面向天域感知领域的知识图谱构建技术研究

Research on Space Domain Awareness Domain-specific Knowledge Graph Construction Technology

天域感知是太空能力的基础和关键,各军事大国均竞相发展天域感知能力。随着太空目标的迅速增多和感知技术的快速进步,获取的天域感知数据呈指数级增长,实体关联关系高度复杂化。传统的知识获取和表述方式难以满足天域感知领域应用需求。知识图谱是基于图的知识网络库,具备强大的语义处理能力和开放组织能力。文中在天域感知领域引入知识图谱技术,研究自动化天域感知知识图谱技术架构和构建流程,提出基于本体和三元组模型的领域知识图谱建模技术以及基于RoBERTa-BiLSTM-CRF的领域知识抽取技术,为天域感知上层应用赋能赋慧。

Space Domain Awareness(SDA)is the foundation and key to space capabilities.Major military powers are competing to develop their space domain awareness capabilities.With the rapid increase of space targets and the rapid progress of awareness technology,space domain awareness data obtained shows exponential growth,and entity association relationships are becoming highly complex.Traditional knowledge acquisition and expression methods are difficult to meet the application requirements in the field of space domain awareness.The knowledge graph is a graph based knowledge network base,and has strong semantic processing capability and open organization capability.This paper introduces knowledge graph technology in the field of SDA domain,researches on the technology architecture and construction process of SDA domain-specific knowledge graph with automation,proposes the domain knowledge graph modeling technology base on ontology and triple model and domain knowledge extraction technology based on RoBERTa-BiLSTM-CRF,and endows the power and the wisdom for SDA upper level application.