摘要
在有多轮次、状态空间巨大、81种不同类别的番种、胡牌方式复杂的国标麻将中,普通的神经网络难以对复杂的数据进行表达和拟合。首次将多尺度骨干的深度神经网络用于实现麻将AI,以更好地捕获国标麻将的局部以及全局特征,适用于处理复杂数据,做出更准确的游戏策略。基于IJCAI 2020 Champion的对局数据,对训练数据进行数据增强。采用增强后的数据,在NVIDAI GeForce RTX3090 LapTop GPU上进行了5天的监督学习训练,训练出的模型有52 M参数,动作准确率达到93.47%,弃牌准确率达到83.93%,鸣牌准确率达到97.56%。将提出的模型部署到北京大学开发的Botzone平台上,进入天梯榜前1%。
Chinese standard mahjong involves multiple rounds,immense state space,81 different categories of tiles,and complex winning strategies.Conventional neural networks struggle to express and fit the intricate data of Chinese standard mahjong.For the first time,a multi-scale backbone deep neural network is employed to build a mahjong AI algorithm,to better capture local and global features of national standard mahjong,suitable for processing complex data and making more accurate game strategies.Based on the game data from the IJCAI 2020 Championship,the training dataset undergoes data augmentation.Using the augmented data,the proposed algorithm receives 5 days of supervised learning training on an NVIDIA GeForce RTX 3090 Laptop GPU.The trained model has 52 million parameters,achieving an action accuracy of 93.47%,discard accuracy of 88.93%,and declaration accuracy of 97.56%.The proposed model is deployed on Botzone platform developed by Peking University and it has entered the top 1%of the leaderboard.
作者
代君学
李霞丽
刘博
王昭琦
DAI Junxue;LI Xiali;LIU Bo;WANG Zhaoqi(Key Laboratory of Ethnic Language Intelligent Analysis and Security Governance of MOE,Minzu University of China,Beijing 100081,China;School of Information Engineering,Minzu University of China,Beijing 100081,China)
出处
《重庆理工大学学报(自然科学)》
CAS
北大核心
2024年第5期137-144,共8页
Journal of Chongqing University of Technology:Natural Science
基金
国家自然科学基金项目(62276285,62236011)。