Anomaly Detection Of Game Artificial Intelligence Based On Hierarchical Temporal Memory Model

With the rapid development of artificial intelligence technology,the artificial intelligence technology in game is constantly developing.Excellent game AI performance similar to human intelligence,bringing more challenges and fun for the players.Since the game AI is different from the traditional Artificial Intelligence,so in practical applications,game AI technology focuses differently.Game AI techniques in some way include perception,decision-making and actions.Firstly,the framework of the game is optimized on the basis of the general artificial intelligence architecture.After analyzing the game AI features of perception,decision and action algorithm technology,optimization and improvement are made in practical applications.Automatic combat for the complex roles in game is designed to enhance the performance.Secondly,the anomaly models of game artificial intelligence in auto find way are summarized and the feasibility and advantage of using hierarchical temporal memory algorithm to detect the anomaly is analyzed.Hierarchical Temporal Memory algorithm is proposed model from bionics and have many advantages compared to other neural network algorithm.The process of the recognition and learning anomaly pattern and self-adaptive of the hierarchical temporal memory algorithm are analyzed.The spatial pooling and temporal pooling are studied.Finally,in order to ensure the normal performance of the game artificial intelligence in the game world,the idea of using software to test the game artificial intelligence is presented.Firstly,the game development platform and the HTM algorithm development platform are analyzed,then the method of platform transformation is studied.The functional interface of the two platforms are analyzed to summarize the process mode of the anomaly detection.Secondly,the feasibility and adaptive advantage of hierarchical temporal memory algorithm to detect anomaly of in game AI are verified by experiments.