Graph Normalizing Flows
This work addresses the challenge of scalable and efficient graph neural networks for both prediction and generation tasks, representing an incremental improvement with specific architectural advantages.
The authors tackled the problem of graph prediction and generation by introducing graph normalizing flows, a reversible graph neural network model that performs similarly to message passing neural networks on supervised tasks but with significantly reduced memory footprint, enabling scaling to larger graphs, and achieves competitive results with state-of-the-art auto-regressive models in unsupervised graph generation while being better suited to parallel architectures.
We introduce graph normalizing flows: a new, reversible graph neural network model for prediction and generation. On supervised tasks, graph normalizing flows perform similarly to message passing neural networks, but at a significantly reduced memory footprint, allowing them to scale to larger graphs. In the unsupervised case, we combine graph normalizing flows with a novel graph auto-encoder to create a generative model of graph structures. Our model is permutation-invariant, generating entire graphs with a single feed-forward pass, and achieves competitive results with the state-of-the art auto-regressive models, while being better suited to parallel computing architectures.