🚨#AIIDE25 full paper deadline extended! Due to requests from authors, we have extended the deadline for full paper submission to June 27, 2025! 🔗 More information: sites.google.com/ualberta.ca/ai…

Get your demos, prototypes and case studies ready because the #AIIDE25 case studies deadline is coming up in a week (EOD July 9th)! sites.google.com/ualberta.ca/ai…

🏆 Best Paper – 2010 Realistic Fireteam Movement in Urban Environments by Darken et al. The authors introduce novel approaches for a realistic simulation of visual scanning and coordinated movement of infantry units in a 3D urban environment.

🏆 Best Paper – 2011 Goal Recognition with Markov Logic Networks for Player-Adaptive Games by Ha et al. Sometimes players may have ill-defined goals and do unrelated exploring, so inferring their goal becomes tricky. Ha et al. explore using MLNs to learn player goals from actions

🏆 Best Paper – 2012 POMCoP: Belief Space Planning for Sidekicks in Cooperative Games by Macindoe et al. The authors present POMCoP, a system for online belief space planning suitable for creating sidekicks for cooperative pursuit games.

🏆 Best Paper – 2013 The Combinatorial Multi-Armed Bandit Problem and its Application to Real-Time Strategy Games by Santiago Ontañón The branching factor of some games makes MCTS unsuitable - until you try some of these alternative sampling strategies!

🏆 Best Paper–2014 Spice It Up! Enriching Open World NPC Simulation Using Constraint Satisfaction By Černý et al. Meaningful NPC interactions make games feel alive! This paper presents a constraint satisfaction method to apply pre-scripted interactions in open-world environments.

🏆 Best Paper – 2015 Predicting Purchase Decisions in Mobile Free-to-Play Games by Sifa et al. Understanding who is most likely to make purchases in a freemium mobile game is important for many reasons, and Sifa et al. proposed the first regression model to address this problem.

🏆 Best Paper – 2016 Combining Gameplay Data with Monte CarloTree Search to Emulate Human Play by Devlin et al. In collaboration with game developers, the authors improve MCTS by biasing decisions based on player data, making the agent more human-like without losing performance.