Document worth reading: “Graph Neural Networks for Small Graph and Giant Network Representation Learning: An Overview”

Graph neural networks denote a bunch of neural neighborhood fashions launched for the illustration finding out duties on graph information notably. Graph neural networks have been demonstrated to be environment friendly for capturing neighborhood development knowledge, and the realized representations can receive the state-of-the-art effectivity on node and graph classification duties. Besides the utterly totally different software program eventualities, the architectures of graph neural neighborhood fashions moreover depend on the studied graph kinds masses. Graph information studied in evaluation could also be usually categorized into two principal kinds, i.e., small graphs vs. massive networks, which differ from each other masses throughout the measurement, event amount and label annotation. Several varied sorts of graph neural neighborhood fashions have been launched for finding out the representations from such varied sorts of graphs already. In this paper, for these two varied sorts of graph information, we’re going to introduce the graph neural networks launched currently. To be additional specific, the graph neural networks launched on this paper embrace IsoNN, SDBN, LF&ER, GCN, GAT, DifNN, GNL, GraphSage and seGEN. Among these graph neural neighborhood fashions, IsoNN, SDBN and LF&ER are initially proposed for small graphs and the remaining ones are initially proposed for massive networks as an alternative. The readers are moreover suggested to consult with these papers for detailed knowledge when finding out this tutorial paper. Graph Neural Networks for Small Graph and Giant Network Representation Learning: An Overview