Natural language question-answering method and device

Zou Lei,Liu Tingting, Lu Yanxiong, Liu Huaijun,Huang Ruizhe

user-60ab1d9b4c775e04970067d6(2021)

引用 0|浏览10
暂无评分
摘要
The embodiment of the invention discloses a natural language question-answering method and device and belongs to the field of information retrieval and processing. The natural language question-answering method comprises the steps that a natural language question N is obtained; the natural language question N is converted to an inquiry semantic graph Qs, wherein each edge in the inquiry semantic graph Qs represents a semantic relation in the natural language question N; a subgraph matched with the inquiry semantic graph Qs is found in a RDF graph G; according to the subgraph, the answer of the natural language question N is obtained. The natural language question-answering method and device solve the problem that in the stage of understanding a question, a large amount of calculation amount is consumed to eliminate ambiguity and achieve a brand new natural question-answering mechanism. When the natural language question N is converted to the inquiry semantic graph Qs, elimination of ambiguity is not needed, and ambiguity is naturally completed in the process of searching for the subgraph matched with the inquiry semantic graph Qs.
更多
查看译文
关键词
Natural language,Graph (abstract data type),Ambiguity,Natural language processing,Computer science,Artificial intelligence,Natural language question answering,Rdf graph,Semantic relation
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要