Adversarial Attacks on Neural Networks for Graph Data

This page is about our paper "Adversarial Attacks on Neural Networks for Graph Data" by Daniel Zügner, Amir Akbarnejad, Stephan Günnemann, published at the ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2018.

[Paper | Poster | Talk (KDD2018) | Github | BibTeX]

Please cite our paper if you use our code, results, model, slides or poster.

An extended version of this work was also published in the ACM TKDD journal. The journal version is available here.


  • In Eq. (7) in the KDD 2018 version the plus sign before (αx - 1) should be a minus sign. This is fixed in the latest version on Arxiv as well as in the code (Dec 2021).