提出一种可证明隐私的图生成方法，兼顾数据效用与安全。
Private Synthetic Graph Generation and Fused Gromov-Wasserstein Distance
- 基于随机连接模型，实现顶点级差分隐私的属性图生成。
- 使用融合格罗莫夫-沃瑟斯坦距离评估，保证合成图与真实图相似性。
- 适合需要隐私保护的社交网络、生物网络等场景研究。
网络广泛用于表示复杂数据，尤其在方法与算法开发中，对差分隐私的合成网络需求迫切。本文从复杂数据出发，联合提供网络表示与合成网络生成器。通过随机连接模型，设计了一种有效的算法，可在顶点级别实现ε-差分隐私，同时在我们提出的距离度量下保持数据效用。该距离度量为融合格罗莫夫-沃瑟斯坦距离（fused Gromov-Wasserstein distance），将经典的Wasserstein度量扩展至结构化数据。理论分析表明，所生成的私有合成图具有良好的准确性。方法灵感源自He et al. (2023) 的PSMM方法。
原文摘要 · Abstract (English)
Networks are popular for representing complex data. In particular, differentially private synthetic networks are much in demand for method and algorithm development. The network generator should be easy to implement and should come with theoretical guarantees. Here we start with complex data as input and jointly provide a network representation as well as a synthetic network generator. Using a random connection model, we devise an effective algorithmic approach for generating attributed synthetic graphs which is $ε$-differentially private at the vertex level, while preserving utility under an appropriate notion of distance which we develop. We provide theoretical guarantees for the accuracy of the private synthetic graphs using the fused Gromov-Wasserstein distance, which extends the Wasserstein metric to structured data. Our method draws inspiration from the PSMM method of \citet{he2023}.
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。