Deep Neural Networks (DNNs) have found extensive applications across diverse fields, such as image classification, speech recognition, and natural language processing. However, their susceptibility to various adversarial attacks, notably the backdoor attack, has repeatedly been demonstrated in recent years. The backdoor attack aims to misclassify inputs with specific trigger pattern(s) into the pre-determined label(s) by training the model on the poisoned dataset. Backdoor attacks on DNNs can lead to severe real-world consequences, e.g., a deep leaning-based classifier in a self-driving car can be backdoored to misclassify a stop sign as a speed limit sign. With an increasing of real-world data being represented as graphs, Graph Neural Networks (GNNs), a subset of the DNNs, have demonstrated remarkable performance in processing graph data. Despite their efficiency, GNNs, similar to other DNNs, are also vulnerable to backdoor attacks, which can lead to severe results, especially when GNNs are applied in security-related scenarios. Although backdoor attacks have been extensively studied in the image domain, we still need dedicated efforts for the graph domain due to the difference between graph data and other data, e.g., images. This thesis embarks on an exploration of backdoor attacks on GNNs. Chapter 2 focuses on designing and investigating backdoor attacks on centralized GNNs. Specifically, we explore the influence of trigger injecting position on the backdoor attack performance on GNNs. To explore this impact, we propose approaches based on explanation techniques on GNNs, which contributes to exploring the interaction between the explainability and robustness of GNNs. Furthermore, we design a clean-label backdoor attack on GNNs to make the poisoned inputs more challenging to be detected. Considering the growing privacy concern, we focus on backdoor attacks on federated GNNs in Chapter 3. We propose a label-only membership inference attack on GNNs in the scenario that the attacker can only get label output from the GNN models. Moreover, we investigate centralized and distributed backdoor attacks on federated GNNs. Besides designing efficient backdoor attacks on GNNs, we also explore the possibility of leveraging backdoor attacks for defensive purposes for GNNs. Chapter 4 introduces a watermarking framework for GNNs based on backdoor attacks. Our research outcomes will deepen the understanding of backdoor attacks on GNNs and push the GNN model designers to develop more secure models. @en

Graph neural networks (GNNs) have gained significant popularity as powerful deep learning methods for processing graph data. However, centralized GNNs face challenges in data-sensitive scenarios due to privacy concerns and regulatory restrictions. Federated learning has emerged as a promising technology that enables collaborative training of a shared global model while preserving privacy. Although federated learning has been applied to train GNNs, no research focuses on the robustness of Federated GNNs against backdoor attacks.

This article bridges this research gap by investigating two types of backdoor attacks in Federated GNNs: centralized backdoor attack (CBA) and distributed backdoor attack (DBA). Through extensive experiments, we demonstrate that DBA exhibits a higher success rate than CBA across various scenarios. To further explore the characteristics of these backdoor attacks in Federated GNNs, we evaluate their performance under different scenarios, including varying numbers of clients, trigger sizes, poisoning intensities, and trigger densities. Additionally, we explore the resilience of DBA and CBA against two defense mechanisms. Our findings reveal that both defenses cannot eliminate DBA and CBA without affecting the original task. This highlights the necessity of developing tailored defenses to mitigate the novel threat of backdoor attacks in Federated GNNs.@en

Graph Neural Networks (GNNs) have achieved promising performance in various real-world applications. Building a powerful GNN model is not a trivial task, as it requires a large amount of training data, powerful computing resources, and human expertise. Moreover, with the development of adversarial attacks, e.g., model stealing attacks, GNNs raise challenges to model authentication. To avoid copyright infringement on GNNs, verifying the ownership of the GNN models is necessary.This paper presents a watermarking framework for GNNs for both graph and node classification tasks. We 1) design two strategies to generate watermarked data for the graph classification task and one for the node classification task, 2) embed the watermark into the host model through training to obtain the watermarked GNN model, and 3) verify the ownership of the suspicious model in a black-box setting. The experiments show that our framework can verify the ownership of GNN models with a very high probability (up to 99%) for both tasks. We also explore our watermarking mechanism against an adaptive attacker with access to partial knowledge of the watermarked data. Finally, we experimentally show that our watermarking approach is robust against a state-of-the-art model extraction technique and four state-of-the-art defenses against backdoor attacks.

