TAMIS: Tailored Membership Inference Attacks on Synthetic Data

2025-04-01

AI Summary

This work addresses the challenge of privacy evaluation for differentially private (DP) graph synthetic data. We propose a novel membership inference attack (MIA) that requires no auxiliary data—marking the first method to reconstruct the original graph structure model solely from synthetic data. Our approach integrates graph structure learning with statistical hypothesis testing, yielding a theoretically interpretable MIA scoring function equipped with a natural binary threshold. Evaluated on the SNAKE benchmark, our method achieves performance comparable to or better than the state-of-the-art MAMA-MIA, while significantly reducing computational overhead and reliance on attacker-side prior knowledge. Key contributions are: (1) the first auxiliary-free mechanism for reverse-engineering graph structural models from DP synthetic graphs; (2) a mathematically rigorous, directly decidable MIA scoring framework; and (3) a new empirical privacy assessment paradigm for DP graph synthesis—efficient, assumption-light, and grounded in practical attackability.

Technology Category

• Machine Learning: Privacy
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

Membership Inference Attacks (MIA) enable to empirically assess the privacy of a machine learning algorithm. In this paper, we propose TAMIS, a novel MIA against differentially-private synthetic data generation methods that rely on graphical models. This attack builds upon MAMA-MIA, a recently-published state-of-the-art method. It lowers its computational cost and requires less attacker knowledge. Our attack is the product of a two-fold improvement. First, we recover the graphical model having generated a synthetic dataset by using solely that dataset, rather than shadow-modeling over an auxiliary one. This proves less costly and more performant. Second, we introduce a more mathematically-grounded attack score, that provides a natural threshold for binary predictions. In our experiments, TAMIS achieves better or similar performance as MAMA-MIA on replicas of the SNAKE challenge.

Problem

• Assessing privacy risks in synthetic data generation
• Reducing computational cost of membership inference attacks
• Improving attack performance with mathematical scoring

Innovation

• Recovers graphical model from synthetic dataset
• Introduces mathematically-grounded attack score
• Lowers computational cost and knowledge