NET-VISA
NET-VISA (NETwork processing Vertically Integrated Seismic Analysis) is a generative model for the creation of seismic events (natural or man-made) the propagation of seismic energy in multiple phases through the earth, and the detection (or mis-detection) of these phases at seismic stations plus the detection of noise. The model also includes infrasound and hydro-acoustic propagation.
This project has been deployed by the CTBT organization of the U.N. to aid human analysts to locate nuclear explosions. It is under evaluation for full deployment to produce automatic bulletins for distribution to all states that have signed the treaty.
Posts
NET-VISA improvements for regional and aftershock event identification at the IDC
Presented in AGU-2018 (S43B-06).
Experience with the NET-VISA “button” and VSEL3 bulletin at the IDC
Presented in the WEG (Waveform Expert Group), August 2018.
Enhancements for Regional Events in NET-VISA v 2.3.6
Presented in the WEG (Waveform Expert Group), August 2018.
Operational experience with next-generation automatic association software NET-VISA.
The ongoing progress with the deployment of NET-VISA at the IDC was presented at the Seismological Society of America (SSA) 2018 meeting.
Progresses with the IDC Infrasound system
The aspects of the NET-VISA model that are specific to infrasound processing were presented at the Acoustical Society of America (ASA) 2018 meeting in Minneapolis.
Evaluation of NET-VISA association and location performance using Ground Truth events and RSTT model based SSSCs
Istvan Bondar has recently completed an extensive analysis of the location errors of NET-VISA using a catalog of ground truth events collected by the ISC. This work was presented at the EGU 2018 conference.
The Bayesian Approach to Association
Nimar presented an invited talk to a panel discussion on the Association problem at American Geophysical Union (AGU) 2017.
Global Infrasound Association at the IDC, advances & performances
A poster presentation was given at the S&T 2017 conference describing all of the progress made with NET-VISA and other infrasound technologies at the IDC.
Fine-Scale Event Location and Error Analysis in NET-VISA
Improvements made to the location search by using a finer search grid, presented in AGU 2016.
Global Infrasound Association Based on Probabilistic Clutter Categorization
The infrasound model in NET-VISA was extended to account for local clutter sources from the oceans or gas flares, for example. This work was presented in AGU 2015.
Using ISC Data to Build a Prior of Seismicity for NET-VISA
The event location prior in NET-VISA is now learned using 10 years of ISC data. This work was presented in S&T 2015.
Fusion of Infrasound with Hydro-Acoustic and Seismic Data in NET-VISA
The first attempt at incorporating Infrasound into NET-VISA.
Threshold Monitoring Maps For Under-water Explosions
The hydro-acoustic model in NET-VISA was extended to display threshold monitoring maps. This work was presented in AGU 2014 (S41B-4470).
Fusion of Seismic and Hydro-Acoustic Propagation
This poster describes the complete NET-VISA hydro-acoustic model. It was presented at the RMR 2014 (Review of Monitoring Research) conference.
A Fusion Model of Seismic and Hydro-Acoustic Propagation for Treaty Monitoring
Includes late breaking results on Ahyl Seamount, an active under-water volcano which erupted on April 23, 2014. (EGU2014-15796).
A Probabilistic Model of Global-Scale Seismology with Veith-Clawson Amplitude Corrections
Veith Clawson amplitude correction curves were added to NET-VISA to allow for the amplitude prediction to be grounded in geophysics. Presented in AGU 2013.
NET-VISA Progress and Enhancements
Status update on current developments in NET-VISA presented at the S&T 2013 conference.
NET-VISA: Network Processing Vertically Integrated Seismic Analysis
Journal paper published in the Bulletin of the Seismological Society of America (BSSA) April 2013, vol. 103 no. 2A pp709-729. This paper won the 2014 Mitchell Prize.
Incorporating seismic phase correlations into a probabilistic model of global-scale seismology
An attempt was made to learn a correlation between seismic phase arrival attributes. However, the work showed that apart from the relative order of the arrivals no other correlations could be inferred. Geophysical Research Abstracts. Vol 15, EGU 2013-13094.
Model-based Bayesian Seismic Monitoring
Nimar’s PhD thesis. Official Link
ANALYST EVALUATION OF MODEL-BASED BAYESIAN SEISMIC MONITORING AT THE CTBTO
An extensive analyst evaluation of NET-VISA was done with multiple analysts looking at the NET-VISA bulletin in parallel with the existing GA-based SEL3 bulletin to see which one was better. This was presented in MRR (Monitoring Research Review) 2012.
A model of seismic coda arrivals to suppress spurious events
A coda model was added to NET-VISA to prevent the model from bulding multiple events in the vicinity of a large events. After this change to the model there was no need for a seprate post-processing to filter out the event clusters. Geophysical Research Abstracts, Vol. 14, EGU2012-6763.
Scalable Probabilistic Inference for Global Seismic Monitoring
This poster won the Best Paper Award in Seismology in AGU 2011.
Signal-Based Bayesian Monitoring
Presented at Science & Technology (S&T), 2011.
NET-VISA: Model and Inference Improvements
Presented at Science & Technology (S&T), 2011.
Modeling Aftershock Sequences in NET-VISA
Presented in a workshop before the main S&T 2011 conference.
Global Seismic Monitoring: A Bayesian Approach
Presented at the American Association of Artificial Intelligence (AAAI), 2011.
Machine Learning at the CTBTO. Testing, and Evaluation of the False Events Identification (FEI) and Vertically Integrated Seismic Association (VISA) Projects
Presented at the Monitoring Research Review (MRR), 2011.
Bayesian Treaty Monitoring: Preliminary Report
Presented at Monitoring Research Review (MRR), 2011
Global seismic monitoring as probabilistic inference
Presented at Neural Information Processing Systems (NIPS), 2010.
Vertically Integrated Seismological Analysis
Presented at European Seismological Conference (ESC) 2010.
Automatic Inference in BLOG.
Presented at Statistical Relational AI (StarAI) Workshop. American Association of Artificial Intelligence (AAAI), 2010.
Gibbs Sampling in Open-Universe Stochastic Languages
Published in Uncertainty in AI (UAI), 2010.
Status of the machine learning efforts at the international data centre of the CTBTO
Presented in Monitoring Research Review 2010.
Vertically Integrated Seismological Analysis 2 : Inference.
Presented at American Geophysical Union (AGU), 2009.
Vertically Integrated Seismological Analysis I : Modeling.
Presented at American Geophysical Union (AGU), 2009.
Joint probabilistic detection, association, & localization II: MCMC Inference.
Presented in International Scientific Studies (ISS), 2009.
Joint probabilistic detection, association, & localization I: Hierarchical modeling.
Presented in International Scientific Studies (ISS), 2009.
Automatic Inference in PyBLOG
Presented in Probabilistic Programming Workshop. Neural Information Processing Systems (NIPS), 2008.
subscribe via RSS