Amazon Web Services Inc. said today the National Football League will use its cloud platform to model contributing factors to player injuries and explore ways of reducing health risks using machine learning.

The partnership, announced at AWS’ annual re:Invent conference today in Las Vegas, expands upon the organizations’ existing multiyear relationship. Since 2017, the NFL has been using AWS to power its Next Gen Stats system, which provides football followers with statistics about matches and player performance.

The league is planning several new projects as part of the health initiative. Among them is The Digital Athlete, a simulation system that will enable the NFL to place virtual clones of players in various game scenarios to identify potential health risks. The platform will generate models by combining video footage from matches with data on player position, plays, equipment choice, playing surface and environmental factors, as well as anonymized information about player injuries.

The initial goal of The Digital Athlete is to help the NFL develop better ways to rehabilitate players who are hurt on the playing field. In the longer term, the league hopes to use the system to predict injury risks and intervene ahead of time.

NFL engineers will build The Digital Athlete in collaboration with AWS’ ML Solutions Lab consulting practice. The system is set to make use of the provider’s newly upgraded SageMaker machine learning development toolkit and the Rekognition computer vision service.

Jeff Crandall, chairman of the NFL’s Engineering Committee, said the idea with The Digital Athlete is to take a virtual representation of this population of athletes and do simulations to identify and predict injury risk scenarios and come up with innovations to reduce injuries. “We believe this will revolutionize the way we look at reducing player injuries,” he said.

The two machine learning offerings will also be used as building blocks for a second analytics system the NFL is building that will focus on detecting concussions. According to the league, the goal is to create a platform that can automatically spot head injuries when they occur on the field and identify the forces that cause them. The league hopes that the concussion identification techniques it will develop as part of the project will prove useful for speeding up the detection of other injuries as well, potentially even in areas outside football.  

“The outcomes of our collaboration with AWS – and what we will learn about the human body and how injuries happen – could reach far beyond football,” NFL Commissioner Roger Goodell (above) said in a statement.

Goodell added in a press conference today that he hopes the NFL will be able to reduce concussions even further as well as other, lower-extremity injuries. We can do that with this partnership to … make predictive analysis to determine if this (particular) athlete will be more prone to an injury,” he said.

With reporting from Robert Hof

Image: AWS