Learning with Graph Neural Networks

In this paper I will discuss machine learning with graphs and its different use cases. Machine-learning prediction challenges at the node level, edge level, and graph level were discussed. The selection of a graph representation was then discussed in terms of directed and undirected graphs, bipartite graphs, weighted and unweighted graphs, adjacency matrices, and some definitions from graph theory, such as the connectivity of graphs, weak connectivity, and strong connectivity, as well as the concept of node degree