Machine Learning Methods to Catalog Sources from Diverse, Widely Distributed Sensors

The efficient detection and understanding of clandestine nuclear device testing is critical to the security and credibility of the organizations currently tasked with this important work. At present, the Air Force Tactical Applications Center (AFTAC) and likewise, the Comprehensive Test Ban Treaty Organization (CTBTO) are running methods and algorithms that are out of date. These approaches use a complex Export-Transform-Load (ETL) that is focused on a central database, constituting a single point of failure as well as being a restriction to increasing the quantity and quality of the analytical effort. th Detecting and understanding of clandestine nuclear device testing is a problem that has been around for a very long time. Most of the work in this area has been based on physical attributes as revealed in the structure of seismic waves propagating through the earth. What advancement can be made in a system that has been established for more than half a century? The opportunity is to apply machine learning (ML) to improve the analysis of seismic waves. This proposal will build off a proven seismic processing system that provides adaptable streaming analytics and eventually into a Distributed HPC Seismic System fully integrated with today’s and future machine learning.