针对差分隐私生成的合成数据，提出更高效精准的成员推理攻击
TAMIS: Tailored Membership Inference Attacks on Synthetic Data
- 仅用合成数据恢复图模型，无需额外影子模型
- 新攻击评分有天然阈值，准确率与现有方法相当
- 适合研究数据隐私泄露风险的学者和安全评估人员
成员推理攻击（MIA）可实证评估机器学习算法的隐私性。本文提出TAMIS，一种针对基于图模型的差分隐私合成数据生成方法的新MIA。该攻击在近期最先进的MAMA-MIA基础上改进：首先，仅通过合成数据即可恢复生成图模型，无需依赖辅助影子模型，降低计算成本且性能更优；其次，提出更数学严谨的攻击得分，具备自然二分类阈值。实验表明，TAMIS在复现SNAKE挑战任务时，性能优于或等同于MAMA-MIA。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。