Word vector representation acquisition method and device, equipment, and readable medium

Abstract

The invention relates to a word vector representation acquisition method and device, equipment, and a readable medium. The acquisition method comprises the steps of constructing a word dependence graph corresponding to a corpus; inputting the text features of the word dependence graph into a graph neural network model suitable for a directed graph for training to obtain a word vector graph structure model, wherein the graph neural network model suitable for the directed graph is obtained by constructing embedded representation of an in-out chain of a target node, feature vector representationof an out-chain neighbor node and feature vector representation of an in-chain neighbor node of the graph neural network model; and performing graph embedding processing on the word dependence graph by utilizing the word vector graph structure model, and determining word vector representation of each word in the word dependence graph. According to the method, the word vector graph structure modelcan be applied to the undirected graph to obtain the word vector representation of the undirected graph, and can also be applied to the directed graph to obtain the word vector representation of the directed graph.

Description

technical field [0001] The present application relates to the technical field of data processing, and in particular to a method, device, device, and readable medium for acquiring word vector representations. Background technique [0002] With the rapid development of the Internet era, data storage is more convenient and faster, and the graph algorithm has been further developed. The graph structure itself has relevance. Among them, the link that defines the data structure means the relationship between the linked entities. [0003] However, the scope of application of some graph algorithms in the prior art has certain limitations. For example, the graphsage model in the prior art is aimed at an isotropic network structure, which is only applicable to the embedded representation of undirected graph nodes and cannot be applied to Embedded representations of directed graph nodes; [0004] For the GCN (Graph Convolutional Network) model, when a new node enters the graph, all no...

AI Technical Summary

Problems solved by technology

[0003] However, the scope of application of some graph algorithms in the prior art has certain limitations. For example, the graphsage model in the prior art is aimed at an isotropic network structure, which is only applicable to the embedded representation of undirected graph nodes and cannot be applied to Embedded representations of directed graph nodes;

[0004] For the GCN (Graph Convolutional Network) model, when a new node enters the graph, all nodes in the whole graph need to be recalculated, which leads to the need to calculate the full quantifier and cause slow calculation speed

Images