Soldiers face life and death situations every day, but the data scientists at Booz Allen Hamilton are determined to reduce the risks with the help of AI.
“The days of us sending soldiers forward into the field alone are over,” said Aaron Sant-Miller, lead data scientist at Booz Allen Hamilton, a management and consulting firm.
In a conversation with AI Podcast host Noah Kravitz, Sant-Miller explained how his team was working with soldiers to gather data by using at the edge devices to help troops analyze potential signs of combat.
“We have so much technology on our soldiers that our soldiers are bringing out to help them, and we have a lot of overhead video feeds that’s linked backed to analysts who are fusing intelligence to help our soldiers know what’s going on before, after, and during a mission,” Sant-Miller said.
One challenge that Sant-Miller’s team faces is adapting their machine learning models to be able to function in a low-connectivity environment. A military environment won’t always have the best connectivity. So the deployed devices need to be able to learn off their data.
But, according to Sant-Miller, their biggest challenge is trust.
“How do we get people to sign off on these algorithms? How do we get people to trust these algorithms?” Sant-Miller asked.
A graduate of the University of Notre Dame, Sant-Miller was drawn to Booz Allen and this specific area of work in AI due to its high impact, fast-moving nature.
But while diving deeper into AI and its different applications, Sant-Miller discovered the importance of ensuring that you also have the perspectives from those in the industry that AI would impact.
“It’s not just a math problem, it’s not just a compute problem, it’s not just a data science or machine learning problem,” Sant-Miller said. “You have to have people who are experts in this specific domain that you’re applying these things to fully embedded in the process.”