A Caltech Library Service

HypoSVI: Hypocenter inversion with Stein variational inference and Physics Informed Neural Networks

Smith, Jonathan D. and Ross, Zachary E. and Azizzadenesheli, Kamyar and Muir, Jack B. (2021) HypoSVI: Hypocenter inversion with Stein variational inference and Physics Informed Neural Networks. . (Unpublished)

[img] PDF - Submitted Version
See Usage Policy.


Use this Persistent URL to link to this item:


We introduce a scheme for probabilistic hypocenter inversion with Stein variational inference. Our approach uses a differentiable forward model in the form of a physics-informed neural network, which we train to solve the Eikonal equation. This allows for rapid approximation of the posterior by iteratively optimizing a collection of particles against a kernelized Stein discrepancy. We show that the method is well-equipped to handle highly non-convex posterior distributions, which are common in hypocentral inverse problems. A suite of experiments is performed to examine the influence of the various hyperparameters. Once trained, the method is valid for any network geometry within the study area without the need to build travel time tables. We show that the computational demands scale efficiently with the number of differential times, making it ideal for large-N sensing technologies like Distributed Acoustic Sensing.

Item Type:Report or Paper (Discussion Paper)
Related URLs:
URLURL TypeDescription Paper ItemSoftware
Smith, Jonathan D.0000-0002-1684-1344
Ross, Zachary E.0000-0002-6343-8400
Azizzadenesheli, Kamyar0000-0001-8507-1868
Muir, Jack B.0000-0003-2617-3420
Additional Information:HypoSVI is avaliable at github, with additional runable Coalab code supplied at the Github. This project was partly supported by a grant from the United States Geological Survey (USGS). We would like to thank Jack Wilding and Bing Q. Li for interesting discussions about the implementation and software usage.
Group:Seismological Laboratory
Funding AgencyGrant Number
Record Number:CaltechAUTHORS:20210225-132745364
Persistent URL:
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:108212
Deposited By: George Porter
Deposited On:01 Mar 2021 15:13
Last Modified:01 Mar 2021 15:13

Repository Staff Only: item control page