PyG w/ SBERT Sentence Transformers for Node Classification in heterogeneous Graphs, coded in PyG (PyTorch geometric) on a free COLAB NB. ML on GRAPHS. Graph-structured data such as social graphs, networks in cybersecurity, or molecular representations are our real-world scenarios which generate heterogeneous Graphs, on which to apply our ML models (Node2Vec, Message Passing MP-GNN, GCN - Graph Convolutional Networks) for prediction of node classification or simply classical link prediction. Detecting fraudulent entities in the network in cybersecurity can be a node classification problem. Therefore we will focus on NODE CLASSIFICATION on heterogeneous Graphs. And code our algorithms in PyG. Most real-world datasets are stored as heterogeneous graphs, like graphs in the area of recommendation (social graphs) are indeed heterogeneous. They store information about different types of entities (nodes) and their different types of relations. This tutorial introduces how heterogeneous graphs and their Node and Edge information (node feature tensor and edge feature tensor) are mapped to PyG and how they can be transformed as input to ML on Graph Neural Network models. All rights of code w/ PyTorch geometric: https://pytorch-geometric.readthedocs.io/en/latest/notes/heterogeneous.html #pytorch #geometric #heterogeneous #graphs