Generates 9 million grid points of 3D spatiotemporal physical fields in seconds, a 10,000x speedup over traditional physics simulations.
March 19, 2026
Original Paper
Large-Scale 3D Ground-Motion Synthesis with Physics-Inspired Latent Operator Flow Matching
arXiv · 2603.17403
The Takeaway
Demonstrates the power of latent operator flow matching for complex scientific simulations. This level of speedup transforms physical hazard analysis from a supercomputing task to a real-time engineering tool.
From the abstract
Earthquake hazard analysis and design of spatially distributed infrastructure, such as power grids and energy pipeline networks, require scenario-specific ground-motion time histories with realistic frequency content and spatiotemporal coherence. However, producing the large ensembles needed for uncertainty quantification with physics-based simulations is computationally intensive and impractical for engineering workflows. To address this challenge, we introduce Ground-Motion Flow (GMFlow), a ph