Document worth reading: “Graph Representation Learning: A Survey”

Research on graph illustration learning has acquired quite a few consideration currently since many info in real-world functions can be found in sort of graphs. High-dimensional graph info are generally in irregular sort, which makes them harder to analysis than image/video/audio info outlined on widespread lattices. Various graph embedding strategies have been developed to remodel the raw graph info proper right into a low-dimensional vector illustration whereas preserving the intrinsic graph properties. In this evaluation, we first make clear the graph embedding exercise and its challenges. Next, we evaluation quite a lot of graph embedding strategies with insights. Then, we take into account quite a lot of state-of-the-art methods in the direction of small and large datasets and study their effectivity. Finally, potential functions and future directions are supplied. Graph Representation Learning: A Survey