CUQ-GNN: Committee-based Graph Uncertainty Quantification using Posterior Networks
This work addresses uncertainty estimation in graph-based node classification, which is incremental as it builds on and improves prior methods like GPN.
The authors tackled the problem of predictive uncertainty quantification on graph data by showing that existing Graph Posterior Network (GPN) axioms are not always satisfied in practice, and they proposed CUQ-GNN, which combines Graph Neural Networks with Normalizing Flows to adapt flexibly to domain-specific demands, demonstrating effectiveness on common benchmarks.
In this work, we study the influence of domain-specific characteristics when defining a meaningful notion of predictive uncertainty on graph data. Previously, the so-called Graph Posterior Network (GPN) model has been proposed to quantify uncertainty in node classification tasks. Given a graph, it uses Normalizing Flows (NFs) to estimate class densities for each node independently and converts those densities into Dirichlet pseudo-counts, which are then dispersed through the graph using the personalized Page-Rank algorithm. The architecture of GPNs is motivated by a set of three axioms on the properties of its uncertainty estimates. We show that those axioms are not always satisfied in practice and therefore propose the family of Committe-based Uncertainty Quantification Graph Neural Networks (CUQ-GNNs), which combine standard Graph Neural Networks with the NF-based uncertainty estimation of Posterior Networks (PostNets). This approach adapts more flexibly to domain-specific demands on the properties of uncertainty estimates. We compare CUQ-GNN against GPN and other uncertainty quantification approaches on common node classification benchmarks and show that it is effective at producing useful uncertainty estimates.