Junyuan Fang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Junyuan Fang has a total of 18 co-authors. Here are the top co-authors by number of shared publications:
- Jiajing Wu
Sun Yat-sen University, Zhuhai, China
11 shared publications - Zibin Zheng
Sun Yat-Sen University, Guangzhou, China
9 shared publications - Chi Kong Tse
City University of Hong Kong
7 shared publications - Han Yang
School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China
4 shared publications - Biwei Li
Department of Electrical Engineering, City University of Hong Kong, Kowloon Tong, Hong Kong
1 shared publication - Dong Liu
Department of Electrical Engineering, City University of Hong Kong, Kowloon, Hong Kong SAR, China
1 shared publication
Latest Publications
- Open Access
Unnoticeable Community Deception via Multi-objective Optimization
Community detection in graphs is crucial for understanding the organization of nodes into densely connected clusters. While numerous strategies have been developed to identify these clusters, the success of community detection can lead to privacy and information security concerns, as individuals may not want their personal information exposed. To address this, community deception methods have been proposed to reduce the effectiveness of detection algorithms. Nevertheless, several limitations, such as the rationality of evaluation metrics and the unnoticeability of attacks, have been ignored in current deception methods. Therefore, in this work, we first investigate the limitations of the widely used deception metric, i.e., the decrease of modularity, through empirical studies. Then, we propose a new deception metric, and combine this new metric together with the attack budget to model the unnoticeable community deception task as a multi-objective optimization problem. To further improve the deception performance, we propose two variant methods by incorporating the degree-biased and community-biased candidate node selection mechanisms. Extensive experiments on three benchmark datasets demonstrate the superiority of the proposed community deception strategies.
Open AccessUnnoticeable Community Deception via Multi-objective Optimization
Community detection in graphs is crucial for understanding the organization of nodes into densely connected clusters. While numerous strategies have been developed to identify these clusters, the success of community detection can lead to privacy and information security concerns, as individuals may not want their personal information exposed. To address this, community deception methods have been proposed to reduce the effectiveness of detection algorithms. Nevertheless, several limitations, such as the rationality of evaluation metrics and the unnoticeability of attacks, have been ignored in current deception methods. Therefore, in this work, we first investigate the limitations of the widely used deception metric, i.e., the decrease of modularity, through empirical studies. Then, we propose a new deception metric, and combine this new metric together with the attack budget to model the unnoticeable community deception task as a multi-objective optimization problem. To further improve the deception performance, we propose two variant methods by incorporating the degree-biased and community-biased candidate node selection mechanisms. Extensive experiments on three benchmark datasets demonstrate the superiority of the proposed community deception strategies.
- Open Access
Quantifying the Noise of Structural Perturbations on Graph Adversarial Attacks
Graph neural networks have been widely utilized to solve graph-related tasks because of their strong learning power in utilizing the local information of neighbors. However, recent studies on graph adversarial attacks have proven that current graph neural networks are not robust against malicious attacks. Yet much of the existing work has focused on the optimization objective based on attack performance to obtain (near) optimal perturbations, but paid less attention to the strength quantification of each perturbation such as the injection of a particular node/link, which makes the choice of perturbations a black-box model that lacks interpretability. In this work, we propose the concept of noise to quantify the attack strength of each adversarial link. Furthermore, we propose three attack strategies based on the defined noise and classification margins in terms of single and multiple steps optimization. Extensive experiments conducted on benchmark datasets against three representative graph neural networks demonstrate the effectiveness of the proposed attack strategies. Particularly, we also investigate the preferred patterns of effective adversarial perturbations by analyzing the corresponding properties of the selected perturbation nodes.
Open AccessQuantifying the Noise of Structural Perturbations on Graph Adversarial Attacks
Graph neural networks have been widely utilized to solve graph-related tasks because of their strong learning power in utilizing the local information of neighbors. However, recent studies on graph adversarial attacks have proven that current graph neural networks are not robust against malicious attacks. Yet much of the existing work has focused on the optimization objective based on attack performance to obtain (near) optimal perturbations, but paid less attention to the strength quantification of each perturbation such as the injection of a particular node/link, which makes the choice of perturbations a black-box model that lacks interpretability. In this work, we propose the concept of noise to quantify the attack strength of each adversarial link. Furthermore, we propose three attack strategies based on the defined noise and classification margins in terms of single and multiple steps optimization. Extensive experiments conducted on benchmark datasets against three representative graph neural networks demonstrate the effectiveness of the proposed attack strategies. Particularly, we also investigate the preferred patterns of effective adversarial perturbations by analyzing the corresponding properties of the selected perturbation nodes.
- 21 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 55 (6) , 3989-4000
What Contributes More to the Robustness of Heterophilic Graph Neural Networks?
