GRID: Protecting Training Graph from Link Stealing Attacks on GNN Models
Authors:
- 12 May 2025
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2095-2113
Abstract
Graph neural networks (GNNs) have exhibited superior performance in various classification tasks on graph-structured data. However, they encounter the potential vulnerability from the link stealing attacks, which can infer the presence of a link between two nodes via measuring the similarity of its incident nodes' prediction vectors produced by a GNN model. Such attacks pose severe security and privacy threats to the training graph used in GNN models. In this work, we propose a novel solution, called Graph Link Disguise (GRID), to defend against link stealing attacks with the formal guarantee of GNN model utility for retaining prediction accuracy. The key idea of GRID is to add carefully crafted noises to the nodes' prediction vectors for disguising adjacent nodes as n-hop indirect neighboring nodes. We take into account the graph topology and select only a subset of nodes (called core nodes) covering all links for adding noises, which can avert the noises offset and have the further advantages of reducing both the distortion loss and the computation cost. Our crafted noises can ensure 1) the noisy prediction vectors of any two adjacent nodes have their similarity level like that of two non-adjacent nodes and 2) the model prediction is unchanged to ensure zero utility loss. Extensive experiments on five datasets are conducted to show the effectiveness of our proposed GRID solution against different representative link-stealing attacks under transductive settings and inductive settings respectively, as well as two influence-based attacks. Meanwhile, it achieves a much better privacy-utility trade-off than existing methods when extended to GNNs.Keywords
All Related Versions
Funding Information
- NSF (2019511,2348452,2315613,1937787,2325564)
This publication has 27 references indexed in Scilit:
- LINKTELLER: Recovering Private Edges from Graph Neural Networks via Influence Analysis2022
- Adapting Membership Inference Attacks to GNN for Graph Classification: Approaches and Implications2021
- TransMIA: Membership Inference Attacks Using Transfer Shadow Training2021
- Comprehensive Privacy Analysis of Deep Learning: Passive and Active White-box Inference Attacks against Centralized and Federated Learning2019
- Differentially Private Model Publishing for Deep Learning2019
- Stealing Hyperparameters in Machine Learning2018
- Membership Inference Attacks Against Machine Learning Models2017
- Distinguishing Enzyme Structures from Non-enzymes Without AlignmentsJournal of Molecular Biology, 2003
- Structural, Syntactic, and Statistical Pattern Recognition2002
- Stolen memories2002