Modeling Edge Features with Deep Bayesian Graph Networks

TL;DR

This paper extends a deep probabilistic graph model to incorporate edge features using a Bayesian network, leading to richer graph representations and improved performance in classification, regression, and link prediction tasks.

Contribution

It introduces an architectural extension with a Bayesian network for modeling edge features, enhancing the original model's capabilities and performance.

Findings

Improved graph classification accuracy on benchmarks.

Enhanced link prediction performance in multiple tasks.

Maintains linear computational complexity for large graphs.

Abstract

We propose an extension of the Contextual Graph Markov Model, a deep and probabilistic machine learning model for graphs, to model the distribution of edge features. Our approach is architectural, as we introduce an additional Bayesian network mapping edge features into discrete states to be used by the original model. In doing so, we are also able to build richer graph representations even in the absence of edge features, which is confirmed by the performance improvements on standard graph classification benchmarks. Moreover, we successfully test our proposal in a graph regression scenario where edge features are of fundamental importance, and we show that the learned edge representation provides substantial performance improvements against the original model on three link prediction tasks. By keeping the computational complexity linear in the number of edges, the proposed model is amenable to large-scale graph processing.