Natural language question answering method and device based on knowledge graph

Abstract

The invention discloses a natural language question and answer method and device based on a knowledge graph. The method comprises the steps of obtaining an entity name and a relationship path name corresponding to at least one entity reference text in an input natural query statement and at least one relationship reference text between the entity reference texts in a preset knowledge graph; and according to the at least one entity name and the at least one relationship path name, obtaining a relationship graph corresponding to the natural query statement, and querying the obtained relationshipgraph to obtain a question and answer result of the natural query statement. According to the method, the relation graph corresponding to the natural query statement is constructed based on the knowledge graph, the dependence degree on the syntactic relation in the natural query statement is greatly reduced, the accuracy of constructing the relation graph is improved, and therefore the accuracy of questioning and answering and the generalization ability of the natural language are improved.

Description

technical field [0001] The present application relates to the technical field of natural language processing, in particular to a method and device for natural language question answering based on knowledge graphs. Background technique [0002] In the context of the large-scale emergence of artificial intelligence technology and applications, a large number of knowledge map data has emerged, providing a strong driving force for the development of artificial intelligence technology. How to make full use of the large-scale high-quality structured data of knowledge graph is the focus of many artificial intelligence application research. The question answering technology based on the knowledge map can use the knowledge in the knowledge map to answer questions raised in natural language, and can be used in scenarios such as intelligent customer service and frequently asked questions. How to convert natural language questions into knowledge graph query language and answer question...

AI Technical Summary

Problems solved by technology

[0007] However, for the rule-based question answering method, this method requires a lot of manpower to write templates, and because it is difficult to exhaust all possible question answer templates by manpower, such methods usually can only answer a small number of question sentence patterns, and the generalization ability is poor. ;

[0008] For the question-and-answer method based on semantic analysis, since this method needs to construct query sentences through a syntactic tree, it relies heavily on syntactic analysis. Due to the complexity of natural language itself, syntactic analysis will make mistakes in sentences with slightly complicated expressions. Based on the wrong syntax tree, it is very likely to construct the wrong query statement, so the correct answer cannot be obtained, resulting in a lower correct rate of question and answer;

[0009] For the end-to-end question answering method, the deep learning model in this method is difficult to obtain due to the need for a large amount of training corpus, and the current related technical research based on the deep learning model has a low question answering accuracy rate

Images