人工知能学会研究会資料 SIG-KBS-B508-09 Analysis of Voting Behavior in One Night Werewolf 1 2 Ema Nishizaki 1 Tomonobu Ozaki 2 1 1 Graduate School of Integrated Basic Sciences, Nihon University 2 2 College of Humanities and Sciences, Nihon University Abstract: The werewolf game is a multiplayer party game, and it is recognized widely as a new standard problem for artificial intelligence. One promising research strategy towards the realization of the intelligent artificial agents for the werewolf games is to reveal useful insight and knowledge of human players behaviors by discovering distinctive characteristic and tendency from large volume of human play logs. In this paper, we focus on a simplified variant of werewolf game named One Night Werewolf, and prepare the play log data of hundred games by ten human players. Intensive analyses are applied to the log data from the aspect of voting behaviors. They include comparisons of the distribution of voting destinations, analyses of the reasons for voting, voting network analyses, and vote prediction by machine learning. Through the analyses, we discuss characteristic behaviors on the vote in One Night Werewolf. 1 [1, 2, 3] 1 BBS 1 1 10 Chittaranjan [4] 156-8550 3-25-40 E-mail: tozaki@chs.nihon-u.ac.jp 1 http://www.wolfg.x0.com/ [5] BBS BBS [6] [7] 3 3 [8] [9, 10] [11, 12] - 52 -
10 100 3 4 5 6 2 2.1 2 1 1 2.2 2 10 5 2 5 10 100 2 1 1 2 1. 2. 5 3. 1 4. 1 5., CO 6. 2.3 43% 57 43 BBS 1571 38.2%[?] 5% 300 188 966 186 1 1 7 2 1 2 1 1-53 -
1: 0.41 0.50 0.50 2: 100 0.13 0.17 0.20 0.44 0.02 0.00 0.04 0.22 0.11 0.06 0.00 0.27 0.72 0.75 0.74 0.05 3 1 2 1 10% (1) 100 (2) (3) 3 3.1 100 2 100 70% 1 10% 44% 22% 27% 3: 0.73 0.69 0.001 13 2 1 0 2 0.46 0.54 2 21 2 0 1 2 0.38 0.62 3 37 2 1 1 1 0.56 0.44 4 29 1 1 1 2 0.76 0.24 4: 0.15 0.31 0.69 0.12 0.00 0.31 0.24 0.00 0.40 0.19 0.29 0.48 0.09 0.05 0.00 0.22 0.15 0.24 0.27 0.51 0.01 0.00 0.03 0.27 0.57 0.71 0.12 0.74 0.70 0.70 0.00 0.10 0.10 0.00 0.34 0.03 0.00 0.07 0.26 0.00 0.00 0.10 0.33 0.86 0.90 0.83 0.07 3.2 100 4 3 3 4 1 2 3 1 3 2 4 14 4 1 2 2 2 1 1 2 30% 70% 1 73% - 54 -
5: 0.05 0.04 0.06 0.32 0.00 0.00 0.04 0.27 0.13 0.06 0.00 0.33 0.81 0.90 0.90 0.08 0.23 0.39 0.41 0.60 0.05 0.00 0.05 0.17 0.10 0.06 0.00 0.21 0.63 0.55 0.54 0.03 2 57% 2 70% 2 4 20 30% 70% 4 26% 33% 34% 3.3 5 1 1.8 2 100 2 6: 92 88 61 53 38 38 CO 26 2 20 19 17 17 3 28 4 500 6 13 4.1 100 92 20% 92 77 6 9 79 35 20% - 55 -
2 88 9 9 1 1 16 3 BBS 3 4 61 53 21 20 10 40 11 2 3.3 163 25% 4.2 2 3 1 2 63 42 20 9 13 2 30% 21 13 9 9 5 59 6 1 7 0 4 1 4 1 1 1 2 2 1 2 2 1 1-56 -
1 2 3 4 5 6 7 6 5 3 3 3 1: 7: 4 1 33 15 19 14 3 2 31 16 20 11 2 2 1 19 16 7 12 3 1 1 13 8 7 6 2 1 1 1 4 4 3 1 5.1 1 1, 3, 5, 6 3 2 1 1 1 1, 5, 6 4 3 4 5.2 2 4 1 3 2 4 1 3 2 2 4 110 7 3 28 8 3 2 1 3 2 1 8 3 2 6 1 3 37 4 13 2 21 57% 1 4 1 6 3 2 11 10 3 2 2 1 2 11 3 1 4 8 6 3 1 1 2 7 5 2 1 1 1 3 1 1 2 1 1 1 17 6 500 100 5 T 241 F 259-57 -
15 co ord 2 {1,2,3,4,5} co role{ } CO {True, False} agrdisagrest vstw qst{none, q 0, q 1,, q 8 } none q 0 q 1 q 2 q 3 q 4 q 5 CO q 6 q 7 q 8 2 F 0.519 0.574 16 6 21 co ordco co ord CO 1 CO 3 CO CO 2 CO CO 4 CO 7 100 BBS BBS BBS BBS [1] (2016) [2] AI (2017) - 58 -
2: [3] 28 2C4-OS-22a-3 (2014) [4] G. Chittaranjan and H. Hung : Are you Awerewolf? Detecting deceptive roles and outcomes in a conversational role-playing game, 2010 IEEE International Conference on Acoustics, Speech and Signal Processing, pp.5334-5337 (2010) 2016-GI-35(12), pp.1-6 (2016) [12] 31 1N1-1in2 (2017) [5] 28 1E4-OS-23a-1 (2014) [6] 29 1F2-1 (2015) [7] 30 2F4-3 (2016) [8] HAI 2013P19 (2013) [9] 2015-GI-33(18), pp.1-5 (2015) [10] 2012 2012(6), pp.144-147 (2012) [11] - 59 -