People and culture change are essential to fully leverage the power of high-performance computing (HPC), according to experts from the National Geospatial-Intelligence Agency (NGA) and Oak Ridge National Laboratory (ORNL).

USGIF hosted a Geocomputational Breakout Breakfast Feb. 15 at NGA as a follow on to a summit NGA and ORNL hosted the previous day.

The prior day’s summit focused on the power of HPC and the data challenges it could help the agency address. USGIF’s breakfast was an opportunity for those who were unable to attend the summit to hear a summarized overview.

NGA Director of Research Peter Highnam kicked off the breakfast, sharing his thoughts on the potential of HPC for GEOINT analysis.

“Automation is a driving opportunity,” Highnam said. “Machine and deep learning are just pieces of the puzzle. I see three [HPC] objectives: automation, transition in practice, and enriching GEOINT to present information.”

NGA also shared its announcement of the Magellan Project, which allows an analyst to view any location in the world in the past, present, and future to predict scenarios in order to make more informed decisions.

Shaun Gleason, director of ORNL’s Computational Sciences and Engineering Division, recounted some key takeaways from the summit, which included insight from industry and academia partners using HPC. One example Gleason described was how the University of Illinois uses its ROGER (Resourcing Open Geospatial Education and Research) supercomuter in its CyberGIS Center.

“People are the true expertise,” Gleason said. “The [Illinois] campus champions research groups to learn about problems to bring to HPC. There’s constant communication to create an ecosystem of people to talk to each other.”

Gleason also said people—forward-thinking analysts and subject matter experts—are essential help drive the transition to HPC.

The breakfast concluded with a panel discussion including Justin Poole, director of NGA’s Source Directorate, and Frank Avila, a senior scientist with the agency’s Office of Sciences and Methodologies.

Both agreed there are not enough eyeballs on the explosion of GEOINT information and that they’re shifting to a quantitative approach.

“At the end of the day, a human will make judgment of what’s going to happen,” Avila said. “Analysts need to feel confident in how we bring [in HPC] technology. … It’s exciting and there’s going to be big challenges, but it’s an opportunity to do things better.”

Poole emphasized how community support is necessary to fully take advantage of high-performance computing at the agency.

“It is a culture change and we need to embrace and adopt [HPC] in order to move ahead,” Poole said.

Photo Credit: Oak Ridge National Laboratory


Posted by Lindsay Tilton Mitchell