Mining a haystack of data to find the golden needle
Computational expertise is essential to LLNL’s Global Security (GS) organization. GS develops multidisciplinary science and technology solutions to protect the nation and connect the world. Many of GS’s programmatic and scientific applications depend on making sense of vast quantities of data, so the 125 computer scientists, researchers, and computer technicians in the GS Computing Applications Division (GS-CAD) are often tasked with designing architectures, developing algorithms, and writing user interfaces to achieve that goal.
According to Jamie Van Randwyk, GS-CAD division leader, GS-CAD staff members support many of the more than 800 active projects under way in GS at any one time, in such diverse areas as atmospheric monitoring, bioinformatics, energy, intelligence, bioterrorism, and nonproliferation. “In many of these areas, we are developing tools to help scientists efficiently look through a lot of unstructured text and glean useful information,” says Jamie. “This is a particularly hard computer science problem.”
GS-CAD computer scientists help organize and display data in such a way that intelligence and nonproliferation analysts can more easily pinpoint anomalous behavior in a huge quantity of data. Jamie likens the task to finding a needle in a haystack. The team also performs modeling and simulation for energy and defense efforts.
One realistic, high-resolution simulation developed by a cross-Lab team, including some GS-CAD personnel, has become the most widely used conflict model in the Defense Department for strategic planning. GS-CAD also supports clean energy research, a major growth area for LLNL, with power grid modeling and program development work. “Additionally, we collaborate with Russian scientists on several projects to assist in nonproliferation of nuclear weapons,” Jamie says.
GS-CAD Computer Scientist Dean Williams works with a team on the Earth System Grid Federation, an immense computerized climate model that standardizes and organizes data collected from 23 countries. “I work with a great team, the best team you could ever hope to work with,” Dean says. “The software that we have developed allows scientists around the world to share data in a standard format so they can more easily conduct their research and publish results.”
Adds Jamie, “In Computing, we want to advance the discipline of computing and provide the very best computer scientists to our programs. The challenge in supporting Global Security is that most of the programs have very short funding cycles. We need employees with very versatile skill sets, as they have the opportunity to change projects frequently.”
One of the more unusual methods initiated to sharpen skills and promote creativity and retention among GS-CAD’s computer scientists is “ShipIt Day,” an optional coding marathon that is held several times a year. Inspired by similar “hackathons” in industry, participants are given 24 hours to create a project, which they then present to their peers. Jamie notes that several leading ideas from the events are being implemented by GS programs. “These activities offer big benefits to employees and to our programmatic partners.”
Biodefense Knowledge Center
The “go-to” source for vetted information on biothreats
Sleuthing seismic signals
Developing core technologies to change the way geophysical monitoring systems work and data is processed
The ground truth
High-resolution, data-driven conflict simulation and modeling tools
Big machines, data, and ideas
Advancing the state-of-the-art of our nation’s data science capabilities