current position:Home>Heterogeneous network entry

Heterogeneous network entry

2022-11-24 23:07:38Wait for the scene, it is better to look for the scene

A Survey of Heterogeneous Information Network Analysis

Heterogeneous Graph Attention Network

Is heterogeneous network popular?
In a network, there are no different types of nodes, that's for sure.
Therefore, heterogeneous networks are more suitable for representing complex situations.

In natural language processing, is graph network used a lot?

Is it only used when there is a knowledge graph?
But in reading SCI journal papers, there are more.
But in the top conference papers, I haven't seen much..

How to code the picture?

A graph consists of nodes and edges.
When coding a node, the information of the neighbor nodes is combined as the representation of the current node.
Compared with the most basic network structure of the neural network, the fully connected layer (MLP), the feature matrix is ​​multiplied by the weight matrix, and the graph neural network has an additional adjacency matrix.The calculation form is very simple, three matrices are multiplied and a nonlinear transformation is added (below).
Insert image description here

GCN can be described as the "pioneering work" of the graph neural network. For the first time, the convolution operation in image processing is simply used in graph structure data processing, and a specific derivation is given, which involves complexThe spectrogram theory, for details, refer to [6][7].The derivation process is still relatively complicated, but the final result is very simple (Figure 5).

Insert image description here

Refer to the original link:

copyright notice
author[Wait for the scene, it is better to look for the scene],Please bring the original link to reprint, thank you.

Random recommended