In this work, we propose the first backdoor attack to graph neural networks (GNN). Specifically, we propose a subgraph based backdoor attack to GNN for graph classification. In our backdoor attack, a ...

Backdoor attacks in neural networks represent a significant security threat in the field of machine learning. These attacks involve embedding hidden triggers into models during the training phase ...

This repository is the official implementation of Single-Node Attack For Fooling Graph Neural Networks. SINGLE NODE attack will produce a 2d matrix of SINGLE approaches such as (hops, GradChoice, ...

Graph Neural Networks (GNNs) and network embedding techniques have emerged as powerful tools for analyzing and interpreting complex data structures represented as graphs. These methods are ...

Therefore, we propose a novel approach called the preferential selective-aware graph neural network (PSAGNN) model to simultaneously defend against feature and structural nontarget poisoning attacks ...

Abstract: In this study, we present a novel approach to adversarial attacks for graph neural networks (GNNs), specifically addressing the unique challenges posed by graphical data. Unlike traditional ...

They introduced a novel graph neural network-based architecture, Graph-In, Graph-Out (GIGO)-ToM, designed to predict adversarial targets and attack trajectories across variable-sized and ...