Graph classification method based on U-shaped nested network

Abstract

The invention discloses a graph classification method based on a U-shaped nested network. According to the method, starting from extraction of more complex multi-scale targets, Graph U-Nets are nested, so that the structural information of overall data can be captured, and local information and global information can be considered at the same time. The invention provides an improved graph neural network for a graph classification task, and the improved graph neural network can extract complex structure features in data, thereby improving the prediction accuracy.

Description

technical field [0001] The invention belongs to the field of graph neural network in deep learning, and is mainly used for tasks related to graph classification. Background technique [0002] With the rapid development of deep learning in recent years, deep learning has achieved great success in the fields of image, video, natural language processing and speech. However, the data processed by these tasks is usually European-style data. In real life, there are many complex and irregular data that naturally exist in the form of graphs (such as social networks, molecular structures, etc.), and the classic CNN and other algorithms on images cannot be simple. Migrations are applied to graph data. Faced with these challenges, many researchers began to seek new deep learning methods to solve graph-related problems. [0003] In recent years, graph neural networks have achieved great success in graph representation learning. The essence of current GNNs is message passing, however,...

AI Technical Summary

Problems solved by technology

However, the depth of Graph U-Nets is only simple 2 to 3 layers, and only one convolution operation is performed per layer, which limits its ability to extract complex features

Images