A visualization tool developed by PhD student Shivam Agarwal, which supports analysis of Pommerman games in a competition, was awarded with $1k in Google Credits. The competition was held as a workshop of the NeurIPS 2019 conference held December in Vancouver, Canada. The competition organizers commented: "We used this extensively to analyze the games and encourage everyone else to give it a try as well." (https://twitter.com/Pommerman/status/1206101858336395264)

The Pommerman game environment (https://www.pommerman.com/) is a variant of classic Bomberman games and serves as a testbed for artificial intelligence (AI). The goal is to train AI players who can co-operate with teammate while competing with the other team. Our visualization tool helps the developers of AI players to find the strategies executed by the players and explore their behavior from the recorded gameplay data. Benefits of the tool include: improving the trained AI players, understanding the underlying AI algorithm, and communicating the strategies learned by the players. 

The visualization tool was developed in collaboration with Günter Wallner who is an Assistant Professor at the Eindhoven University of Technology and Senior Scientist at the University of Applied Arts Vienna.

Check out the tool here: https://vis-tools.paluno.uni-due.de/pom/