@en

Backdoor attacks have been demonstrated as a security threat for machine learning models. Traditional backdoor attacks intend to inject backdoor functionality into the model such that the backdoored model will perform abnormally on inputs with predefined backdoor triggers and still retain state-of-the-art performance on the clean inputs. While there are already some works on backdoor attacks on Graph Neural Networks (GNNs), the backdoor trigger in the graph domain is mostly injected into random positions of the sample. There is no work analyzing and explaining the backdoor attack performance when injecting triggers into the most important or least important area in the sample, which we refer to as trigger-injecting strategies MIAS and LIAS, respectively. Our results show that, generally, LIAS performs better, and the differences between the LIAS and MIAS performance can be significant. Furthermore, we explain these two strategies’ similar (better) attack performance through explanation techniques, which results in a further understanding of backdoor attacks in GNNs.@en

Deep learning found its place in various real-world applications, where many also have security requirements. Unfortunately, as these systems become more pervasive, understanding how they fail becomes more challenging. While there are multiple failure modes in machine learning, one category received significant attention in the last few years-backdoor attacks. Backdoor attacks aim to make a model misclassify some of its inputs to a preset-specific label while other classification results would behave normally. While many works investigate various backdoor attacks and defenses for different domains, no works aim to provide a systematic comparison of backdoor attacks for different scenarios. This work considers backdoor attacks in image, sound, text, and graph domains and provides a comparative analysis of their respective strengths.

@en

Graph Neural Networks (GNNs) are a class of deep learning-based methods for processing graph domain information. GNNs have recently become a widely used graph analysis method due to their superior ability to learn representations for complex graph data. Due to privacy concerns and regulation restrictions, centralized GNNs can be difficult to apply to data-sensitive scenarios. Federated learning (FL) is an emerging technology developed for privacy-preserving settings when several parties need to train a shared global model collaboratively. Although several research works have applied FL to train GNNs (Federated GNNs), there is no research on their robustness to backdoor attacks.

This paper bridges this gap by conducting two types of backdoor attacks in Federated GNNs: centralized backdoor attacks (CBA) and distributed backdoor attacks (DBA). Our experiments show that the DBA attack success rate is higher than CBA in almost all cases. For CBA, the attack success rate of all local triggers is similar to the global trigger, even if the training set of the adversarial party is embedded with the global trigger. To explore the properties of two backdoor attacks in Federated GNNs, we evaluate the attack performance for a different number of clients, trigger sizes, poisoning intensities, and trigger densities. Finally, we explore the robustness of DBA and CBA against two state-of-the-art defenses. We find that both attacks are robust against the investigated defenses, necessitating the need to consider backdoor attacks in Federated GNNs as a novel threat that requires custom defenses.@en

Graph Neural Networks (GNNs), inspired by Convolutional Neural Networks (CNNs), aggregate the message of nodes' neighbors and structure information to acquire expressive representations of nodes for node classification, graph classification, and link prediction. Previous studies have indicated that node-level GNNs are vulnerable to Membership Inference Attacks (MIAs), which infer whether a node is in the training data of GNNs and leak the node's private information, like the patient's disease history. The implementation of previous MIAs takes advantage of the models' probability output, which is infeasible if GNNs only provide the prediction label (label-only) for the input.

