CHICAGO– February 27, 2018 – STATS, the worldwide leader in sports data and intelligence, was recognized as the runner-up for the 2018 MIT Sloan Sports Analytics Conference research competition. STATS has become a keystone of the prestigious sports conference with seven research papers – pioneering the use artificial intelligence in sports – being selected as finalists over the last three years.
Of the eight finalists, “Bhostgusters: Realtime Interactive Play Sketching with Synthesized NBA Defenses” received a second-place finish. The paper was a collaboration with Dr. Patrick Lucey, Director of Artificial Intelligence and Data Science at STATS; Thomas Seidl, PhD student at the Technical University of Munich; Aditya Cherukumudi, Research Associate at Disney Research Pittsburgh; Andrew Hartnett, Research Associate at Disney Research Pittsburgh; and Peter Carr, Senior Staff Engineer at Argo AI, formerly with Disney Research.
“This second-place finish was the result of a fruitful collaboration between STATS, Disney Research and the Technical University of Munich,” said Dr. Patrick Lucey, Director of Artificial Intelligence and Data Science at STATS. “This also marks the third year in a row that STATS has been highly recognized in the Sloan paper track, which reaffirms that we are a leader in artificial intelligence and data science.”
The paper presents a new way for NBA coaches to communicate in-game strategy to their players through play-sketching from a data-driven perspective. By using a powerful analytics framework built on deep-imitation learning, the team created an intelligent and highly intuitive user interface.
In addition to the runner-up best research paper, STATS was also recognized with three of the 12 research paper poster finalists. They include: The Problem with Win Probability, Mythbusting Set- Pieces in Soccer and Quantifying the Value of Transition in Soccer via Spatiotemporal Trajectory Clustering.
For more information on Patrick Lucey and the Data Science and Artificial Intelligence team, visit: stats.com/data-science/