Graph Normalizing Flows
arXiv:1905.13177
Abstract
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.
References in corpus (12)
- Inductive Representation Learning on Large Graphs
- Neural Message Passing for Quantum Chemistry
- NICE: Non-linear Independent Components Estimation
- Variational Inference with Normalizing Flows
- Variational Graph Auto-Encoders
- Density estimation using Real NVP
- Predicting multicellular function through multi-layer tissue networks
- GraphRNN: Generating Realistic Graphs with Deep Auto-regressive Models
- Learning Deep Generative Models of Graphs
- Gated Graph Sequence Neural Networks
- Generative Moment Matching Networks
- Adversarial Autoencoders
Cited by in corpus (22)
- Equivariant flow-based sampling for lattice gauge theory
- E(n) Equivariant Graph Neural Networks
- Equivariant Flows: Exact Likelihood Generative Learning for Symmetric Densities
- Graph Neural Ordinary Differential Equations
- CaSPR: Learning Canonical Spatiotemporal Point Cloud Representations
- Graph Residual Flow for Molecular Graph Generation
- Transfer Learning of Graph Neural Networks with Ego-graph Information Maximization
- SLAPS: Self-Supervision Improves Structure Learning for Graph Neural Networks
- AdvFlow: Inconspicuous Black-box Adversarial Attacks using Normalizing Flows
- Graph Generation with Variational Recurrent Neural Network
- Categorical Normalizing Flows via Continuous Transformations
- Text to Image Synthesis using Stacked Conditional Variational Autoencoders and Conditional Generative Adversarial Networks
- Latent Variable Modelling with Hyperbolic Normalizing Flows
- E(n) Equivariant Normalizing Flows
- UFO-BLO: Unbiased First-Order Bilevel Optimization
- The Convolution Exponential and Generalized Sylvester Flows
- GraphITE: Estimating Individual Effects of Graph-structured Treatments
- Graph-based Normalizing Flow for Human Motion Generation and Reconstruction
- Deep Generative Modeling in Network Science with Applications to Public Policy Research
- Learning Augmentation Distributions using Transformed Risk Minimization
- GFCN: A New Graph Convolutional Network Based on Parallel Flows
- TD-GEN: Graph Generation With Tree Decomposition