Communications of the ACM
Computer Algorithm Automatically Recognizes Soccer Formations and Defensive Strategies
A new deep-learning algorithm automatically recognizes the formations of soccer teams when analyzing player tracking data.
Credit: www.sportsessionplanner.com
Researchers at Disney Research, the California Institute of Technology, and STATS have developed a deep-learning algorithm that automatically recognizes formations of teams when analyzing player tracking data.
"This is the first time an imitation learning approach has been applied to jointly learn cooperative multi-agent policies at large scale," says Disney's Peter Carr. He says this enables a computer to better analyze how each athlete plays in any team sport by understanding how players coordinate with each other to change roles.
The technology also can be applied to robot movement, autonomous vehicle planning, and modeling of collective animal behavior, says Disney's Markus Gross.
The researchers tested the algorithm on data from 45 games of European professional soccer teams and found it significantly outperformed conventional imitation learning methods.
The team also ran experiments on a predator-prey simulation game, in which four predators must coordinate their actions to capture one prey in the least amount of time.
From EurekAlert
View Full Article
Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA
No entries found