 - 14:45 SCG71-04 Development of the J-SHIS AI Chatbot Yuma Matsumoto
14:45 - 15:15 SCG71-05 Data-driven approaches in weather prediction and climate change adaptation Daisuke Matsuoka
Oral Presentation May 27 PM2
15:30 - 16:00 SCG71-06 Graph Nueral Networks and Solid Earth Science Yuya Sasaki
16:00 - 16:15 SCG71-07 Simultaneous estimation of moment tensors for numerous earthquakes and site corrections using a physics-guided unsupervised neural network Kazutoshi Imanishi
16:15 - 16:30 SCG71-08 A Fully Automated Deep-Learning Framework for Daily Regional Earthquake Cataloging: Design and Cross-Validation Wu-Yu Liao
16:30 - 16:45 SCG71-09 Untangle the Complexities: Multivariate Statistical Insights into the Geochemistry of the Molucca Sea and Philippine Arcs Prya Arif Rahman
Presentation No Title Presenter
Poster Presentation May 27 PM3
SCG71-P01 State Space Model Based Trend Decomposition and Its Application to Machine Learning Based Anomaly Detection Mayu Tsuchiya
SCG71-P02 Enhancing Eruption Predictability Through Machine Learning in Quantitative Volcanology and Geophysics Flavio Cannavo
SCG71-P03 Application of Federated Learning to Earthquake Ground Motion Prediction Models Sho Akagi
SCG71-P04 Development of PINN based method to estimate coseismic slip distribution in a horizontally stratified elastic viscoelastic two dimensional medium Shohei Takai
SCG71-P05 Classification of rock sample derivedAE events using a Convolutional Neural Network (CNN) Yuki Tsurumachi
SCG71-P06 Approaches to Advancing Earthquake Catalogs Based on Routine Seismic Networks in Japan Makoto Naoi
SCG71-P07 Application of PhaseNetWC to MeSO-net Acceleration waveforms and its Performance Evaluation Ichiro Oishi
SCG71-P08 Optimization of the threshold for the phase detection by PhaseNet along with parameters for phase association Kengo Shimojo