基于桥牌叫牌体系的模仿学习模型研究
首发时间:2019-11-19
摘要:桥牌拥有相比德州扑克大得多的隐藏信息,又包含4个玩家之间的分工与合作,是棋牌类游戏中最为复杂的游戏之一。本文针对桥牌叫牌阶段所使用的自然叫牌体系,提出了使用模仿学习的技术进行研究,并且根据桥牌叫牌体系的特点,设计神经网络模型,让模型能够输出更多有助于局面理解的特征。通过学习大量的自然叫牌体系下的专家经验数据,训练出一个自然叫牌体系下的桥牌理解模型。该模型不仅能够预测合理的叫牌,还能够根据输出当前体系下对局面的理解。
For information in English, please click here
Research on Imitation Learning Model of Bridge Bidding System
Abstract:Bridge has a much larger hidden information than Texas Hold\'em and includes a division of labor and cooperation between the four players. It is one of the most complicated games in board games. This paper proposes the use of imitation learning technology for the natural bidding system used in the bridge bidding stage, and according to the characteristics of the bridge bidding system, the design makes the model output more features that contribute to the understanding of the situation. By learning a large number of expert experience data under the natural bidding system, a bridge understanding model under the natural bidding system is trained well and the model not only predicts reasonable bidding, but also provides an understanding of the situation based on the current system.
Keywords: artificial intelligence bridge game bidding system
基金:
引用
No.****
动态公开评议
共计0人参与
勘误表
基于桥牌叫牌体系的模仿学习模型研究
评论
全部评论0/1000