Army Capitalizing on Machine Learning

The U.S. Army is seeing success in implementing machine learning (ML)— evidenced by improved classification of previously unknown data by 20 percent— and is developing a workforce culture where artificial intelligence (AI) adoption is possible.

Speaking during the AI Experience 2020 Webcast, Deputy Assistant Secretary of Financial Information Management John Bergin spoke about how embracing shifting attitudes from top-level leadership, bringing the right team and tools together, and adopting a fail-fast mentality are vital to ML adoption.

“We have to ground ourselves in some common starting point,” Bergin said.

Bergin said that ML and reskilling the workforce to use AI work together to improve the Army’s classification of data, adding that it’s up to leadership to guide the workforce through the changes.

“Machine learning doesn’t work in a vacuum. You have to train the algorithm,” Bergin said, emphasizing the role humans play in adopting AI.

Going forward, the Army wants to identify opportunities where automation and machine learning can enable transformation. Among ways to do this, Bergin highlighted:

  • Understanding operating environments first;
  • Finding errors that current solutions can’t fix; and
  • Employing new methods by testing and analyzing.

ML adoption is an evolving process and as Bergin said: “we can’t let perfect be the enemy of great.” Adopting a fail-fast mentality and a willingness to be unsuccessful, but trying again is how leadership can usher in ML adoption.

Jordan Smith
About Jordan Smith
Jordan Smith is a MeriTalk Staff Reporter covering the intersection of government and technology.

Categories

Recent