Answer selection method and device based on web text, server and storage medium

Abstract

The embodiment of the invention relates to the field of artificial intelligence, and discloses an answer selection method and device based on a web text, a server and a storage medium. The method comprises the following steps: selecting M candidate paragraphs related to a question from a pre-stored document; respectively splicing and encoding the M first features to obtain M vectors correspondingto the M first features, each first feature comprising each candidate paragraph and question; interacting the M vectors through a self-attention mechanism to obtain M updated vectors; the self-attention mechanism pays attention to the internal correlation between the data, the M vectors are interacted by using the self-attention mechanism so as to update the M vectors, and the mutual influence between the vectors corresponding to the features and the vectors corresponding to the features is considered, so that the updated vectors are more accurate, and N golden paragraphs can be selected fromthe M candidate paragraphs corresponding to the M updated vectors more accurately; and according to the N golden paragraphs, the position information of the answer is acquired more accurately.

Description

Technical field[0001]The embodiments of the present invention relate to the field of artificial intelligence, and in particular to an answer selection method, device, server and storage medium based on network text.Background technique[0002]Machine Reading Comprehension (MRC) is a technology that uses algorithms to make a computer understand the semantics of a text and answer related questions. Traditional machine reading comprehension is just to find the answer from a single document for a given question, while multi-hop machine reading comprehension is to reason about the content in multiple candidate documents step by step and select questions from multiple documents. Relevant paragraphs to find the answer to the question. At present, multi-hop machine reading comprehension is mainly accomplished through the following methods: selecting candidate paragraphs from the document, selecting golden paragraphs related to the question based on the pre-training model, and then selecting t...

AI Technical Summary

Problems solved by technology

However, in the process of selecting the golden paragraphs related to the question based on the pre-trained model, only the golden paragraphs are selected by the similarity score between the question and the candidate paragraphs, and then the answer to the question is selected from the golden paragraphs, which does not take into account The interaction between paragraphs ignores the process of multi-hop reasoning, which makes the accuracy of the selected golden paragraphs low, which in turn makes the accuracy of the selected answers low

Images