Bloody Mahjong is a kind of mahjong. It is very popular in China in recent years. It not only has the characteristics of mahjong's conventional state space, huge hidden information, complicated rules, and large randomness of hand cards but also has special rules such as Change three, Hu must lack at least one suit , and Continue playing after Hu . These rules increase the difficulty of research. These special rules are used as the input of the deep learning DenseNet model. DenseNet is used to extract the Mahjong situation features. The learned features are used as the input of the classification algorithm XGBoost, and then the XGBoost algorithm is used to derive the card strategy. Experiments show that the fusion model of deep learning and XGBoost proposed in this paper has higher accuracy than the single model using only one of them in the case of high‐dimensional sparse features. In the case of fewer training rounds, accuracy of the model can still reach 83%. In the games against real people, it plays like human.

Doudizhu poker is a very popular and interesting national poker game in China, and now it has become a national competition in China. As this game is a typical example of incomplete information game problem, it has received more and more attention from artificial intelligence experts. This paper proposes a multirole modeling-based card-playing framework. This framework includes three parts: role modeling, cards carrying, and decision-making strategies. Role modeling learns different roles and behaviors by using a convolutional neural network. Cards carrying can calculate reasonable rules especially for “triplet” by using an evaluation algorithm. Decision making is for implementing different card strategies for different player roles. Experimental results showed that this card-playing framework makes playing decisions like human beings, and it can to some extent learn, collaborate, and reason when facing an incomplete information game problem. This framework won the runner-up in the 2018 China Computer Game Competition.

This paper has studied the game rules, victory or defeat rules, and the key techniques of Dots-and-Boxes, and has designed and realized one 6×6 Dots-and-Boxes based on the representing method of Strings and Coins. By corresponding Dots-and-Boxes chessboard to strings and coins chessboard, the proposed representing method is not only intuitive and convenient in chessboard representation, but also has other 3 merits which include: (1) can reduce the size of representing data largely, and simplify the chessboard stored matrix, thus easy to store and analysis the data, (2) help to judge the key chains and rings among elements of chessboard, and (3) help to calculate the number of chains, and apply the long-chain rule. In the 4th machine gambling tournament took in 2010, the realized Dots-and-Boxes system gained the second best, which indicates the feasibility of the proposed chessboard representing method and the validity of the Dots-and-Boxes system

In recent years, deep learning has developed rapidly and gradually infiltrated into various fields. As a rookie, generation‐based confrontation networks based on deep learning show excellent characteristics in many aspects. This study presents two innovative ideas, the same card force and generation cards algorithm. The generation of the equivalent cards force based on the generation of confrontation networks is studied. For the same period of time in the regular game, players will be issued different types of cards with similar card force, so as to distinguish the player level, the theory of the equal force is proposed. Based on the average and variance of scores obtained after the completion of a game, the cards are divided into ten different types of cards. On this basis, the use of generating a counterfeit network Generative Adversarial Nets generates a large number of cards of the equal card force. In the actual game, a network model is generated to generate a large number of game cards with the same force and distributed to different table numbers, so that the points scored by the landlords in different matches can have the characteristics of mutual appraisal.