Home

Welcome to the Section on Statistics in Defense and National Security

The purpose of this section is to provide a forum for statisticians working in the areas of defense, military research, national security, homeland security, and counterterrorism. This includes assisting in the advancement of knowledge in defense and national security by fostering wider and more effective use of statistical techniques, promoting the statistical profession and statistical best practices within the defense and national security governmental and affiliated organizations, encouraging professional review of statistical methods and statistical activities carried out in support of national security, particularly statistical research that is not available for public review, and disseminating information and providing training opportunities for individuals involved in the production and use of defense and national security data and statistics.


The Winter 2024 Newsletter is now available!

Newsletters are available under the News page.


Upcoming Webinar

Dr. Annie Sauer Booth, Department of Statistics, NC State University
Deep Gaussian Process Surrogates for Computer Experiments  
March 19, 2024

Registration link: https://jhuapl.zoomgov.com/meeting/register/vJIsdO6urT8tGmITiZiDc7r_Pr4gXgFmsbM

Recording will posted on the SDNS YouTube Channel when available


Abstract:  

This talk provides an overview of Bayesian deep Gaussian processes (DGPs) as surrogate models for computer experiments.  Computer experiments are invaluable tools for replacing and/or supplementing direct experimentation, particularly in settings where physical experimentation is restricted by ethical, time, financial, or practicality constraints.  Such simulations are necessarily complex and require statistical “surrogate” models, trained on a limited budget of simulator evaluations, which can provide predictions and uncertainty quantification at untried input configurations.  Gaussian process (GP) surrogates are the canonical choice, but they are limited by stationarity constraints.  DGPs upgrade ordinary GPs through functional composition, in which intermediate GP layers warp the original inputs, providing flexibility to model non-stationary dynamics.  In large data settings, we integrate Vecchia approximation for faster computation.  In small data settings, we utilize strategic active learning/sequential designs with a variety of objectives including variance reduction, Bayesian optimization, and reliability analysis.  We showcase implementation in the “deepgp” package for R on CRAN.


Recent Shared Files

No Data Found

Either the content you're seeking doesn't exist or it requires proper authentication before viewing.

Latest Discussions

There are no discussions to which you have access

Either the content you're seeking doesn't exist or it requires proper authentication before viewing.