NGA director eyes automation in response to data deluge
Seventy-three years ago today, approximately 156,000 American, British, and Canadian troops stormed five beaches on the northern coast of France. Their assault—D-Day—was a pivotal turning point in World War II and the largest amphibious military assault in history. That it succeeded was a credit not only to the thousands of men who left their lives on the sand that day, but also to the many intelligence analysts on whom General Dwight D. Eisenhower relied to plan and execute the Allied forces’ invasion of Normandy, codenamed “Operation Overlord.”
“Operation Overlord was fueled by some of the most important maps and charts and imagery intelligence in history,” National Geospatial-Intelligence Agency (NGA) Director Robert Cardillo said Monday during a GEOINT 2017 keynote address, which he opened with a reflection on D-Day that he called a “pointed reminder of who we are and who we must be as an Intelligence Community.”
This is our time. In a world that has image scarcity in its rearview mirror and a data tidal wave on its horizon, we’ll sink, we’ll swim, or we’ll ride this rising tide. I say we ride.”
—Robert Cardillo, Director, NGA
“We exist for one reason: to advantage our decision-makers and the warfighters that deploy their decisions,” Cardillo told the morning crowd. “And when those deployments transition from defend to defeat, we must ensure that their fight is not fair, and that we have the upper hand … We’re our own version of D minus 1. This is our time. In a world that has image scarcity in its rearview mirror and a data tidal wave on its horizon, we’ll sink, we’ll swim, or we’ll ride this rising tide. I say we ride.”
The director devoted the balance of his 30-minute address to explaining how NGA plans to “ride” the growing wave of geospatial data in order to deliver strategic advantage in what he called a “race for space and for time.”
Automation and augmentation are at the heart of NGA’s plans. “If we attempted to manually exploit all of the imagery we’ll collect over the next 20 years, we’d need 8 million imagery analysts,” explained Cardillo, who said NGA already collects with a single sensor—every day—the data equivalent of three NFL seasons recorded in high-definition video. “Imagine you’re a coach trying to understand the strategy of his opponent by watching every game and every play for three seasons, all in a single day … That’s exactly what we ask our analysts to do when we don’t augment them with automation. All this data, combined with dramatic improvements in computing power, represents a phenomenal opportunity.”
Cardillo announced several new ways in which NGA is seizing that opportunity, the goal of which is to eventually automate 75 percent of the tasks NGA analysts perform, enabling them to spend less time completing rote tasks and more time answering complex intelligence questions.
For one, the agency is launching new GEOINT tools and services that make the case for automation, such as Beachfront, which automatically creates new coastline data using commercial satellite imagery. For example, vectors of the vast river delta system that separates India from Bangladesh would have taken a human analyst five hours to produce, but took Beachfront less than six minutes, Cardillo said.
In addition to new tools that leverage automation, NGA is appointing new leaders whose mission will be to realize its potential: William “Buzz” Roberts and Dr. David Bray.
Roberts, who currently serves as head of special programs for NGA Research, will become director of artificial intelligence, automation, and augmentation. His charge, Cardillo said, will be to successfully apply automation to one of NGA’s most challenging data sources: full-motion video (FMV).
“If we continue to exploit FMV the way we do it today it will become an existential threat to NGA as it eats up more and more of our analytic capacity,” Cardillo said. “It’s time-consuming, manually intensive, redundantly processed, and it leaves a great deal of data undiscovered and unexploited. In other words, while it remains essential to our national security it’s both very expensive and extremely inefficient. We must change this.”
Bray, who is currently chief information officer at the Federal Communications Commission (FCC), will lead a new Office of Ventures and Innovation that will be tasked with driving innovation both within NGA and externally with its commercial partners. A major new initiative led by that office will be a public-private partnership between NGA and industry that simultaneously supports and advances the capabilities of both. The product of the partnership will be what Cardillo described as a “data brokerage” that will furnish industry with reams of historic data from NGA archives in exchange for cutting-edge algorithms produced by applying machine learning to decades worth of geospatial imagery and analysis.
“NGA sits on an amazing and broadly untapped resource,” Cardillo said. “In a world of deep learning, historic data and ground truth are of immense value. Some have even called such data ‘the new oil.’”
It’s time-consuming, manually intensive, redundantly processed, and it leaves a great deal of data undiscovered and unexploited. In other words, while it remains essential to our national security it’s both very expensive and extremely inefficient. We must change this.” —Robert Cardillo, Director, NGA
In a press conference after his speech, the director said NGA’s data brokerage could potentially operate as a government-run nonprofit corporation, which would be managed by an outside entity on behalf of federal stakeholders. “There’s something called a B corporation [which is] essentially a government company,” he said, acknowledging that such a structure would require legislative support to proceed. “I’ve begun those communications with the Hill, and I’ve gotten good early reception to it.”
To demonstrate how a public-private partnership with NGA might work, Cardillo said the agency will soon launch a “Trafficking Data Challenge” whereby NGA will provide industry with geospatial data to use in developing solutions that model, track, and deter trafficking of humans, wildlife, drugs, and weapons.
“We’ll run this in the same manner as our hackathons, with a team of experts assembled to judge the submissions,” Cardillo said. “Let’s consider the challenge a preview of that public-private partnership, because for the prizes we’re not simply awarding money—you would be able to participate in a cooperative data project with NGA.”
Three-quarters of a century after D-Day, GEOINT looks remarkably different. As it storms digital rather than physical beaches, however, the U.S. can secure victory now the same way as it did then, Cardillo said: by uniting around a single objective.
“In a world where so many things divide us, our mission brings us together,” he concluded. “It unites our community and our focus. It reminds us just how closely connected we all are. And if we can do more than just see these connections—if we can honor them, if we can strengthen them—then that unity of purpose will inspire us and propel us onward. That unity will develop and deliver the next generation of intelligence that the world demands and our customers deserve.”
Collaboration and planning integral to best serve the GEOINT Community