In this paper, we propose a label-only MIA against GNNs for node classification with the help of GNNs' flexible prediction mechanism, e.g., obtaining the prediction label of one node even when neighbors' information is unavailable. Our attacking method achieves around 60\% accuracy, precision, and Area Under the Curve (AUC) for most datasets and GNN models, some of which are competitive or even better than state-of-the-art probability-based MIAs implemented under our environment and settings. Additionally, we analyze the influence of the sampling method, model selection approach, and overfitting level on the attack performance of our label-only MIA. All of those three factors have an impact on the attack performance. Then, we consider scenarios where assumptions about the adversary's additional dataset (shadow dataset) and extra information about the target model are relaxed. Even in those scenarios, our label-only MIA achieves a better attack performance in most cases. Finally, we explore the effectiveness of possible defenses, including Dropout, Regularization, Normalization, and Jumping knowledge. None of those four defenses prevent our attack completely. @en

This work explores backdoor attacks for automatic speech recognition systems where we inject inaudible triggers. By doing so, we make the backdoor attack challenging to detect for legitimate users and, consequently, potentially more dangerous. We conduct experiments on two versions of a speech dataset and three neural networks and explore the performance of our attack concerning the duration, position, and type of the trigger. Our results indicate that less than 1% of poisoned data is sufficient to deploy a backdoor attack and reach a 100% attack success rate. We observed that short, non-continuous triggers result in highly successful attacks. Still, since our trigger is inaudible, it can be as long as possible without raising any suspicions making the attack more effective. Finally, we conduct our attack on actual hardware and saw that an adversary could manipulate inference in an Android application by playing the inaudible trigger over the air.

@en

Graph Neural Networks (GNNs) have achieved impressive results in various graph learning tasks. They have found their way into many applications, such as fraud detection, molecular property prediction, or knowledge graph reasoning. However, GNNs have been recently demonstrated to be vulnerable to backdoor attacks. In this work, we explore a new kind of backdoor attack, i.e., a clean-label backdoor attack, on GNNs. Unlike prior backdoor attacks on GNNs in which the adversary can introduce arbitrary, often clearly mislabeled, inputs to the training set, in a clean-label backdoor attack, the resulting poisoned inputs appear to be consistent with their label and thus are less likely to be filtered as outliers. The initial experimental results illustrate that the adversary can achieve a high attack success rate (up to 98.47%) with a clean-label backdoor attack on GNNs for the graph classification task. We hope our work will raise awareness of this attack and inspire novel defenses against it.

@en

The current navigation systems used in many autonomous mobile robotic applications, like unmanned vehicles, are always equipped with various sensors to get accurate navigation results. The key point is to fuse the information from different sensors efficiently. However, different sensors provide asynchronous measurements, some of which even appear to be nonlinear. Moreover, some sensors are vulnerable in specific environments, e.g., GPS signal is likely to work poorly in interior space, underground, and tall buildings. We propose a multi-sensor information fusion method based on a factor graph to fuse all available asynchronous sensor information and efficiently and accurately calculate a navigation solution. Assuming the sensor measurements and navigation states in a navigation system as factor nodes and variable nodes in a factor graph, respectively, the update of the states can be implemented in the framework of the factor graph. The proposed method is experimentally validated using two different datasets. A comparison with Federated Filter, which has been widely used in integrated navigation systems, demonstrates the proposed method’s effectiveness. Additionally, analyzing the navigation results with data loss verifies that the proposed method could achieve sensor plug and play in software.@en

Backdoor attacks represent a serious threat to neural network models. A backdoored model will misclassify the trigger-embedded inputs into an attacker-chosen target label while performing normally on other benign inputs. There are already numerous works on backdoor attacks on neural networks, but only a few works consider graph neural networks (GNNs). As such, there is no intensive research on explaining the impact of trigger injecting position on the performance of backdoor attacks on GNNs. To bridge this gap, we conduct an experimental investigation on the performance of backdoor attacks on GNNs. We apply two powerful GNN explainability approaches to select the optimal trigger injecting position to achieve two attacker objectives - high attack success rate and low clean accuracy drop. Our empirical results on benchmark datasets and state-of-the-art neural network models demonstrate the proposed method's effectiveness in selecting trigger injecting position for backdoor attacks on GNNs. For instance, on the node classification task, the backdoor attack with trigger injecting position selected by GraphLIME reaches over 84% attack success rate with less than 2.5% accuracy drop.

@en