Saved in:
Bibliographic Details
Main Authors: Zhang, Yexin, Ma, Zhongtian, Zhang, Qiaosheng, Wang, Zhen
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.01987
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914527421923328
author Zhang, Yexin
Ma, Zhongtian
Zhang, Qiaosheng
Wang, Zhen
author_facet Zhang, Yexin
Ma, Zhongtian
Zhang, Qiaosheng
Wang, Zhen
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2605_01987
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Misclassification Rate and Privacy-Utility Trade-offs in Graph Convolutional Networks via Subsampling Stability
Zhang, Yexin
Ma, Zhongtian
Zhang, Qiaosheng
Wang, Zhen
Machine Learning
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.
title Misclassification Rate and Privacy-Utility Trade-offs in Graph Convolutional Networks via Subsampling Stability
topic Machine Learning
url https://arxiv.org/abs/2605.01987