Leakage-Aware Defense Against Training-Trace Link Inference in Vertical Federated Graph Learning
DOI:
https://doi.org/10.70088/30syrv41Keywords:
vertical federated graph learning, link inference, privacy defense, graph neural networks, leakage-aware trainingAbstract
Vertical federated graph learning (VFGL) shares intermediate training signals that can reveal graph links even when data are partitioned. We present LAD-VFGL, an empirical defense for link inference from a feature-side representation and its backward gradient. In a feature-partitioned two-branch VFGL implementation, a trace-based risk proxy combines prior protected gradients and current feature-side embeddings. An exponential moving average (EMA) of node-level risk controls Gaussian perturbation of both the forward embedding and its backward gradient; pair scores weight a decorrelation loss, while historical EMA traces preserve task utility. Across six public graph datasets and five random seeds, the gradient-channel attack area under the receiver operating characteristic curve (AUC) decreases from 0.7852 to 0.5128, while average node-classification accuracy remains 0.8758, only 0.28 percentage points below undefended training. At the evaluated operating points, LAD-VFGL attains lower attack AUC and higher average accuracy than fixed Gaussian noise. These findings are limited to the evaluated passive attack and do not constitute a formal privacy guarantee.References
T. N. Kipf and M. Welling, "Semi-supervised classification with graph convolutional networks," in Proceedings of the International Conference on Learning Representations (ICLR), 2017.
W. L. Hamilton, R. Ying, and J. Leskovec, "Inductive representation learning on large graphs," in Advances in Neural Information Processing Systems, vol. 30, pp. 1024–1034, 2017.
C. He, K. Balasubramanian, E. Ceyani, Y. Rong, P. Zhao, J. Huang, M. Annavaram, and S. Avestimehr, "FedGraphNN: A federated learning system and benchmark for graph neural networks," in Proceedings of the 1st MLSys Workshop on Graph Neural Networks and Systems (GNNSys), 2021.
C. Wu, F. Wu, Y. Cao, L. Lyu, Y. Huang, and X. Xie, "FedGNN: Federated graph neural network for privacy-preserving recommendation," presented at the International Workshop on Federated Learning for User Privacy and Data Confidentiality (FL-ICML), 2021.
H. B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. Agüera y Arcas, "Communication-efficient learning of deep networks from decentralized data," in Proceedings of the 20th International Conference on Artificial Intelligence and Statistics (AISTATS), vol. 54, pp. 1273–1282, 2017.
Q. Zhang, B. Gu, C. Deng, and H. Huang, "Secure bilevel asynchronous vertical federated learning with backward updating," in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 35, no. 12, pp. 10896–10904, 2021.
R. Liu, P. Xing, Z. Deng, A. Li, C. Guan, and H. Yu, "Federated graph neural networks: Overview, techniques, and challenges," IEEE Transactions on Neural Networks and Learning Systems, vol. 36, no. 3, pp. 4279–4295, 2025.
Z. Liu, L. Yang, Z. Fan, H. Peng, and P. S. Yu, "Federated social recommendation with graph neural network," ACM Transactions on Intelligent Systems and Technology, vol. 13, no. 4, Art. no. 55, pp. 1–24, 2022.
G. Mei, Z. Guo, S. Liu, and L. Pan, "SGNN: A graph neural network based federated learning approach by hiding structure," in 2019 IEEE International Conference on Big Data (Big Data), pp. 2560–2568, 2019.
L. Zhu, Z. Liu, and S. Han, "Deep leakage from gradients," in Advances in Neural Information Processing Systems, vol. 32, pp. 14774–14784, 2019.
J. Geiping, H. Bauermeister, H. Dröge, and M. Moeller, "Inverting gradients: How easy is it to break privacy in federated learning?" in Advances in Neural Information Processing Systems, vol. 33, pp. 16937–16947, 2020.
M. Nasr, R. Shokri, and A. Houmansadr, "Comprehensive privacy analysis of deep learning: Passive and active white-box inference attacks against centralized and federated learning," in 2019 IEEE Symposium on Security and Privacy (SP), pp. 739–753, 2019.
O. Zari, C. Xu, J. Parra-Arnau, A. Ünsal, and M. Önen, "Link inference attacks in vertical federated graph learning," in 2024 Annual Computer Security Applications Conference (ACSAC), pp. 761–777, 2024.
M. Abadi, A. Chu, I. Goodfellow, H. B. McMahan, I. Mironov, K. Talwar, and L. Zhang, "Deep learning with differential privacy," in Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security, pp. 308–318, 2016.
P. Sen, G. Namata, M. Bilgic, L. Getoor, B. Gallagher, and T. Eliassi-Rad, "Collective classification in network data," AI Magazine, vol. 29, no. 3, pp. 93–106, 2008.
O. Shchur, M. Mumme, A. Bojchevski, and S. Günnemann, "Pitfalls of graph neural network evaluation," in Relational Representation Learning Workshop, Advances in Neural Information Processing Systems (NeurIPS), 2018.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Jian Zhang, Jiaqi Liu (Author)

This work is licensed under a Creative Commons Attribution 4.0 International License.









