Method and device for mining entity-hypernym relations

Abstract

The embodiment of the present application discloses a method and a device for mining entity-subordinate relations. An embodiment of the method includes: obtaining a question-answer text pair to be mined, the question-answer text pair to be mined includes an entity question text and a candidate answer text to obtain an answer to the entity question therefrom, wherein the entity question text includes a question intention; The question-and-answer text pair input the pre-acquired entity question-answering model to obtain the answer to the entity question, wherein the entity question-answering model is used to determine the answer to the entity question from the candidate answer text; the question intention is used as the hypernym of the answer to the entity question. The scheme of the present application can determine a pair of hyponyms, the question intention and the answer word of the substantive question, from the question-answer text pair to be mined. Further, the hypernym can be further mined from the entity question text of the question-answer text to be mined and the hyponym of the hypernym can be mined from the candidate answer text.

Description

technical field [0001] The embodiments of the present application relate to the field of computers, in particular to the field of natural language processing, and in particular to methods and devices for mining entity-hypernym relationships. Background technique [0002] The hypernym of an entity is a generalization of a class of entities with common characteristics. For example, the hypernym of "red", "yellow", and "green" is "color", and the hypernym of "car" and "truck" is "car" . [0003] In practice, it can be noticed that the entity-hypernym system has great significance for answer recognition in entity question answering. Specifically, having this hyponymy relationship as prior knowledge helps to quickly narrow down the range of entity answers, and helps entity question answering algorithms to quickly and accurately find out the answer. [0004] For example, for the question "what color is the sky?", in the process of finding an answer, it is usually first analyzed ...

AI Technical Summary

Problems solved by technology

However, there are usually a large number of noun entities in the article, and it is very difficult to directly locate the answer

Images