Room: Key Ballroom 10
Date: 4/15/2025
Session Time: 4:30 PM to 5:45 PM (local time)
Scientific Machine Learning for Forward and Inverse Wave Equation Problems
The synergy between scientific machine learning (SciML) and computational mechanics is transforming our approach to forward and inverse problems governed by complex partial differential equations, particularly in seismic wave propagation. This session brings together experts from various disciplines -including seismology, computational geomechanics and dynamics, and the broader SciML community - to explore cutting-edge methods and real-world applications in this interdisciplinary arena. The goal is to bridge the gap between traditional computational approaches and emerging AI-driven techniques in the modeling and analysis of wave phenomena. We welcome contributions that leverage data-driven and physics-inspired machine-learning techniques to enhance the modeling, simulation, and interpretation of seismic wave phenomena across different fields. Topics of interest include but are not limited to (1) Applications of PINNs in solving forward and inverse wave problems, handling complex boundary conditions and subsurface velocity models; (2) Advancements in neural operators, such as Deep Operator Networks and Fourier Neural Operators for efficient and accurate wavefield simulations; and (3) Innovative uses of SciML in related fields such as acoustics and ultrasound imaging, elastodynamics, structural health monitoring, planetary seismology, environmental monitoring, and natural hazard assessment. We particularly encourage submissions highlighting how SciML techniques applied to wave equation problems can be adapted or inspire solutions in adjacent fields, fostering the exchange of ideas and methodologies. This session provides a dynamic forum for attendees to discuss theoretical developments, share practical experiences, and identify future research directions that transcend traditional domain boundaries. Through these interactive exchanges, we aim to advance the capabilities of computational models and unlock new potentials in scientific research and engineering applications.
Conveners
Tariq Alkhalifah, King Abdullah University of Science and Technology (tariq.alkhalifah@kaust.edu.sa)
Arash Fathi, ExxonMobil Technology and Engineering (arash.fathi@exxonmobil.com)
Lu Lu, Yale University (lu.lu@yale.edu)
Kami Mohammadi, University of Utah (kami.mohammadi@utah.edu)
Harpreet Sethi, NVIDIA (hasethi@nvidia.com)
Oral Presentations
| Participant Role | Details | Start Time | Minutes | Action |
|---|---|---|---|---|
| Submission | Helmholtz Neural Operator for Full Waveform Inversion Tomography of California | 04:30 PM | 15 | View |
| Submission | Simulating Seismic Wavefields Using Generative Artificial Intelligence | 04:45 PM | 15 | View |
| Submission | Efficient Solutions to the Acoustic Wave Equation Using Extreme Learning Machines With Domain Decomposition | 05:00 PM | 15 | View |
| Submission | An End-to-end Physics-based Deep Learning Approach for Robust Seismic Inversion | 05:15 PM | 15 | View |
| Submission | Seismic Geotechnical Imaging Using Full-waveform Inversion and Physics-informed Neural Networks | 05:30 PM | 15 | View |
| Total: | 75 Minute(s) |
Scientific Machine Learning for Forward and Inverse Wave Equation Problems - I
Description