检索规则说明:AND代表“并且”;OR代表“或者”;NOT代表“不包含”;(注意必须大写,运算符两边需空一格)
检 索 范 例 :范例一: (K=图书馆学 OR K=情报学) AND A=范并思 范例二:J=计算机应用与软件 AND (U=C++ OR U=Basic) NOT M=Visual
作 者:代鹏程 李淑琴[1,2] 郑蓝舟 孟坤 丁濛[1,2] DAI Pengcheng;LI Shuqin;ZHENG Lanzhou;MENG Kun;DING Meng(Beijing Information Science&Technology University,Beijing 100010,China;Perception and Computation Intelligence Joint Lab,Beijing 100010,China;WEIZHIYU(Beijing)Technology Co.,Ltd.,Beijing 100010,China)
机构地区:[1]北京信息科技大学,北京100010 [2]感知与计算智能联合实验室,北京100010 [3]微智娱(北京)科技有限公司,北京100010
出 处:《重庆理工大学学报(自然科学)》2021年第3期159-165,共7页Journal of Chongqing University of Technology:Natural Science
基 金:北京信息科技大学科技项目(5212010937,KM201911232002,5112011019,5112011041)。
摘 要:基于深度学习模型的有监督训练依赖于大量高质量标定数据,针对非完全信息博弈中二打一智力游戏问题,根据不同阶段回合局面数据的特点,提出了通过Alpha-Beta完全搜索获得共包含400万带标定二打一智力游戏局面样本的数据集,根据得到的标定样本训练CNN模型,使其能够对二打一智力游戏残局进行局面评估,为进一步将牌类游戏向棋类游戏的转化提供了保障,也为其他非完全信息博弈训练数据的标定提供了有价值的借鉴。The success of AlphaGo has made the deep learning method widely concerned in the field of computer games.However,in the incomplete information game,since the game participants only have private information and cannot obtain all the state information of the current situation,it is difficult to make a reasonable evaluation of the game situation.How to apply some methods in the field of complete information game to the field of incomplete information game is one of the research hotspots in the industry.Supervised training based on deep learning models relies on a large amount of high quality calibration data.Aiming at the problem of Two-on-One intelligence game in incomplete information game,according to the characteristics of round data in different stages,this paper proposes a complete search of alpha beta,and obtains a total of 4 million data sets with the situation samples of the Two-on-One intelligence games.According to the obtained calibration sample,CNN model is trained so that it can evaluate the situation of the landlords’endgame,and provide a guarantee for further conversion of card games to board games.This paper also provides valuable reference for the data calibration of other incomplete information games.
关 键 词:数据标定 二打一智力游戏 局面评估 计算机博弈 非完全信息
分 类 号:TP18[自动化与计算机技术—控制理论与控制工程]
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在链接到云南高校图书馆文献保障联盟下载...
云南高校图书馆联盟文献共享服务平台 版权所有©
您的IP:216.73.216.200