通过采样稳定性分析图卷积网络的隐私-效用权衡，给出误分类率上界。
Misclassification Rate and Privacy-Utility Trade-offs in Graph Convolutional Networks via Subsampling Stability
- 基于采样稳定性框架，推导误分类率与采样概率的关系。
- 发现采样概率过大会导致隐私保证失效，过小则影响模型准确率。
- 首次为GCN在差分隐私下的采样稳定性提供理论保障，适合隐私计算研究者。
本文通过采样稳定性框架研究图卷积网络（GCNs）中的差分隐私（DP）。推导出误分类率的上界，该上界显式依赖于采样概率 $p_s$。进一步刻画了隐私-效用权衡：若 $p_s$ 过大，基于稳定性的隐私条件难以满足，导致隐私保证无效；若 $p_s$ 过小，模型准确率下降。研究成果首次为GCNs在差分隐私下提供严谨的理论分析框架。
原文摘要 · Abstract (English)
We study differential privacy (DP) in Graph Convolutional Networks (GCNs) through the framework of \textit{subsampling stability}. We derive upper bounds on the misclassification rate that depend explicitly on the subsampling probability $p_s$. Furthermore, we characterize the \textit{privacy--utility trade-off} by identifying feasible ranges of $p_s$; if $p_s$ is too large, the stability-based privacy condition becomes difficult to satisfy, yielding vacuous guarantees, whereas if it is too small, accuracy deteriorates. Our results provide the first rigorous theoretical framework for understanding subsampling stability in GCNs under DP.
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。