In recent years, graph neural networks (GNNs) have gained significant attention due to their outstanding performance on graph-related tasks by utilizing neighborhood aggregation. However, traditional GNNs are primarily designed based on the homophily assumption, which means that they show poor performance on heterophilic networks where dissimilar nodes prefer to connect. To address this issue, several heterophilic GNNs have been proposed that employ techniques, such as extending local neighbors and improving GNN architectures. From another perspective, recent studies have shown that, unlike homophilic GNNs, heterophilic GNNs exhibit higher robustness against graph adversarial attacks. In these attacks, the attackers try to inject small perturbations into the graph. Therefore, in this study, we delve into the core designs of heterophilic GNNs, including high-order neighbor, potential neighbor, ego-neighbor separation, and interlayer combination designs. We further analyze the influence of these key designs on the robustness of GNNs. We conducted comprehensive experiments to compare the impact of different designs on baseline and real-world GNN models. The findings of this work can serve as a reference for future studies aiming to design more robust and universal GNNs.
- 21 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 55 (6) , 3989-4000
What Contributes More to the Robustness of Heterophilic Graph Neural Networks?
In recent years, graph neural networks (GNNs) have gained significant attention due to their outstanding performance on graph-related tasks by utilizing neighborhood aggregation. However, traditional GNNs are primarily designed based on the homophily assumption, which means that they show poor performance on heterophilic networks where dissimilar nodes prefer to connect. To address this issue, several heterophilic GNNs have been proposed that employ techniques, such as extending local neighbors and improving GNN architectures. From another perspective, recent studies have shown that, unlike homophilic GNNs, heterophilic GNNs exhibit higher robustness against graph adversarial attacks. In these attacks, the attackers try to inject small perturbations into the graph. Therefore, in this study, we delve into the core designs of heterophilic GNNs, including high-order neighbor, potential neighbor, ego-neighbor separation, and interlayer combination designs. We further analyze the influence of these key designs on the robustness of GNNs. We conducted comprehensive experiments to compare the impact of different designs on baseline and real-world GNN models. The findings of this work can serve as a reference for future studies aiming to design more robust and universal GNNs.
- 10 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Systems Journal
- Vol. 19 (1) , 258-269
Failure Propagation Graphs for Studying Cascading Failure Propagation in Power Networks
Cascading failure, characterized by the widespread propagation of failure events, is a common cause for severe blackouts in power networks. Strengthening critical branches in a power network is crucial for mitigating the risk of blackouts resulting from cascading failures. In this article, we propose a time-efficient greedy search method to identify critical branches in a power network. We address the challenge of computational constraints by using a failure propagation graph, which accurately captures the critical failure propagation patterns based on cascading failure simulation. Our approach minimizes cascading failure risk while strategically reinforcing a limited number of branches. The failure-propagation-graph greedy-search (FPG-GS) algorithm selects candidate branches based on cascading failure simulation and iteratively identifies the most crucial branches. Our experimental results on different power systems demonstrate the superior performance and efficiency of the FPG-GS algorithm compared to existing methods. In addition, our study highlights the importance of strategic branch selection, showing that reinforcing one-fifth of the branches can achieve a mitigation rate exceeding 80%.
- 10 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Systems Journal
- Vol. 19 (1) , 258-269
Failure Propagation Graphs for Studying Cascading Failure Propagation in Power Networks
Cascading failure, characterized by the widespread propagation of failure events, is a common cause for severe blackouts in power networks. Strengthening critical branches in a power network is crucial for mitigating the risk of blackouts resulting from cascading failures. In this article, we propose a time-efficient greedy search method to identify critical branches in a power network. We address the challenge of computational constraints by using a failure propagation graph, which accurately captures the critical failure propagation patterns based on cascading failure simulation. Our approach minimizes cascading failure risk while strategically reinforcing a limited number of branches. The failure-propagation-graph greedy-search (FPG-GS) algorithm selects candidate branches based on cascading failure simulation and iteratively identifies the most crucial branches. Our experimental results on different power systems demonstrate the superior performance and efficiency of the FPG-GS algorithm compared to existing methods. In addition, our study highlights the importance of strategic branch selection, showing that reinforcing one-fifth of the branches can achieve a mitigation rate exceeding 80%.
- 20 June 2024
- Published by Springer Nature
- p. 101-124
Account Classification Based on the Homophily-Heterophily Graph Neural Networks
- 20 June 2024
- Published by Springer Nature
- p. 101-124
Account Classification Based on the Homophily-Heterophily Graph Neural Networks
- 22 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. 11 (4) , 5374-5387
GANI: Global Attacks on Graph Neural Networks via Imperceptible Node Injections
Graph neural networks (GNNs) have found successful applications in various graph-related tasks. However, recent studies have shown that many GNNs are vulnerable to adversarial attacks. In a vast majority of existing studies, adversarial attacks on GNNs are launched via direct modification of the original graph such as adding/removing links, which may not be applicable in practice. In this article, we focus on a realistic attack operation via injecting fake nodes. The proposed global attack strategy via node injection (GANI) is designed under the comprehensive consideration of an unnoticeable perturbation setting from both structure and feature domains. Specifically, to make the node injections as imperceptible and effective as possible, we propose a sampling operation to determine the degree of the newly injected nodes, and then generate features and select neighbors for these injected nodes based on the statistical information of features and evolutionary perturbations obtained from a genetic algorithm, respectively. In particular, the proposed feature generation mechanism is suitable for both binary and continuous node features. Extensive experimental results on benchmark datasets against both general and defended GNNs show strong attack performance of GANI. Moreover, the imperceptibility analyses also demonstrate that GANI achieves a relatively unnoticeable injection on benchmark datasets.
