Summary of Adversarial Robustness Of Link Sign Prediction in Signed Graphs, by Jialong Zhou et al.
Adversarial Robustness of Link Sign Prediction in Signed Graphs
by Jialong Zhou, Xing Ai, Yuni Lai, Tomasz Michalak, Gaolei Li, Jianhua Li, Kai Zhou
First submitted to arxiv on: 19 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Cryptography and Security (cs.CR)
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 The paper investigates signed graphs, which represent positive and negative relationships in social networks. Signed graph neural networks (SGNNs) are a primary tool for analyzing these graphs, but balance theory, used to model signed relationships, introduces vulnerabilities to black-box attacks. The authors propose a novel adversarial strategy called balance-attack and develop an efficient algorithm to solve the associated optimization problem. They also introduce Balance Augmented-Signed Graph Contrastive Learning (BA-SGCL), which combines contrastive learning with balance augmentation techniques to achieve robust graph representations. The paper demonstrates the effectiveness of the proposed framework on multiple SGNN architectures and real-world datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper looks at how we can represent relationships in social networks, like friendships or rivalries. They use something called signed graphs and neural networks to do this. But they found that these approaches have a problem – they’re not very good at dealing with bad information. So, they came up with some new ideas to make it better. They created a way to mess up the relationships on purpose (called balance-attack) and then developed a way to fix the problems that causes. This new approach is called BA-SGCL and it helps keep the relationships stable even when there’s bad information. |
Keywords
* Artificial intelligence * Optimization
