Text similarity matching method, electronic equipment and computer readable medium

Abstract

The invention discloses a text similarity matching method, electronic equipment and a computer readable medium, and the text similarity matching method comprises the following steps: obtaining an input text which comprises at least one input vocabulary; inputting the input text into a text similarity matching model to perform matching prediction, respectively introducing general context information of the at least one input vocabulary into the input text through a graph neural network, and performing matching prediction according to the introduced input text; and outputting the label of the predicted text matched and predicted by the text similarity matching model. According to the text similarity matching method, the similarity matching task of the short text can be effectively completed,so that the matching accuracy and the matching efficiency are greatly improved.

Description

technical field [0001] The invention relates to the technical field of artificial intelligence, in particular to a short text similarity matching method based on fusion of vocabulary context features. Background technique [0002] In recent years, thanks to the rapid development of deep learning, text similarity matching tasks (such as search, retrieval question answering, etc.) have made remarkable progress. The task of text similarity matching can be defined as: given a piece of input text as Query, and then matching the most relevant documents or answers among the candidate documents or answer set candidates. [0003] At present, the text similarity matching model usually uses vocabulary and its context information in the input text to analyze the meaning expressed in the input text, and then uses some deep learning architectures to match the most relevant documents or answers in the candidate set candidates. [0004] However, when faced with a short text query consistin...

AI Technical Summary

Problems solved by technology

[0005] The technical problem to be solved by the present invention is to overcome the defect that the text similarity matching model in the prior art cannot effectively match short texts, resulting in reduced matching accuracy and matching efficiency, and to provide a text similarity matching method and electronic equipment and computer readable media, especially a short text similarity matching method based on fusion of lexical context features

Images