Knowledge Question Answering System In Artificial Intelligence Field Based On Knowledge Graph

Question-Answer System is an advanced form of information retrieval system.It can answer users’ questions in natural language with accurate and concise natural language.The main reason for its rise is people’s demand for fast and accurate information acquisition.Currently,the most popular technological routes are the relational database-based method,the standard question-and-answer pair-based method and the knowledge graph-based method.Compared with the relational database-based method,which has a slightly worse performance in multi-table federated scenarios,the standard template-based method requires a large amount of labor costs to build datasets.The knowledge graph-based method can effectively improve the search quality and accuracy,and has better performance than the other two technical routes,and does not need to build standard template question-and-answer pairs manually.It has received wide attention.At the same time,artificial intelligence technology as a hot topic in recent years,with its related data increasing,the difficulty of retrieving user target information is also increasing,therefore,this paper combines the project group’s work accumulation in knowledge graph,and develops a question and answer system based on knowledge graph,which relies on the field of artificial intelligence data.This paper focuses on the improvement of question parsing method based on knowledge graph,and puts forward a question parsing method that based pre-training model with knowledge graph to extract the central entity of the question and sort the subgraph paths.Compared with template-based question parsing,this method can reduce a lot of manual template building and has a higher accuracy.In this way,by combining knowledge graph technology with domain knowledge,domain knowledge questions and answers can be achieved,so as to facilitate the efficient use of current domain knowledge by users.Further,this paper takes the question parsing method proposed as the core,designed and implemented a question and answer system in the field of artificial intelligence based on knowledge graph by analyzing the actual demand scenarios,and optimized the clear,accurate and friendly visual user feedback.The system is divided into six modules: question answer retrieval,chart construction,result visualization,graph data management,synonym maintenance and user management.The Question Answer Retrieval module obtains the direct answer based on the method of question parsing.The result visualization module summarizes and lists the origin of the answer knowledge entries,the related entities and the related statements in the database that contain the text information of the question by visualizing the relationship,so that users can get the answer and related information that the question points to at a glance.The chart construction module is used to assist the visualization module to make the results more detailed.At the same time,in order to ensure the extensibility of the system data,this paper designs and implements data maintenance related modules such as graph data management,so that administrators can manage and maintain the data in the database.This paper structured a validation dataset for the field of artificial intelligence,compared and analysed the accuracy and response time of the question parsing method proposed in this paper,and tests the question and answer system implemented in this paper.By analyzing the experimental results,we can see that the question parsing method proposed in this paper achieves a higher accuracy with only a small increase in response time than the template-based parsing method.At present,this system has passed the test and has been delivered to the user.