Disentangling Interpretable Generative Parameters of Random and Real-World Graphs

Abstract

While a wide range of interpretable generative procedures for graphs exist, matching observed graph topologies with such procedures and choices for its parameters remains an open problem. Devising generative models that closely reproduce real-world graphs requires domain knowledge and time-consuming simulation. While existing deep learning approaches rely on less manual modelling, they offer little interpretability. This work approaches graph generation (decoding) as the inverse of graph compression (encoding). We show that in a disentanglement-focused deep autoencoding framework, specifically Beta-Variational Autoencoders (Beta-VAE), choices of generative procedures and their parameters arise naturally in the latent space. Our model is capable of learning disentangled, interpretable latent variables that represent the generative parameters of procedurally generated random graphs and real-world graphs. The degree of disentanglement is quantitatively measured using the Mutual Information Gap (MIG). When training our Beta-VAE model on ER random graphs, its latent variables have a near one-to-one mapping to the ER random graph parameters n and p. We deploy the model to analyse the correlation between graph topology and node attributes measuring their mutual dependence without handpicking topological properties.

33rd Conference on Neural Information Processing Systems (NeurIPS 2019), Workshop on Graph Representation Learning

References in corpus (10)

• Emergence of scaling in random networks
• Geometric deep learning: going beyond Euclidean data
• Evolution of networks
• The Dynamics of Viral Marketing
• Kronecker Graphs: An Approach to Modeling Networks
• Learning Deep Generative Models of Graphs
• Constrained Graph Variational Autoencoders for Molecule Design
• Accurately modeling the Internet topology
• Towards Interpretable Sparse Graph Representation Learning with Laplacian Pooling
• Independent Subspace Analysis for Unsupervised Learning of Disentangled Representations

Cited by in corpus (2)

• GraphTune: A Learning-based Graph Generative Model with Tunable Structural Features
• Deep Generative Modeling in Network Science with Applications to Public Policy Research