Graph neural networks has gain much popularity recently. Many domains can be readily modeled as graphs, such as social networks, molecular graph structures, and recommender system. In this article, I will briefly introduce the graph neural networks.

- what is graph network
- the general architecture of graph neural networks

Graph network is a structure that composed of **vertices** and **edges**.

Usually vertices are instances. People in social networks, products in recommender system, and institutions in organizations can all be vertices. In order to embed more information in the graph, the vertices can include many attributes as well. …

A graduate student in machine learning, data science