HURLBURT FIELD, FL – Warfighting innovation is making its way to Air Force Special Operations Command in the form of Artificial Intelligence.
The Project Maven team presented data-tagging technology to 11th Special Operations Intelligence Squadron intelligence analysts here Aug. 24.
Project Maven is an Artificial Intelligence effort to assist the Department of Defense with organizing data collected for intelligence gathering.
“Data-tagging is looking at individual images and selecting features on those images for identification,” said Master Sgt. Daniel Warden, AFSOC Initiatives Group superintendent. “It’s training the machine to recognize what a building, car or person looks like. The more instances we have of each object, the ‘smarter’ the machine becomes. Once enough images (roughly 100,000 per object) are tagged, the machine will be able to make the determinations on its own.”
The end result of this initiative will be speeding intelligence analysts’ decision-making processes.
“Once the machine is trained, it will be able to look for specific things,” said Warden. “These things could be a red sedan, black pickup truck, person with a blue hat, person with a weapon, or any other descriptive identifier.”
The technology these machines use to tag images is similar to how Facebook will recognize faces of friends and family in newly uploaded photos.
Enhanced focus on the mission of analyzing data will result in more efficient and effective workplace practices.
“All of this information now gives an analyst more time to make tough decisions because the computer makes the easy decisions for us by sorting through all of the data out there,” said Warden. “There’s not enough people in the world to look at all of the data. This helps us solve that problem as we go through and teach these machines how to do their job and automate some of our processes so we can more efficiently use the data we’re being presented.”
Staff Sgt. Ashlie Robledo, 11th SOIS intelligence analyst, said this technology will help focus analytical skills on something that needs human brainpower instead of mundane, day-to-day tasks that a machine can pick up.
“It will help with the multi-tasking aspect,” said Robledo. “We can prioritize different mission sets, and if something is slow-going, we can use this technology to meet the intent of that mission. If something pops up and needs the attention of a person, then that person can be right there to do the analysis and whatever the job entails.”
The project will assist commanders as well.
“It gives commanders options in how they use technology to improve efficiency and become more effective,” said Diego Charles, AFSOC Intelligence Directorate, Intelligence, Surveillance and Reconnaissance Strategy and Plans Division technical advisor. “This, in turn, allows commanders to apply their resources to other problem sets.”