Current graph neural networks (GNNs) that tackle node classification on graphs tend to only focus on nodewise scores and are solely evaluated by nodewise metrics. This limits uncertainty estimation on graphs since nodewise marginals do not fully characterize the joint distribution given the graph structure. In this work, we propose novel edgewise metrics, namely the edgewise expected calibration error (ECE) and the agree/disagree ECEs, which provide criteria for uncertainty estimation on graphs beyond the nodewise setting. Our experiments demonstrate that the proposed edgewise metrics can complement the nodewise results and yield additional insights. Moreover, we show that GNN models which consider the structured prediction problem on graphs tend to have better uncertainty estimations, which illustrates the benefit of going beyond the nodewise setting.
@article{arxiv.2210.15575,
title = {A Graph Is More Than Its Nodes: Towards Structured Uncertainty-Aware Learning on Graphs},
author = {Hans Hao-Hsun Hsu and Yuesong Shen and Daniel Cremers},
journal= {arXiv preprint arXiv:2210.15575},
year = {2022}
}
Comments
Presented at NeurIPS 2022 New Frontiers in Graph Learning Workshop (NeurIPS GLFrontiers 2022)