Abstract
Experiments on groups of Checkers programs, playing by majority voting, were performed to investigate performance and stability. Homogeneous groups, copies of the same program, was used to perform these experiments instead of heterogeneous group that was more complicated by factors of different programs. Experiments were performed based on a search-depth of 5, 10 and 12 using the Samuel checkers program. Games of checkers were played between groups of size ranging from 1 up to 10 for each side. Experiment results suggest that group performance increases as a kind of logarithm function as the group size gradually increases for stronger player, and the performance slowly decreases in the case of a weaker player, stability seems to increase as the group size increases.
Highlights
Studies on group benefit and performance have been conducted since 1898 (Triplett, 1898)
The advantage of a player depends on search depth, as shown in ‘An analysis of majority voting in homogeneous groups for checkers: Understanding group performance through unbalance’ (Carvalho, Nguyen & Iida, 2017), which shows the winning-rate of players in each depth and color
In summary, experimental results, of majority voting in Checkers, suggest that group performance increases as a kind of logarithm function as the group size gradually increases for stronger player, and the performance slowly decreases in the case of a weaker player
Summary
Studies on group benefit and performance have been conducted since 1898 (Triplett, 1898). Good opening play in board games is quite important for developing superhuman AI as well as understanding the nature of opening book (Muangkasem, Iida & Spoerer, 2013). Another example is the use of game elements in the educational context (Huynh & Iida, 2017). A new trend of AI research using games is the notion of mixture-of-experts in the deep-learning context. It relates to a resarch idea known as group peformance in game playing on which we concentrate in this paper. The idea is that the probability of the majority making the same mistake is lower than the probability of an individual making the mistake, and therein lies the strength of the method
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: Asia-Pacific Journal of Information Technology & Multimedia
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
