Adversarial Attack is a type of adversarial method that “fools” the discriminator by adding small perturbation to the original data. For example, Nettack modifies graph structures and nodes attributes, and uses a loss function that aims to find the best legitimate change to cause a node to be misclassified.

University of Notre Dame

Adversarial Methods are a promising field and have drawn lots of attention. And there are mainly two ways of applying adversarial methods to graphs:

1) Adversarial Training

2) Adversarial Attacks


Graph Adversarial Methods

Zhang, Z., Cui, P., & Zhu, W. (2020). Deep learning on graphs: A survey. IEEE Transactions on Knowledge and Data Engineering.

Deep Learning on Graphs: A Survey

In adversarial training, there is one generator and one discriminator. The generator aims to generate fake data and “fool” the discriminator into thinking it’s real data, and the discriminator aims to tell fake data from real data. So both model benefit from the improvement of each other.

Learn Before

Related