.png){kind=icon}
返回
Backdoor attacks on unsupervised graph representation learning
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neunet.2024.106668.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Unsupervised graph learning techniques have garnered increasing interest among researchers. These methods employ the technique of maximizing mutual information to generate representations of nodes and graphs. We show that these methods are susceptible to backdoor attacks, wherein the adversary can poison a small portion of unlabeled graph data (e.g., node features and graph structure) by introducing triggers into the graph. This tampering disrupts the representations and increases the risk to various downstream applications. Previous backdoor attacks in supervised learning primarily operate directly on the label space and may not be suitable for unlabeled graph data. To tackle this challenge, we introduce GRBA,(1) a gradient-based first-order backdoor attack method. To the best of our knowledge, this constitutes a pioneering endeavor in investigating backdoor attacks within the domain of unsupervised graph learning. The initiation of this method does not necessitate prior knowledge of downstream tasks, as it directly focuses on representations. Furthermore, it is versatile and can be applied to various downstream tasks, including node classification, node clustering and graph classification. We evaluate GRBA on state-of-the-art unsupervised learning models, and the experimental results substantiate the effectiveness and evasiveness of GRBA in both node-level and graph-level tasks.
Keyword:
Backdoor attack
Unsupervised graph learning
Trigger
期刊
IF:
6.3
论文数:
8.3K
被引数:
3.0W
机构
查看更多机构
引用论文
ModuleDiscoverer: Identification of regulatory modules in protein-protein interaction networks.png)
SCIENTIFIC REPORTS
OAAI
SCIENTIFIC REPORTSIF3.9
OAAI分享收藏Unsupervised graph-level representation learning with hierarchical contrasts具有层次对比的无监督图级表示学习NEURAL NETWORKS
NEURAL NETWORKSIF6.3
PREAI
分享收藏Mechanical characterization of neural electrodes based on PDMS-parylene C-PDMS sandwiched system
PREAI
分享收藏Location-aware convolutional neural networks for graph classification用于图分类的位置感知卷积神经网络NEURAL NETWORKS
NEURAL NETWORKSIF6.3
PREAI
分享收藏MGLNN: Semi-supervised learning via Multiple Graph Cooperative Learning Neural NetworksMGLNN: 基于多图合作学习神经网络的半监督学习NEURAL NETWORKS
NEURAL NETWORKSIF6.3
PREAI
分享收藏A Scalable Permutation Approach Reveals Replication and Preservation Patterns of Network Modules in Large DatasetsCELL SYSTEMSOAAI
CELL SYSTEMSIF7.7
OAAI分享收藏Multi-graph Fusion Graph Convolutional Networks with pseudo-label supervision具有伪标签监督的多图融合图卷积网络NEURAL NETWORKS
NEURAL NETWORKSIF6.3
PREAI
分享收藏 没有更多内容