Answer selection method and system for reading comprehension by reading robot

Abstract

The invention discloses an answer selection method and system for reading comprehension by a reading robot, which is designed to improve the correct rate of the answer selection of the robot for reading comprehension. The answer selection method for the reading comprehension of the reading robot of the present invention includes locating the paragraphs related to the question stem; using the bag-of-words model, extracting the sentence with the most matching words with the options as evidence materials; analyzing the semantic consistency between the options and the evidence materials; Based on the results of the consistency analysis, the best answer to the question is selected. The invention can be applied to reading robots, and realizes intelligent answering of answer selection questions through automatic analysis and reasoning.

Description

technical field [0001] The invention relates to the field of natural language processing research, in particular to an answer selection method and system for reading comprehension by a reading robot. [0002] technical background [0003] Reading comprehension is an important way for people to obtain information from a large amount of text. Many times, users do not want to browse the full text, but want to know the exact answer to a certain question. Reading comprehension question answering technology emerged as the times require. In the reading comprehension question and answer, there is a type of selective reading comprehension, that is, a question is asked based on the background material of reading comprehension, and multiple options are provided for the question, and the computer is required to automatically "read" a piece of material. Choose the correct answer from multiple candidate answers based on your "understanding" of the material. In the field of natural languag...

AI Technical Summary

Problems solved by technology

[0004] At present, for answer-choice questions, the method of determining the correct answer based on similarity or correlation is mainly used. This method determines the correct answer by calculating the most similarity or correlation between the options and the sentences of the background material. However, in the semantic The above-equivalent sentences are often expressed in different syntactic forms. The method based on similarity and correlation can only find sentences with a high similarity to the optional grammatical structure or semantic expression in the background material, and cannot understand the subtle differences in semantics. , and the nuances between sentences are the first priority of language processing

At the same time, this type of method analyzes the correlation between the options and the entire background material, but the options are often related to one or a few sentences in the article. Therefore, analyzing the options and the full text introduces a large amount of noise data, which affects the accuracy of the answer.

Images