Research And Application Of Hotspot Discovery Technology Based On Knowledge Graph

With the increasing number of scientific publications,it may take a long time to digest a scientific paper,which poses a great challenge to the number of papers that people can read.With the rapid development of artificial intelligence technology,Knowledge graphs are more widely used in various fields.In this paper,the graphical knowledge structure of the knowledge graph is transformed into a tree structure with a hierarchical system by using the association between entities in the knowledge graph and the pruning algorithm of the graph,and the statistical scientific knowledge is relied on to mine the hidden information on the tree structure to find the technology hotspots that exist on the tree.This can help scientific researchers quickly understand the hotspot technologies related to a certain professional field.In response to the above problems,this paper constructs a knowledge graph in the field of artificial intelligence based on the abstract data of the field of artificial intelligence,studies the generation method of the tree structure of the conceptual system,and realizes the function of discovering technology hotspot nodes in the tree structure.The main work is as follows:(1)Crawler technology is used to crawl the abstracts of related papers in the field of artificial intelligence and process the data.Based on this data,an end-to-end joint entity relation extraction model based on pre-training model is used to complete two tasks: entity recognition and relation extraction.(2)Then,the lack of triples in the knowledge graph is studied,and the method of knowledge representation is proposed to complete the triples by link prediction.The nodes in the knowledge graph are studied by representation learning,and the related experiments of map completion are carried out.(3)To address the problem that the graphical structure of the knowledge graph is too complex to discover hotspot information,we study how to use the information on the constructed knowledge graph to generate a conceptual spanning tree with a good hierarchical structure.(4)For the concept system spanning tree,the hot entities in the tree knowledge structure are discovered by relying on the corresponding analysis indexes,and the current hot technologies are found to complete the hot spot discovery function.(5)Finally,this paper uses graph database to store data.Based on the constructed knowledge graph of AI domain,four functional modules are completed: entity relationship recognition,entity query,relationship retrieval and hotspot discovery.Through this system,users can more convenient,faster,more comprehensive understanding of the field of artificial intelligence related information.