A Graph Is More Than Its Nodes: Towards Structured Uncertainty-Aware Learning on Graphs
This work addresses uncertainty estimation for researchers and practitioners using GNNs in graph-based tasks, though it is incremental as it builds on existing nodewise methods.
The paper tackled the limitation of graph neural networks (GNNs) in uncertainty estimation for node classification by proposing novel edgewise metrics, such as edgewise expected calibration error (ECE), which complement nodewise results and show that models considering structured prediction improve uncertainty estimations.
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.