S011
From task specific Machine Learning to Foundation Models in Seismology and Geodesy
Session ID#: 247860
Session Description:
Machine learning is rapidly advancing scientific research and solid Earth geophysics has benefited in this data driven modeling transformation. Many existing models are task-specific without a generalization of the data, therefore limiting widespread application across different settings. Inspired by natural language processing and computer vision, we can extend the foundation model concept to seismology and geodesy. Foundation models are large-scale, pre-trained deep neural networks that learn general representations from vast amounts of data and easily adapt to a range of downstream task-specific applications. This session aims to highlight advancements in foundation models for solid Earth geophysics. We invite contributions that explore machine learning architectures, domain adaptation strategies, large-scale pretraining methodologies, evaluation methods, and downstream applications of foundation models (even existing ones) in seismology and geodesy. We aim to foster discussions on the impact of foundation models on seismology and geodesy research, as well as their contribution to scientific progress.
Index Terms:
1299 General or miscellaneous [GEODESY AND GRAVITY]
1942 Machine learning [INFORMATICS]
4315 Monitoring, forecasting, prediction [NATURAL HAZARDS]
7230 Seismicity and tectonics [SEISMOLOGY]
Primary Convener: Laura Laurenti, ETH Swiss Federal Institute of Technology Zurich, EAPS, Zurich, Switzerland
Convener: Christopher W Johnson, Los Alamos National Laboratory, Los Alamos, NM, United States
A Geologically-Informed Regularizer for Enhanced Seismic Fault Detection in the Indian Krishna Godavari Basin (1908427)
Tiash Ghosh, Indian Institute of Technology Kharagpur, Department of Electrical Engineering, Kharagpur, India, Sanjai Kumar Singh, GEOPIC, ONGC, Dehradun, India, Mamata Jenamani, Indian Institute of Technology, Kharagpur, Department of Industrial & Systems Engineering, Kharagpur, India and Aurobinda Routray, Indian Institute of Technology, Kharagpur, Department of Electrical Engineering, Kharagpur, India
A Purely Data-driven Generative Model for Seismic Waveform Synthesis with Minimal Conditional Information (1981208)
Donghun Lee1, Jaeheun Jung1, Jaehyuk Lee1, Chang-Hae Jung1,2, Hanyoung Kim1, Bosung Jung1, Jongwon Han3 and Seongryong Kim4, (1)Korea University, Department of Mathematics, Seoul, South Korea, (2)MakinaRocks Inc., Seoul, South Korea, (3)Korea Institute of Geoscience and Mineral Resources (KIGAM), Earthquake Research Center, Daejeon, South Korea, (4)Korea University, Department of Earth and Environmental Sciences, Seoul, South Korea
A State-Space Foundation Model for Raw Seismic Waveforms (1984992)
David Liu, Boston University, Boston, MA, United States, Laura Laurenti, Sapienza University of Rome, Computer, Control and Management Engineering, Rome, Italy, Christopher W Johnson, Los Alamos National Laboratory, Los Alamos, NM, United States and Brian Kulis, Boston University, ECE, Boston, United States
BiLSTM-Based Low-Frequency Extrapolation: Generalizing from Synthetic Training to Field Airgun Data (1941777)
Ivan Deiana, Stanford University, Stanford, CA, United States, Giacomo Roncoroni, University of Trieste, Trieste, Italy, Biondo Leonardo Biondi, Stanford University, Geophysics, Stanford, CA, United States and Shuki Ronen, Stanford University, Stanford, United States
Comparison of the Earthworm real-time associator binder and PhaseNet+GaMMA for small scale seismic arrays (1924497)
Paul A Friberg1, Sebin John1,2, Alex Dzubay1, Josh Stachnik3, Scott Fertig1 and sebin john, (1)Instrumental Software Technologies, Inc., Saratoga Springs, United States, (2)Fairbanks, Alaska, UNITED STATES, (3)Instrumental Software Technologies, Inc., Saratoga Springs, NY, United States
