Finding MNEMON: Reviving Memories of Node Embeddings
Date
07 November 2022
Author
Yun Shen, NetApp, Bristol, England UK; Yufei Han, Inria, Rennes, France; Zhikun Zhang, CISPA Helmholtz Center for Information Security, Saarbrücken, Germany; Min Chen, CISPA Helmholtz Center for Information Security, Saarbrücken, Germany ; Ting Yu, Qatar Computing Research Institute, Doha, Qatar; Michael Backes, CISPA Helmholtz Center for Information Security, Saarbrücken, Germany; Yang Zhang, CISPA Helmholtz Center for Information Security, Saarbrücken, Germany; Gianluca Stringhini, Boston University, Boston, MA, USA
CCS '22: Proceedings of the 2022 ACM SIGSAC Conference on Computer and Communications Security November 2022
Previous security research efforts orbiting around graphs have been exclusively focusing on either (de-)anonymizing the graphs or understanding the security and privacy issues of graph neural networks. Little attention has been paid to understand the privacy risks of integrating the output from graph embedding models (e.g., node embeddings) with complex downstream machine learning pipelines. In this paper, we fill this gap and propose a novel model-agnostic graph recovery attack that exploits the implicit graph structural information preserved in the embeddings of graph nodes. We show that an adversary can recover edges with decent accuracy by only gaining access to the node embedding matrix of the original graph without interactions with the node embedding models. We demonstrate the effectiveness and applicability of our graph recovery attack through extensive experiments.
Resources
The paper can be found at: https://dl.acm.org/doi/10.1145/3548606.3559358
