Army working on new cyber project

By Jaspreet Gill / October 29, 2019 at 11:13 AM

The Army Artificial Intelligence Task Force is working on a new project this year that will help inform the future of the cyber network.

Brig. Gen. Matt Easley, director of the A.I. Task Force, told Inside Defense in an interview last week the service is still in the "problem formulation stage" of the project, but it will address where vulnerabilities are present in the network. One of the things the task force will be doing is working with data scientists from Army Cyber Command to manage the service's cyber centers by making IT systems more efficient and creating information awareness tools.

"We can have machines do a task first [and] understand what the network looks like," Easley said. "Just having a good situational awareness of what the Army network looks like today and understanding which operating systems each of those systems are on . . . is another key area."

Easley added he's interested to see what industry partners have to say about artificial intelligence technology that will help shape the cyber project at an industry day slated for Nov. 13 in Silver Spring, MD. The event will include a discussion on artificial intelligence prototyping activities, including algorithms, user interface, user experience, training data, data and inference platforms, testing and integration, according to an Oct. 16 Federal Business Opportunities notice.

The notice states the Defense Department is interested in market research on artificial intelligence capabilities including facial recognition, language translation, object detection, tracking and more.

"It's interesting to see what the future is going to be for A.I. for 10 years out and to see some of those [capabilities]," Easley said. "And it's also good just to see who's been taking the current technology, being object recognition technology, self-driving car technology or predictive maintenance technology, and who's been able to use those techniques and show how they can scale these systems and apply them to real problems."

205324