Method and device for generating paragraph set and reasoning method and device

Abstract

The embodiment of the invention provides a method and device for generating a paragraph set, a reasoning method and device, computing equipment and a computer readable storage medium, so as to efficiently, accurately and comprehensively obtain evidence content sufficiently related to question content from a large amount of text data, avoid loss of valuable information and provide a high-quality data basis for subsequent answer prediction, so that the accuracy and efficiency of a reasoning process during subsequent answer prediction are improved. The method for generating the paragraph set comprises the following steps: obtaining a document set related to a question from a document database; constructing a paragraph graph based on the document set, wherein the paragraph graph comprises a plurality of paragraph levels, the paragraph levels comprise a plurality of paragraphs from the document set, and the paragraphs in the paragraph levels are related to the content of at least one paragraph in the next paragraph level; and obtaining all paragraphs in the paragraph graph to generate a paragraph set.

Description

technical field [0001] The present application relates to the field of computer technology, and in particular to a method and device for generating a paragraph set, a reasoning method and device, a computing device, and a computer-readable storage medium. Background technique [0002] Sentence relationship judgment is one of the four major tasks in the field of natural language processing, which includes reading comprehension question answering. In real-world scenarios, answers to many questions are distributed across multiple texts, and sometimes there is little (or no) overlapping vocabulary between the texts that generate the answers and the questions. Therefore, many tasks require reasoning through evidence scattered in multiple text fragments. At the same time, before reasoning, the text that needs to be processed is usually relatively long. If it is not processed properly, a lot of valuable information may be lost, thereby affecting the inference effect. [0003] For ...

AI Technical Summary

Problems solved by technology

Therefore, many tasks require reasoning through evidence scattered in multiple text fragments. At the same time, before reasoning, the text that needs to be processed is usually relatively long. If it is not processed properly, a lot of valuable information may be lost, thereby affecting the inference effect.

[0003] For a question answering system input query question, the machine may easily find the answer in a text

But in many cases, the answer may be hidden in a pile of text. At this time, even humans need to go through rigorous reasoning and analysis to get the correct answer.

Images