Summary of Can Graph Neural Networks Expose Training Data Properties? An Efficient Risk Assessment Approach, by Hanyang Yuan et al.
Can Graph Neural Networks Expose Training Data Properties? An Efficient Risk Assessment Approach
by Hanyang Yuan, Jiarong Xu, Renhong Huang, Mingli Song, Chunping Wang, Yang Yang
First submitted to arxiv on: 6 Nov 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 proposed efficient graph property inference attack leverages model approximation techniques to identify sensitive property information leakage from shared GNN models. This approach reduces computational intensity by only requiring training a small set of models on graphs, generating approximated shadow models for attacks. The method introduces an edit distance-based criterion to evaluate each model’s error and a novel selection mechanism to ensure high diversity and low error. Experimental results demonstrate substantial improvement across six real-world scenarios, with average increases in attack accuracy and ROC-AUC while being faster compared to the best baseline. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper explores how sharing trained GNN models can accidentally leak sensitive information from private graphs. To address this issue, researchers propose an efficient method for identifying potential leaks using model approximation techniques. This approach reduces computational costs by only requiring a small number of model trainings and generates approximated shadow models to simulate attacks. The method also introduces new criteria to evaluate the accuracy of these simulated attacks. Experimental results show that this method improves attack accuracy and reduces computation time. |
Keywords
» Artificial intelligence » Auc » Gnn » Inference
