Dr Ed Powley (The Games Academy/Metamakers) has been awarded the Best Paper Award for his paper Memory Bounded Monte Carlo Tree Search at The AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE).
Written in collaboration with Peter Cowling and Daniel Whitehouse from the University of York, their paper describes how Monte Carlo Tree Search (MCTS) can be modified to run under restricted memory budgets. MCTS is a type of learning algorithm which leads to better AI decision making in a variety of games and real-world problems. The paper shows how this powerful algorithm can be implemented efficiently on limited hardware such as mobile phones.
The AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE) is sponsored by the Association for the Advancement of Artificial Intelligence (AAAI). AIIDE is one of the top international conferences in the field of artificial intelligence for games, and places a particular emphasis on bridging between the communities of academic research and commercial game development. This year’s AIIDE was the 13th annual conference, and brought together many of the world’s top game AI researchers with representatives from companies such as EA, Sony, Facebook and Google.