Graph neural network: Current state of Art, challenges and applications

Several areas in science and engineering have the relationships between their underlying data which can be represented as graphs, for example, molecular chemistry, node prediction, link prediction, computer vision, pattern recognition, social networking and more. In this article, an approach to a model which can handle such type of data is elaborated, which is Graph Neural Networks (GNN). GNN encompasses the neural network technique to process the data which is represented as graphs. Due to its massive success, GNN has made its way into many applications and is a popular architecture to work upon. This paper explains the graph neural networks, its area of applications and its day-to-day use in our daily lives. Some of the very common application is a social networking site which is on our hands regularly, and another could be the recommendation system which recommends us friends, or the products of our interest based on our pat choices and preferences. This paper also demonstrates the basic challenges encountered while implementing GNN. This paper will be a great help to those researchers who are keen to work in the domain of GNN.