Uncertainty Aware Semi-Supervised Learning on Graph Data

TL;DR

This paper introduces a novel uncertainty-aware semi-supervised learning framework for graph neural networks that improves node classification by effectively detecting out-of-distribution nodes and misclassifications using evidence theory.

Contribution

It proposes a multi-source uncertainty framework and the GKDE method for accurate Dirichlet distribution prediction and uncertainty estimation in graph data.

Findings

Discovered dissonance-based detection excels in misclassification detection.

Vacuity-based detection is most effective for out-of-distribution detection.

The proposed model outperforms state-of-the-art methods on six real network datasets.

Abstract

Thanks to graph neural networks (GNNs), semi-supervised node classification has shown the state-of-the-art performance in graph data. However, GNNs have not considered different types of uncertainties associated with class probabilities to minimize risk of increasing misclassification under uncertainty in real life. In this work, we propose a multi-source uncertainty framework using a GNN that reflects various types of predictive uncertainties in both deep learning and belief/evidence theory domains for node classification predictions. By collecting evidence from the given labels of training nodes, the Graph-based Kernel Dirichlet distribution Estimation (GKDE) method is designed for accurately predicting node-level Dirichlet distributions and detecting out-of-distribution (OOD) nodes. We validated the outperformance of our proposed model compared to the state-of-the-art counterparts in terms of misclassification detection and OOD detection based on six real network datasets. We found that dissonance-based detection yielded the best results on misclassification detection while vacuity-based detection was the best for OOD detection. To clarify the reasons behind the results, we provided the theoretical proof that explains the relationships between different types of uncertainties considered in this work.