Research On Military Regulations Knowledge Graph Construction Technology

Military regulations are regulatory and guiding documents that guide the training and operations of Military.These regulations provide a common language and common reference system for command,operations,and training personnels at all levels to communicate as required by missions.These regulations can reflect the combat thinkings and movement patterns of an army.According to the information we have reviewed,at present,no researchers have applied intelligent entity extraction and relationship discovery to assit the manual construction of a knowledge graph with reasoning and expansibility targeting military regulations.This essay takes the construction of the US Army Joint Operations Task List(hereinafter referred to as the "Task List")knowledge graph as an example,and builds the ontology model,combined with natural language processing,knowledge representation and other technologies to demonstrate the method and process of constructing military knowledge graph.This eassy also formally describe the entities and relationships in them,so that the text of the military regulations can be calculated,understood and evaluated by machines in order to lay the technical foundation for the future of intelligent question and answer and decision-making systems based on military regulations.Knowledge graph construction involves many technical fields such as natural language processing and knowledge representation.The innovative research work of this essay mainly includes the following contents:(1)Establishing the technical framework for the construction of the knowledge graph of military regulations.Military regulations involve a large amount of knowledge in the military field.The contents of the different regulations,their fields of application and functions are different.Based on the content and functional characteristics of the military regulations,this eassy describes the technologies involved in the process of constructing their corresponding knowledge graphs,and builds the knowledge graph of the task list as a case to instantiate the practicality and expandability of the construction technology framework.(2)Developing the ontology model of the Task ListDue to the wide distribution of the knowledge involved in the “Task List”,the description methods are diverse and need to be standardized.According to the description and definition of tasks in the text of the Task List,and the application background of the knowledge graph,this essay firstly describes the joint tasks in a formalized way.According to the characteristics of the task description and the characteristics of all the entities and information of the text,the ontology model of the Task List is developed,and all classes and relationships in the Task List are defined.(3)Proposing a knowledge representation model improved by multivariate information.This study transforms the rules between the types and relations of entities and semantic correlativities into prior distributions.And projects the entities and relations in triples into vector spaces by the translation model in representation learning.Link prediction experiments were carried out on the dataset of this study.Compared with other transformation-based knowledge representation models,the proposed method has better accuracy on small sample datasets.