Via Leidos

Leidos coders competed last week in the AI Palooza Champion’s Cup, an event that brought autonomous racing simulations onto a real track.

The competition was held in partnership with AWS DeepRacer, the world’s first global artificial intelligence (AI) racing league. In preparation for the event, four Leidos teams qualified for the league’s Pro Division, which represents the top 10% of competitors worldwide.

Tom Boggs, a Leidos AI expert, said the competition is a way to encourage coders of all skill levels to participate in AI and machine learning.

“We wanted to do something that aligned with our enterprise-wide push for reinforcement learning in particular,” Boggs said. “DeepRacer was a natural fit because it provides a good platform to easily build, train and evaluate reinforcement learning models.”

Jim Carlini, Leidos Chief Technology Officer (CTO), said challenges like these are not only a great way to have fun learning, but will also have an important technical impact in the company’s future.

“A strategic focus for Leidos as part of our Trusted AI strategy, reinforcement learning is an area of machine learning that is focused on enabling autonomous agents and robots to simultaneously learn about their environments and make decisions without humans,” Carlini said. “As AI is further integrated into everyday devices and used within uncertain environments, we believe that reinforcement learning will enable us to deliver reliable and previously impossible solutions for our customers.”

The event, which took place at Leidos Global Headquarters in Reston, Va., was held in partnership with Amazon Web Services (AWS) and Intel Corporation.

“Nearly 100 Leidos employees have competed this racing season, representing groups across the company hailing from all over the U.S., the U.K. and Australia,” said Leidos AI expert Andrew Godbehere. “It’s a special networking and team-building opportunity that has galvanized and connected AI teams across Leidos.”