Research On Pattern Extraction Of Tibetan Go And Development Of Multimedia Courseware

Tibetan Go is a traditional game of Tibetan nationality,which is mainly spread in Tibet and ten Tibetan areas.The game rules of Tibetan Go are similar to Go,but have unique characteristic,which is no less difficult than Go.In 2016,Google Alpha Go defeated human superior,which is not only the milestone of artificial intelligence(AI)development,but also caused the world to pay close attention to Go and its cultural development.Unlike Go,Tibetan Go is on the verge of extinction,the study of game is still in its infancy,the information,digitization and intelligence of Tibetan Go are still relatively backward,a grave shortage of Tibetan Go chess manual,unable to provide data basis for study of Tibetan Go game.Therefore,based on Bayesian Network,this paper carries out the mode extraction and transformation research of Tibetan Go,and establishes the mode library,which provides data support for the research of Tibetan Go game.The research on the game of Tibetan Go and the modern education technology,not only in AI and cognitive science,but also in promoting the cultural protection and inheritance of Tibetan Go is of great significance.Unlike Go,Chess and other board games,the education teaching resources of Tibetan Go are scarce.And education on campus is one of the important ways of cultural protection and inheritance.Research for middle schools of Tibetan Go modern education teaching technology,on the basis of national characteristics game to carry out AI based knowledge education,can effectively cultivate students' interest in learning,and improve the students' study enthusiasm.This article provides technical and educational support for the digitization and informatization of Tibetan Go,it is of great significance to protect and inheritance the national treasure of Tibetan Go.The main work is as follows:1.Organize the Tibetan Go chess manual collected during the preliminary investigation,combined with ways of Go walking and expert guidance,to extracted the common modes of Tibetan Go,and constructed the Bayesian Network topology relationship between Tibetan Go modes.Digitizing the Tibetan Go chess manual and counting the corresponding node(mode)probability in the network structure.uses the parameter learning algorithm——the ELM algorithm and EM algorithm to train the network to get the a posterior probability of corresponding nodes,to realize the mode extraction of Tibetan Go based on Bayesian Network.When the mode recognition engine in the game process recognizes that there exist multiple modes in the mode library to match the current state of the game,this probabilistic estimate can be used to guide the mode selection of the mode library.2.Extraction of about one thousand commonly used basic mode,these modes are divided into four categories:Tibetan Go modes,Gaomu modes,Xiaomu modes,Muwai modes.Equivalent transformation of these common basic modes,each of which has fifteen equivalent transformation modes,three kinds of rotation transformation,four symmetric transformation and eight colors.Common basic modes and equivalent transformation modes form a modes library.In Tibetan Go man-machine game process,the machine can use the modes library,the research results based on Bayesian Network mode are used to guide the selection and decision of the mode.In the case of extreme shortage of current samples,which has strong guidance and reference significance for the game practice of Tibetan Go computer.This part implements an extended mode library and mode coding,selection and matching based mode library.3.Guided by the Ministry of Education's advocacy of AI into the classroom and constructivism,using the flash animation design software and Camtasia Studio,premiere multimedia editing software,design and develop a set of Tibetan Go multimedia courseware,which is suitable for the teaching of AI in Tibetan middle schools.The set of courseware has a total of eight classes,the first half introduces the basic knowledge of Tibetan Go,and the second part introduces the AI ideas in Tibetan Go game software——Basis of mode recognition.This will make the AI classroom become more interesting and vivid,which is of great significance to the campus protection of Tibetan Go.This part designs and develops a set of Tibetan Go multimedia courseware,which is suitable for the teaching of AI in Tibetan middle schools.