Formal Verification of Graph Convolutional Networks with Uncertain Node Features and Uncertain Graph Structure

TL;DR

This paper introduces a formal verification method for graph convolutional networks that accounts for uncertainties in node features and graph structure, ensuring robustness in safety-critical applications.

Contribution

It presents a novel reachability analysis approach using polynomial zonotopes to verify GCN robustness under uncertain conditions, filling a key research gap.

Findings

Successfully verified GCN robustness on benchmark datasets.

Explicitly models uncertainties in node features and graph structure.

Demonstrates effectiveness of reachability analysis with polynomial zonotopes.

Abstract

Graph neural networks are becoming increasingly popular in the field of machine learning due to their unique ability to process data structured in graphs. They have also been applied in safety-critical environments where perturbations inherently occur. However, these perturbations require us to formally verify neural networks before their deployment in safety-critical environments as neural networks are prone to adversarial attacks. While there exists research on the formal verification of neural networks, there is no work verifying the robustness of generic graph convolutional network architectures with uncertainty in the node features and in the graph structure over multiple message-passing steps. This work addresses this research gap by explicitly preserving the non-convex dependencies of all elements in the underlying computations through reachability analysis with (matrix) polynomial zonotopes. We demonstrate our approach on three popular benchmark datasets.