Training Stable Graph Neural Networks Through Constrained Learning
This work addresses stability issues in GNNs for applications involving network data, but it is incremental as it builds on existing stability properties of graph filters.
The paper tackles the problem of improving the stability of Graph Neural Networks (GNNs) to graph perturbations by proposing a constrained learning approach that imposes stability constraints, resulting in more stable representations without compromising overall accuracy.
Graph Neural Networks (GNN) rely on graph convolutions to learn features from network data. GNNs are stable to different types of perturbations of the underlying graph, a property that they inherit from graph filters. In this paper we leverage the stability property of GNNs as a typing point in order to seek for representations that are stable within a distribution. We propose a novel constrained learning approach by imposing a constraint on the stability condition of the GNN within a perturbation of choice. We showcase our framework in real world data, corroborating that we are able to obtain more stable representations while not compromising the overall accuracy of the predictor.