13 June 2017

Intel officials pin high hopes on automation, artificial intelligence

By: Amber Corrin

Intelligence analysts are swimming in data pouring in from an array of vehicles and platforms — a problem that isn’t new, but for which government leaders still seek the right solutions.

To help stem the deluge and better position analysts and key mission-critical data, intelligence community officials are targeting automation as a high priority, with a futuristic vision for applications down the road as well.

“A significant chunk of my analytic workforce today, I will send them to a dark room to look at TV monitors to do national security-essential work … but boy, is it inefficient,” Robert Cardillo, director of the National Geospatial-Intelligence Agency, told reporters at the annual GEOINT Symposium in San Antonio, Texas. “The number of people needed to maintain awareness of, if not exploitation of, one sensor is really daunting. I suspect we’re going to get more of those sensors. … I can’t double my human population in those dark rooms.”

The near-term goal particularly centers on analysis of full-motion video that streams in from unmanned aerial systems — “the wolf really close to the door,” as Cardillo put it. But he and other officials also are looking toward future uses for different types of automation, including artificial intelligence. And Cardillo, among others, are looking to partner up for help.

“As the commercial industry and academic think tanks and advanced science and engineering schools move to artificial intelligence and machine learning, they’re all desperate to get a hold of some data with [which] to train their algorithms and teach their machines to learn,” Cardillo said. He added that intelligence community leaders, including those at the NGA, are looking at how to safely expose data sets to accelerate development in automated tools.

But the NGA isn’t just looking outside for solutions. Internally, the agency has launched a new Office of Ventures and Innovation aimed at guiding emerging technology from incubation through the entire life cycle.

“To get to this automated, augmented future that we’re talking about, we need to coordinate across a lot of different parts of the agency,” Anthony Vinci, NGA director of plans and programs, told C4ISRNET. “It’s not just a technology issue, it’s bringing technology and [research and development] into the operational units, into analysis or into the business services units, human development or finance for business analytics.”

Vinci said the NGA is working closely with other government agencies, including the Defense Department, to further automation and AI and get to a new level of intelligence and military operations.

“How can we use automation to take some of that pressure off of the analysts who are putting together those products? How do we buy back some of their time by automating some of these processes so they can focus on the higher-end analysis that they need to be focusing on? That’s a fundamental thing we’re trying to do using all the tools at the agency’s disposal,” Vinci said.

“On the higher end — the exquisite analysis — how can we bring in modeling and AI and some of those tools that are on the high end of technology to support analysis and do missions that weren’t even possible until now because we didn’t have the tools to analyze that amount of data, or because we couldn’t analyze some complex phenomenonologies? How can we bring them those tools and those models?”

In a future operational landscape where autonomous vehicles dominate and interact with adversaries, it’s easy to see where intelligence missions cross over to operational military missions, Vinci noted.

“These tools are not just intelligence tools; they’re operational as well. We’re not just an intelligence agency, we’re a combat support agency, so we help across the spectrum,” he said. “We’re a support function to someone who has to make a decision.

"We have to be able to act faster and get as far left of boom as possible.”

No comments: