Graph neural networks (GNNs) which apply the deep neural networks to graph data have achieved significant performance for the task of semi-supervised node classification. However, only few work has addressed the adversarial robustness of GNNs. In this paper, we first present a novel gradient-based attack method that facilitates the difficulty of tackling discrete graph data. When comparing to current adversarial attacks on GNNs, the results show that by only perturbing a small number of edge perturbations, including addition and deletion, our optimization-based attack can lead to a noticeable decrease in classification performance. Moreover, leveraging our gradient-based attack, we propose the first optimization-based adversarial training for GNNs. Our method yields higher robustness against both different gradient based and greedy attack methods without sacrificing classification accuracy on original graph.
|Original language||English (US)|
|Title of host publication||Proceedings of the 28th International Joint Conference on Artificial Intelligence, IJCAI 2019|
|Publisher||International Joint Conferences on Artificial Intelligence|
|Number of pages||7|
|State||Published - 2019|
|Event||28th International Joint Conference on Artificial Intelligence, IJCAI 2019 - Macao, China|
Duration: Aug 10 2019 → Aug 16 2019
|Name||IJCAI International Joint Conference on Artificial Intelligence|
|Conference||28th International Joint Conference on Artificial Intelligence, IJCAI 2019|
|Period||8/10/19 → 8/16/19|
Bibliographical noteFunding Information:
This work is supported by Air Force Research Laboratory FA8750-18-2-0058 and the MIT-IBM Watson AI Lab.
© 2019 International Joint Conferences on Artificial Intelligence. All rights reserved.