On provable privacy vulnerabilities of graph representations

TL;DR

This paper analyzes privacy vulnerabilities in graph neural models, showing how edge reconstruction attacks can compromise sensitive information, and evaluates methods like noisy aggregation to mitigate these risks.

Contribution

It provides a theoretical analysis of similarity-based edge reconstruction attacks and assesses their effectiveness and mitigation strategies in graph representations.

Findings

SERA can perfectly reconstruct sparse graphs as size increases

Sparsity is critical for SERA's effectiveness

Noisy aggregation can mitigate SERA attacks

Abstract

Graph representation learning (GRL) is critical for extracting insights from complex network structures, but it also raises security concerns due to potential privacy vulnerabilities in these representations. This paper investigates the structural vulnerabilities in graph neural models where sensitive topological information can be inferred through edge reconstruction attacks. Our research primarily addresses the theoretical underpinnings of similarity-based edge reconstruction attacks (SERA), furnishing a non-asymptotic analysis of their reconstruction capacities. Moreover, we present empirical corroboration indicating that such attacks can perfectly reconstruct sparse graphs as graph size increases. Conversely, we establish that sparsity is a critical factor for SERA's effectiveness, as demonstrated through analysis and experiments on (dense) stochastic block models. Finally, we explore the resilience of private graph representations produced via noisy aggregation (NAG) mechanism against SERA. Through theoretical analysis and empirical assessments, we affirm the mitigation of SERA using NAG . In parallel, we also empirically delineate instances wherein SERA demonstrates both efficacy and deficiency in its capacity to function as an instrument for elucidating the trade-off between privacy and utility.