Summary of Rida: a Robust Attack Framework on Incomplete Graphs, by Jianke Yu et al.
RIDA: A Robust Attack Framework on Incomplete Graphs
by Jianke Yu, Hanchen Wang, Chen Chen, Xiaoyang Wang, Lu Qin, Wenjie Zhang, Ying Zhang, Xijuan Liu
First submitted to arxiv on: 25 Jul 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary In this paper, researchers aim to improve the robustness of Graph Neural Networks (GNNs) by developing strong adversarial attacks as benchmarks. They focus on gray-box poisoning attacks, which target GNNs’ reliance on retraining with updated data. To address real-world scenarios where graphs are incomplete, the authors introduce the Robust Incomplete Deep Attack Framework (RIDA), an algorithm for robust gray-box attacks on incomplete graphs. RIDA aggregates distant vertex information and ensures effective data utilization. The authors test RIDA against 9 state-of-the-art baselines on 3 real-world datasets, demonstrating its superiority in handling incompleteness and high attack performance. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary GNNs are important tools for data science, but they can be easily attacked by hackers. To help researchers create stronger GNN models, scientists need to develop better “attack” models that test the strength of these GNNs. One type of attack is called a gray-box poisoning attack, which is very effective and has fewer rules than other attacks. However, current research doesn’t consider what happens when graphs are not complete. To fix this gap, researchers created a new algorithm called RIDA (Robust Incomplete Deep Attack Framework). This algorithm helps create robust gray-box attacks on incomplete graphs by combining information from distant parts of the graph and using data effectively. The scientists tested RIDA against other popular algorithms on real-world datasets and found that it outperformed them in handling incomplete graphs and attacking GNNs. |
Keywords
* Artificial intelligence * Gnn