Exploring Earthquake Rupture Scenarios at the San Andreas-Garlock Fault Junction: A Simulation-Driven Machine Learning Approach (1967074)
Abhijit Ghosh1, Shankho Niyogi2, Evan Marschall3, Roby Douilly3 and David Douglas Oglesby4, (1)University of California Riverside, Department of Earth and Planetary Sciences, Riverside, CA, United States, (2)University of California, Riverside, Department of Earth and Planetary Sciences, Riverside, CA, United States, (3)University of California Riverside, Riverside, United States, (4)University of California, Riverside, Department of Earth and Planetary Sciences, Riverside, United States
Foundation Models and Earth: Toward Universal Representations for Understanding Seismicity (Invited) (2003727)
Ivan Dokmanić, University of Basel, Basel, Switzerland, Tianlin Tianlin, Google Deep Mind, Paris, France, Maarten V De Hoop, Rice University, Houston, United States, Chris Marone, Sapienza University of Rome, Rome, Italy, Laura Laurenti, Sapienza University of Rome, Computer, Control and Management Engineering, Rome, Italy, Jannes Münchmeyer, Helmholtz Centre Potsdam GFZ German Research Centre for Geosciences, Geophysics, Potsdam, Germany and Esteban Chaves, Observatorio Vulcanológico y Sismológico de Costa Rica, Heredia, Costa Rica
Is Regional or Global Model Acceptable for Local Discriminations of Microearthquakes and Quarry Blast? (1910533)
Eunjin Lee1, Kwang-Hee Kim1 and Hyeong-Tae Jou2, (1)Pusan National University, Department of Geological Science, Busan, Korea, Republic of (South), (2)Institute of Geologic Hazard & Industrial Resources, Pusan National University, Department of Geological Science, Busan, South Korea
Machine Learning-Based Quality Control and Categorizations for Seismic Receiver Functions (2002684)
Sheng Dong1, Hanxiao Wu1, Amarjeet Kumar1, Vera Schulte-Pelkum2 and Weisen Shen1, (1)Stony Brook University, Department of Geosciences, Stony Brook, United States, (2)Univ Colorado Boulder, Boulder, United States
Rapid Estimation of Earthquake Location and Magnitude Using a Large Language Model (Invited) (1852083)
Aurora Bassani1, Daniele Trappolini2, Giulio Poggiali1, Elisa Tinti1, Fabio Galasso3, Alberto Michelini4 and Chris Marone1, (1)Sapienza University of Rome, Rome, Italy, (2)Sapienza University of Rome, Department of Computer, Control and Management Engineering,Sapienza4 University of Rome, Rome, Italy, Rome, Italy, (3)Sapienza University of Rome, Computer Science, Roma, Italy, (4)Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy
Reconstruction of Seismic Wavefields using Shallow Recurrent Decoder: Examples on Simulated Wavefields (1939129)
Yiyu Ni1, Marine Denolle1, Mars Gao2 and Nathan Kutz3, (1)University of Washington Seattle Campus, Department of Earth and Space Sciences, Seattle, United States, (2)University of Washington Seattle Campus, Paul G. Allen School of Computer Science & Engineering, Seattle, United States, (3)University of Washington Seattle Campus, Departments of Applied Mathematics and Electrical & Computer Engineering, Seattle, United States
Repurposing Vision Foundation Models for Multi-Modal Earth System Analysis (1991661)
Zhengfa Bi, Lawrence Berkeley National Laboratory, Berkeley, United States, Xinming Wu, USTC University of Science and Technology of China, Hefei, China, Xiaohua Xu, University of Texas, Institute for Geophysics, Austin, United States and Nori Nakata, Massachusetts Institute of Technology, Cambridge, United States
Toward Real-Time Seismic Imaging: Efficient Learning-Based Inference of Subsurface Models (1978293)
Mike Wendels1, Yiyu Ni2, Marine Denolle2 and Nathan Kutz3, (1)University of Washington Seattle Campus, Department of Applied Mathematics, Seattle, United States, (2)University of Washington Seattle Campus, Department of Earth and Space Sciences, Seattle, United States, (3)University of Washington Seattle Campus, Departments of Applied Mathematics and Electrical & Computer Engineering, Seattle, United States
