Design And Implementation Of Military Knowledge Question Answering System Based On Knowledge Graph

With the development of Internet,big data and artificial intelligence technology,the full use and mining of the connection between information and prior knowledge for more research and application is an important research hotspot at present.The graph neural networks,graph theory,and probabilistic graph models are applied to knowledge graphs.Compared with traditional search engines,the question answering system can grasp the user’s intention more accurately when analyzing user questions.A question answering system based on knowledge graph is designed and implemented for the field of military data to promote the transformation of future information warfare.The main contents and innovations of this thesis are listed as follows:A model of entity-relation joint extraction that integrates knowledge graph information is designed and implemented to solve the problem of insufficient utilization of knowledge graph information.After the relation is extracted,the TransE model is applied to the vector representation of the relation and Bert model is applied to vector representation of questions.The vector representations are fed into an attention pooling layer and the triples are decoded in parallel.Experiments show the effectiveness of the information fusion method.A model for constructing knowledge graph subgraphs that may contain answers is designed and implemented to solve the incompleteness of knowledge graph information.A Bert-based named entity recognition model is applied for entity recognition in questions,and the neighbor nodes related to the question are selected as subgraphs of the knowledge graph.A document relevance calculation method based on Bert encoder,LSTM and self-attention mechanism is designed and applied to find out the supplementary knowledge of the knowledge graph.Answers are extracted from knowledge graph subgraphs.Experiments show the effectiveness of this method.The ontology layer is designed under the guidance of expert knowledge and is applied to construct public knowledge graphs in the military field.Equipment is associated with actions,and the construction results are used to promote the digital transformation of the military field and provide data and technical foundations.The main applications of the system are ontology layer editing,knowledge extraction,and knowledge question answering.