Graph neural networks in TensorFlow

Graph neural networks, or GNNs, are a powerful technique that leverage the graph's connectivity and input features to make predictions about graphs. TF-GNN is a production-tested library for building GNNs at large scale. It supports both modeling and training in TensorFlow and provides tooling for subgraph sampling. GNNs can be trained in an unsupervised way to compute a continuous representation or embedding of the graph structure.