Title: Multiagent Challenges in Team Sports Analytics
Abstract: This talk will present several challenges and opportunities within the area of team sports analytics, specifically related to active areas of research within multiagent systems and reinforcement learning. These connections are most prevalent in “invasion games” such as ice hockey, basketball, and soccer, defined as where teams must invade opponent territory to be successful. The talk will detail relevant problems along two axes, short-term strategy (coaching) and long-term planning (management), where problems of player/policy evaluation, strategy, and coalition formation are most present. I will discuss various examples of data with which these problems are typically addressed and show how insights from multiagent systems and reinforcement learning are particularly well suited for these tasks.
David Radke is a Senior Research Scientist for the Chicago Blackhawks of the National Hockey League (NHL) in the Hockey Strategy and Analytics group. His research lies at the intersection of invasion-game sports and artificial intelligence, specifically multiagent systems and reinforcement learning. He completed his PhD at the University of Waterloo with Dr. Kate Larson and Dr. Tim Brecht, titled "The Impact of Teams in Multiagent Systems."
Tiffany Leighton
Email