Revisiting Robustness in Graph Machine Learning
This addresses robustness evaluation issues in graph machine learning for researchers and practitioners, offering a more principled approach that is incremental but impactful.
The paper tackles the problem of evaluating robustness in Graph Neural Networks (GNNs) by introducing a semantics-aware notion of adversarial graphs, revealing that many existing perturbations violate semantic assumptions and that GNNs exhibit over-robustness. It shows that incorporating label-structure into inference reduces over-robustness while improving test accuracy and adversarial robustness, with a theoretical proof of no robustness-accuracy tradeoff for inductive node classification.
Many works show that node-level predictions of Graph Neural Networks (GNNs) are unrobust to small, often termed adversarial, changes to the graph structure. However, because manual inspection of a graph is difficult, it is unclear if the studied perturbations always preserve a core assumption of adversarial examples: that of unchanged semantic content. To address this problem, we introduce a more principled notion of an adversarial graph, which is aware of semantic content change. Using Contextual Stochastic Block Models (CSBMs) and real-world graphs, our results uncover: $i)$ for a majority of nodes the prevalent perturbation models include a large fraction of perturbed graphs violating the unchanged semantics assumption; $ii)$ surprisingly, all assessed GNNs show over-robustness - that is robustness beyond the point of semantic change. We find this to be a complementary phenomenon to adversarial examples and show that including the label-structure of the training graph into the inference process of GNNs significantly reduces over-robustness, while having a positive effect on test accuracy and adversarial robustness. Theoretically, leveraging our new semantics-aware notion of robustness, we prove that there is no robustness-accuracy tradeoff for inductively classifying a newly added node.