Information processing method and device and storage medium

Abstract

The embodiment of the invention discloses an information processing method and device and a storage medium. The method comprises the steps of performing word segmentation processing on a first corpusto obtain a first word set of the first corpus; wherein the first word set comprises at least two words; determining the relevancy between any two words in the first word set; utilizing the determinedrelevancy between any two words in the first word set to construct a first word co-occurrence network; wherein the first word co-occurrence network represents an association relationship between words in the first word set; determining a first feature matrix by utilizing the data of the first word co-occurrence network and combining a graph convolutional neural network (GCN) model; wherein the first feature matrix is a feature matrix corresponding to each node in a first word co-occurrence network; and performing dimension reduction processing on the first feature matrix to obtain a word embedding initialization result. In this way, a reasonable word embedding initialization result can be provided, and therefore the effect of shortening the word embedding training period is achieved through the word embedding initialization result.

Description

technical field [0001] The present invention relates to the field of artificial intelligence, in particular to an information processing method, device and storage medium. Background technique [0002] Word embedding is a general term for language models and representation learning techniques in Natural Language Processing (NLP). Word embedding refers to embedding a high-dimensional space whose dimension is the number of all words into a continuous vector space with a much lower dimension, that is, mapping each word or phrase to a vector on a real number field. The way phrases are represented digitally is considered to be one of the most challenging problems of deep learning in NLP tasks. In related technologies, there is a problem that the word embedding initialization scheme is highly arbitrary, which makes the word embedding training period long. Contents of the invention [0003] In view of this, the embodiments of the present invention expect to provide an informati...

AI Technical Summary

Problems solved by technology

In related technologies, there is a problem that the word embedding initialization scheme is highly arbitrary, which makes the word embedding training cycle long

Images