Verifying message-passing neural networks via topology-based bounds tightening

Christopher Hojny, Shiqiang Zhang, Juan S Campos, Ruth Misener
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:18489-18514, 2024.

Abstract

Since graph neural networks (GNNs) are often vulnerable to attack, we need to know when we can trust them. We develop a computationally effective approach towards providing robust certificates for message-passing neural networks (MPNNs) using a Rectified Linear Unit (ReLU) activation function. Because our work builds on mixed-integer optimization, it encodes a wide variety of subproblems, for example it admits (i) both adding and removing edges, (ii) both global and local budgets, and (iii) both topological perturbations and feature modifications. Our key technology, topology-based bounds tightening, uses graph structure to tighten bounds. We also experiment with aggressive bounds tightening to dynamically change the optimization constraints by tightening variable bounds. To demonstrate the effectiveness of these strategies, we implement an extension to the open-source branch-and-cut solver SCIP. We test on both node and graph classification problems and consider topological attacks that both add and remove edges.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-hojny24a, title = {Verifying message-passing neural networks via topology-based bounds tightening}, author = {Hojny, Christopher and Zhang, Shiqiang and Campos, Juan S and Misener, Ruth}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {18489--18514}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/hojny24a/hojny24a.pdf}, url = {https://proceedings.mlr.press/v235/hojny24a.html}, abstract = {Since graph neural networks (GNNs) are often vulnerable to attack, we need to know when we can trust them. We develop a computationally effective approach towards providing robust certificates for message-passing neural networks (MPNNs) using a Rectified Linear Unit (ReLU) activation function. Because our work builds on mixed-integer optimization, it encodes a wide variety of subproblems, for example it admits (i) both adding and removing edges, (ii) both global and local budgets, and (iii) both topological perturbations and feature modifications. Our key technology, topology-based bounds tightening, uses graph structure to tighten bounds. We also experiment with aggressive bounds tightening to dynamically change the optimization constraints by tightening variable bounds. To demonstrate the effectiveness of these strategies, we implement an extension to the open-source branch-and-cut solver SCIP. We test on both node and graph classification problems and consider topological attacks that both add and remove edges.} }
Endnote
%0 Conference Paper %T Verifying message-passing neural networks via topology-based bounds tightening %A Christopher Hojny %A Shiqiang Zhang %A Juan S Campos %A Ruth Misener %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-hojny24a %I PMLR %P 18489--18514 %U https://proceedings.mlr.press/v235/hojny24a.html %V 235 %X Since graph neural networks (GNNs) are often vulnerable to attack, we need to know when we can trust them. We develop a computationally effective approach towards providing robust certificates for message-passing neural networks (MPNNs) using a Rectified Linear Unit (ReLU) activation function. Because our work builds on mixed-integer optimization, it encodes a wide variety of subproblems, for example it admits (i) both adding and removing edges, (ii) both global and local budgets, and (iii) both topological perturbations and feature modifications. Our key technology, topology-based bounds tightening, uses graph structure to tighten bounds. We also experiment with aggressive bounds tightening to dynamically change the optimization constraints by tightening variable bounds. To demonstrate the effectiveness of these strategies, we implement an extension to the open-source branch-and-cut solver SCIP. We test on both node and graph classification problems and consider topological attacks that both add and remove edges.
APA
Hojny, C., Zhang, S., Campos, J.S. & Misener, R.. (2024). Verifying message-passing neural networks via topology-based bounds tightening. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:18489-18514 Available from https://proceedings.mlr.press/v235/hojny24a.html.

Related Material