Cross-Context Backdoor Attacks against Graph Prompt Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lyu, Xiaoting, Han, Yufei, Wang, Wei, Qian, Hangwei, Tsang, Ivor, Zhang, Xiangliang
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914814722310144
author Lyu, Xiaoting
Han, Yufei
Wang, Wei
Qian, Hangwei
Tsang, Ivor
Zhang, Xiangliang
author_facet Lyu, Xiaoting
Han, Yufei
Wang, Wei
Qian, Hangwei
Tsang, Ivor
Zhang, Xiangliang
contents Graph Prompt Learning (GPL) bridges significant disparities between pretraining and downstream applications to alleviate the knowledge transfer bottleneck in real-world graph learning. While GPL offers superior effectiveness in graph knowledge transfer and computational efficiency, the security risks posed by backdoor poisoning effects embedded in pretrained models remain largely unexplored. Our study provides a comprehensive analysis of GPL's vulnerability to backdoor attacks. We introduce \textit{CrossBA}, the first cross-context backdoor attack against GPL, which manipulates only the pretraining phase without requiring knowledge of downstream applications. Our investigation reveals both theoretically and empirically that tuning trigger graphs, combined with prompt transformations, can seamlessly transfer the backdoor threat from pretrained encoders to downstream applications. Through extensive experiments involving 3 representative GPL methods across 5 distinct cross-context scenarios and 5 benchmark datasets of node and graph classification tasks, we demonstrate that \textit{CrossBA} consistently achieves high attack success rates while preserving the functionality of downstream applications over clean input. We also explore potential countermeasures against \textit{CrossBA} and conclude that current defenses are insufficient to mitigate \textit{CrossBA}. Our study highlights the persistent backdoor threats to GPL systems, raising trustworthiness concerns in the practices of GPL techniques.
format Preprint
id arxiv_https___arxiv_org_abs_2405_17984
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Cross-Context Backdoor Attacks against Graph Prompt Learning
Lyu, Xiaoting
Han, Yufei
Wang, Wei
Qian, Hangwei
Tsang, Ivor
Zhang, Xiangliang
Machine Learning
Graph Prompt Learning (GPL) bridges significant disparities between pretraining and downstream applications to alleviate the knowledge transfer bottleneck in real-world graph learning. While GPL offers superior effectiveness in graph knowledge transfer and computational efficiency, the security risks posed by backdoor poisoning effects embedded in pretrained models remain largely unexplored. Our study provides a comprehensive analysis of GPL's vulnerability to backdoor attacks. We introduce \textit{CrossBA}, the first cross-context backdoor attack against GPL, which manipulates only the pretraining phase without requiring knowledge of downstream applications. Our investigation reveals both theoretically and empirically that tuning trigger graphs, combined with prompt transformations, can seamlessly transfer the backdoor threat from pretrained encoders to downstream applications. Through extensive experiments involving 3 representative GPL methods across 5 distinct cross-context scenarios and 5 benchmark datasets of node and graph classification tasks, we demonstrate that \textit{CrossBA} consistently achieves high attack success rates while preserving the functionality of downstream applications over clean input. We also explore potential countermeasures against \textit{CrossBA} and conclude that current defenses are insufficient to mitigate \textit{CrossBA}. Our study highlights the persistent backdoor threats to GPL systems, raising trustworthiness concerns in the practices of GPL techniques.
title Cross-Context Backdoor Attacks against Graph Prompt Learning
topic Machine Learning
url https://arxiv.org/abs/2405.17984