Goodfellow participates in training AI

U.S. Air Force 2nd  Lt. Dominque M. Smith-Bryant, a 315th student, logs into the Project Maven website to label objects through the program Nov. 21, 2017. As a student waiting to start class she uses some of her time to help the base team earn rewards in the Project Maven website. (U.S. Air Force photo by Airman 1st Class Seraiah Hines/Released)

U.S. Air Force 2nd Lt. Dominque M. Smith-Bryant, a 315th student, logs into the Project Maven website to label objects through the program Nov. 21, 2017. As a student waiting to start class she uses some of her time to help the base team earn rewards in the Project Maven website. (U.S. Air Force photo by Airman 1st Class Seraiah Hines/Released)

GOODFELLOW AIR FORCE BASE, Texas --

At the beginning of November, Goodfellow started participating in Project Maven, a program created by the Secretary of Defense with the intent of helping artificial intelligence learn. By creating a profile, any military service member, or individual working for the Department of Defense, can help the program learn by labeling imagery and video, helping it distinguish particulars.

Phase one of this effort aims to integrate cutting edge computer vision algorithms into full-motion video processes. As a result, some of the brightest minds in the world are developing algorithms for us and our plan is to test them in the field before the end of this year, said Lt. Gen. VeraLinn Jamieson, Deputy Chief of Staff for Intelligence, Surveillance and Reconnaissance.

Phase two’s efforts include bringing the power of AI to bear on other aspects of sifting through information and narrowing down the collected data to be analyzed by the human eye who may then determine whether or not the collected data needs to be acted on or may be discarded.

 “Imagine trying to look for that needle in the proverbial haystack, but when you are overseas in a deployed area there might be 10,000 haystacks and you have to find those handful of needles,” said Senior Master Sgt. Kelly Bales, 17th Training Wing project manager. “Through this program, service members can provide an extra tool, an extra “Airman”, which can process and narrow down the data so the human eye can take a look at it, ’Saving manpower, saving lives’ that is the real point of this program,” said Bales.

A way to propel this effort forward, and have structured, tagged and labeled data to develop and field high fidelity algorithms is a friendly competition throughout the whole DoD continuing to the end of November. The race is to label 1 million objects, the prize, 25 thousand dollars.

 “We are working to train our new ‘Wingmen’, think of your own R2-D2, the algorithm, to join our ranks,” said Jamieson.  

“Between students waiting to start class and permanent party members, Goodfellow has around 350 active participants in the project,” said Bales. “By taking just a couple of minutes out of the day to complete different jobs within the website, individuals can take part in this program. Through a leaderboard within the site the base can track how many objects have been labeled and what the ranking it is within the DoD. It is more than just clicking on images, by participating in this program we are enabling opportunities to save lives.”

This program can be accessed from any computer or mobile device, just read the instructions on the SharePoint before accessing the website off base, said Bales.

If interested in participating individuals may contact Senior Master Sgt. Bales for more information at kelly.bales@us.af.mil.