analiza aiqbench
This research introduces Bayesian Membership Privacy (BMP), a framework that corrects existing Graph Neural Network privacy analyses by accounting for structural correlations and stochastic training sampling. By casting membership inference as a Bayesian hypothesis test, the authors provide a sampling-aware auditing mechanism that quantifies node-level privacy leakage more accurately than global attack metrics. This approach offers a necessary refinement for practitioners aiming to assess fine-grained privacy risks in graph-based machine learning systems.
Bayesian Membership Privacy for Graph Neural Networks
arXiv:2606.04069v1 Announce Type: cross Abstract: Existing privacy analyses for Graph Neural Networks (GNNs) largely inherit assumptions from non-graph settings, overlooking structural correlations and stochastic training-graph sampling. In particular, node-dependent priors make type-I and type-II errors alone insufficient to characterize the best membership inference test. To address this, we introduce Bayesian Membership Privacy (BMP), a sampling-aware formulation of node-level membership privacy that incorporates node-dependent priors and treats graph sampling probabilities as part of the adversary's knowledge. BMP casts membership inference as a Bayesian hypothesis test and accordingly quantifies membership privacy in terms of posterior membership probability. We explore theoretical properties of BMP in relation to the existing definitions in the literature. We further propose a practical, sampling-aware auditing mechanism to estimate the parameters of BMP as a measure of node-level privacy leakage in GNNs. We conduct experiments on benchmark graph datasets and show that BMP yields fine-grained privacy insights that are not visible through global attack accuracy alone.
Abstract:Existing privacy analyses for Graph Neural Networks (GNNs) largely inherit assumptions from non-graph settings, overlooking structural correlations and stochastic training-graph sampling. In particular, node-dependent priors make type-I and type-II errors alone insufficient to characterize the best membership inference test. To address this, we introduce Bayesian Membership Privacy (BMP), a sampling-aware formulation of node-level membership privacy that incorporates node-dependent priors and treats graph sampling probabilities as part of the adversary's knowledge. BMP casts membership inference as a Bayesian hypothesis test and accordingly quantifies membership privacy in terms of posterior membership probability. We explore theoretical properties of BMP in relation to the existing definitions in the literature. We further propose a practical, sampling-aware auditing mechanism to estimate the parameters of BMP as a measure of node-level privacy leakage in GNNs. We conduct experiments on benchmark graph datasets and show that BMP yields fine-grained privacy insights that are not visible through global attack accuracy alone.
Cryptography and Security (cs.CR); Machine Learning (cs.LG)
arXiv:2606.04069 [cs.CR]
(or arXiv:2606.04069v1 [cs.CR] for this version)
https://doi.org/10.48550/arXiv.2606.04069 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 248 cuvinte
Subiecte conexe
Impact of Graph Structure on Membership-Inference Risk for Graph Neural Networks
arXiv · 3 iun. 2026
Are LLM-Enhanced GNNs Privacy-Safe?
arXiv · 27 aug. 2026
Hidden Axis of Uncertainty: Latent-Posterior Alignment in Graph Neural Networks with Bayesian Output Layers
arXiv · 24 aug. 2026
Graph-Based Inference and Topology-Aware Multi-Agent Reinforcement Learning for Large-Scale Railway Network Management
arXiv · 26 sept. 2026