Research And Application Of Elastic Semantic Reasoning For Large-scale Knowledge Graph

As an important branch of artificial intelligence,Knowledge Graph can be seen as a carrier of human characterization and cognition of the world.it can model the entities in the world and their relationships with the formal descriptions to imitate the human to implement logical reasoning.The reasoning technique of knowledge graph is one of the core technology of knowledge graph.It represents the process of deriving unknown conclusions based on known facts.The generalized reasoning includes semantic query and semantic reasoning.The reasoning referred to in this paper is the generalized range of reasoning technology.However,with the rapid development of knowl-edge extraction technology and the continuous enrichment of source data,the scale of knowledge graph is growing rapidly.The current knowledge graph has been used to refer to various large-scale semantic knowledge bases.With the explosive growth of the size of the knowledge graph,knowl-edge graph has shown many new characteristics and challeges such as structure complexity,data dynamics,and real-time querying requirement.Faced with new features and challenges of knowl-edge graph,traditional semantic reasoning methods and tools failed to deal with the challenges,greatly limiting the further development of knowledge graph.The research on the elastic semantic reasoning methods for large-scale knowledge graph becomes an important research topic.Based on the reasoning challenges of large-scale knowledge graph,this paper studies some practical scientific problems of elastic reasoning for large-scale knowledge graph.The main con-tents of this research include the following aspects:(?)Aiming at the challenge of complex knowledge structure of large-scale knowledge graph,this paper studies a method:elastic semantic association reasoning approach based on OWL Property Chain over large-scale knowledge graphs.(?)Aiming at the challenge of real-time query requirement of large-scale knowledge graph,this paper studies a method:elastic semantic querying reasoning approach over large-scale knowl-edge graph based on distributed memory.(?)Aiming at the challenge of dynamics and streaming of large-scale knowledge graph,this paper studies a method:rule-based semantic streaming reasoning framework and approach.(?)Based on the above methods and experience,two practical flexible semantic reasoning sys-tems for large-scale knowledge graph are designed and implemented:a generalized large-scale knowledge graph rule-based reasoning system based on Spark and a domain-related association knowledge system of Chinese and Western medicine knowledge graph.