dK-Projection: Publishing Graph Joint Degree Distribution with Node Differential Privacy

Iftikhar, Masooma; Wang, Qing

doi:10.1007/978-3-030-75765-6_29

Cited by 5 publications

(3 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When calculating the number of triangles in a graph, for instance, maximum change of a D-bounded-degree graph is D 2 which is strictly smaller than n 2 if D < n. Settings that can benefit most from this formulations of DP are those that put an emphasis on the data within the node itself yet additionally privatise the connections between the nodes. This includes studies on social networks [8,92], the publication of higher-order network statistics [46,71,74], and recommendation systems [73].…”

Section: Edge-level Differential Privacymentioning

confidence: 99%

Differentially Private Guarantees for Analytics and Machine Learning on Graphs: A Survey of Results

Mueller,

Usynin,

Paetzold

et al. 2024

JPC

View full text Add to dashboard Cite

We study the applications of differential privacy (DP) in the context of graph-structured data and discuss the formulations of DP applicable to the publication of graphs and their associated statistics as well as machine learning on graph-based data, including graph neural networks (GNNs). Interpreting DP guarantees in the context of graph-structured data can be challenging, as individual data points are interconnected (often non-linearly or sparsely). This connectivity complicates the computation of individual privacy loss in differentially private learning. The problem is exacerbated by an absence of a single, well-established formulation of DP in graph settings. This issue extends to the domain of GNNs, rendering private machine learning on graph-structured data a challenging task. A lack of prior systematisation work motivated us to study graph-based learning from a privacy perspective. In this work, we systematise different formulations of DP on graphs, discuss challenges and promising applications, including the GNN domain. We compare and separate works into graph analytics tasks and graph learning tasks with GNNs. We conclude our work with a discussion of open questions and potential directions for further research in this area.

show abstract

Section: Edge-level Differential Privacymentioning

confidence: 99%

Differentially Private Guarantees for Analytics and Machine Learning on Graphs: A Survey of Results

Mueller,

Usynin,

Paetzold

et al. 2024

JPC

View full text Add to dashboard Cite

show abstract

“…In differential privacy methods, many methods based on differential privacy have been presented for graph data since C. Dwork came up with differential privacy, which was classified into two kinds: preserving specific sensitive statistics of graphs and generating differential private graphs. For publishing higher order network statistics, i.e., joint degree distribution, Iftikhar [29] designed a general framework for releasing dK-distributions under node differential privacy, in which sensitivity was regulated by a graph projection algorithm, which transformed graphs into bounded graphs. To accurately estimate subgraph counts, [30] proposed a novel multi-phase framework under DDP (decentralized differential privacy), which was able to control the minimum local noise scale to preserve the sub-graph counts.…”

Section: Related Workmentioning

confidence: 99%

A Node Differential Privacy-Based Method to Preserve Directed Graphs in Wireless Mobile Networks

Yan

Zhou

2023

Applied Sciences

View full text Add to dashboard Cite

With the widespread popularity of Wireless Mobile Networks (WMNs) in our daily life, the huge risk to disclose personal privacy of massive graph structure data in WMNs receives more and more attention. Particularly, as a special type of graph data in WMNs, the directed graph contains an amount of sensitive personal information. To provide secure and reliable privacy preservation for directed graphs in WMNs, we develop a node differential privacy-based method, which combines differential privacy with graph modification. In the method, the original directed graph is first divided into several sub-graphs after it is transformed into a weighted graph. Then, in each sub-graph, the node degree sequences are obtained by using an exponential mechanism and micro-aggregation is adopted to get the noised node degree sequences, which is used to generate a synthetic directed sub-graph through edge modification. Finally, all synthetic sub-graphs are merged into a synthetic directed graph that can preserve the original directed graph. The theoretical analysis proves that the proposed method satisfies differential privacy. The results of the experiments demonstrate the effectiveness of the presented method in privacy preservation and data utility.

show abstract

“…Here, the removal of a single node results in an upper-bounded change in edges which typically leads to a reduced impact on the output of the algorithm. When calculating the number of triangles in a graph, for instance, maximum change of a D-bounded-degree graph is D 2 which is strictly smaller than n 2 if D < n. Settings that can benefit most from this formulations of DP are those that put an emphasis on the data within the node itself yet additionally privatise the connections between the nodes include studies on social networks [47,33], degree histogram distribution [63,59,50], and recommendation systems [28].…”

Section: Node-level Differential Privacymentioning

confidence: 99%

SoK: Differential Privacy on Graph-Structured Data

Mueller¹,

Usynin²,

Paetzold³

et al. 2022

Preprint

View full text Add to dashboard Cite

In this work, we study the applications of differential privacy (DP) in the context of graph-structured data. We discuss the formulations of DP applicable to the publication of graphs and their associated statistics as well as machine learning on graph-based data, including graph neural networks (GNNs). The formulation of DP in the context of graph-structured data is difficult, as individual data points are interconnected (often non-linearly or sparsely). This connectivity complicates the computation of individual privacy loss in differentially private learning. The problem is exacerbated by an absence of a single, well-established formulation of DP in graph settings. This issue extends to the domain of GNNs, rendering private machine learning on graph-structured data a challenging task. A lack of prior systematisation work motivated us to study graph-based learning from a privacy perspective. In this work, we systematise different formulations of DP on graphs, discuss challenges and promising applications, including the GNN domain. We compare and separate works into graph analysis tasks and graph learning tasks with GNNs. Finally, we conclude our work with a discussion of open questions and potential directions for further research in this area.

show abstract

dK-Projection: Publishing Graph Joint Degree Distribution with Node Differential Privacy

Cited by 5 publications

References 17 publications

Differentially Private Guarantees for Analytics and Machine Learning on Graphs: A Survey of Results

Differentially Private Guarantees for Analytics and Machine Learning on Graphs: A Survey of Results

A Node Differential Privacy-Based Method to Preserve Directed Graphs in Wireless Mobile Networks

SoK: Differential Privacy on Graph-Structured Data

Contact Info

Product

Resources

About