- 22 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. 11 (4) , 5374-5387
GANI: Global Attacks on Graph Neural Networks via Imperceptible Node Injections
Graph neural networks (GNNs) have found successful applications in various graph-related tasks. However, recent studies have shown that many GNNs are vulnerable to adversarial attacks. In a vast majority of existing studies, adversarial attacks on GNNs are launched via direct modification of the original graph such as adding/removing links, which may not be applicable in practice. In this article, we focus on a realistic attack operation via injecting fake nodes. The proposed global attack strategy via node injection (GANI) is designed under the comprehensive consideration of an unnoticeable perturbation setting from both structure and feature domains. Specifically, to make the node injections as imperceptible and effective as possible, we propose a sampling operation to determine the degree of the newly injected nodes, and then generate features and select neighbors for these injected nodes based on the statistical information of features and evolutionary perturbations obtained from a genetic algorithm, respectively. In particular, the proposed feature generation mechanism is suitable for both binary and continuous node features. Extensive experimental results on benchmark datasets against both general and defended GNNs show strong attack performance of GANI. Moreover, the imperceptibility analyses also demonstrate that GANI achieves a relatively unnoticeable injection on benchmark datasets.
- 26 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 19 (15566013) , 3038-3050
Fishing for Fraudsters: Uncovering Ethereum Phishing Gangs With Blockchain Data
As one of the most typical cybercrime types, phishing scams have extended the devil’s hand to the emerging blockchain ecosystem in recent years. Especially huge economic losses have been caused by phishing scams in Ethereum, the second-largest blockchain system. Existing approaches for Ethereum phishing detection, however, typically use machine learning or transaction graph embedding methods to identify phishers in isolation and do not effectively uncover the group of transaction accounts linked to scams (which we term a “gang”). Since accounts are pseudonymous in Ethereum, these undisclosed conspirator accounts have potential risks to the system. In this paper, we conduct the first study that characterizes and detects Ethereum phishing gangs. We first investigate the transaction behaviors in phishing gangs from the perspectives of individuals, pairs, and higher-order patterns. Our analysis reveals that although the Ethereum transaction graph is sparse with a highly skewed degree distribution, phishing accounts in the same gang have closer relationships and share specific transaction patterns. Based on our findings, we formalize the phishing gang detection problem and introduce a novel detection model named PGDetector. Given a risky phishing account as a seed, PGDetector can find out the potential risky accounts sharing close relationships within the seed’s community based on genetic algorithm optimization. Experimental results on large-scale Ethereum transaction data demonstrate the effectiveness of PGDetector.
- 26 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 19 (15566013) , 3038-3050
Fishing for Fraudsters: Uncovering Ethereum Phishing Gangs With Blockchain Data
As one of the most typical cybercrime types, phishing scams have extended the devil’s hand to the emerging blockchain ecosystem in recent years. Especially huge economic losses have been caused by phishing scams in Ethereum, the second-largest blockchain system. Existing approaches for Ethereum phishing detection, however, typically use machine learning or transaction graph embedding methods to identify phishers in isolation and do not effectively uncover the group of transaction accounts linked to scams (which we term a “gang”). Since accounts are pseudonymous in Ethereum, these undisclosed conspirator accounts have potential risks to the system. In this paper, we conduct the first study that characterizes and detects Ethereum phishing gangs. We first investigate the transaction behaviors in phishing gangs from the perspectives of individuals, pairs, and higher-order patterns. Our analysis reveals that although the Ethereum transaction graph is sparse with a highly skewed degree distribution, phishing accounts in the same gang have closer relationships and share specific transaction patterns. Based on our findings, we formalize the phishing gang detection problem and introduce a novel detection model named PGDetector. Given a risky phishing account as a seed, PGDetector can find out the potential risky accounts sharing close relationships within the seed’s community based on genetic algorithm optimization. Experimental results on large-scale Ethereum transaction data demonstrate the effectiveness of PGDetector.
- 14 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Emerging and Selected Topics in Circuits and Systems
- Vol. 13 (3) , 829-840
Both Homophily and Heterophily Matter: Bi-Path Aware Graph Neural Network for Ethereum Account Classification
In recent years, the cryptocurrency market has been booming with an ever-increasing market capitalization. However, due to the anonymity of blockchain technology, this market has become a hotbed of financial crimes. As the largest blockchain platform supporting smart contracts, financial crimes including scams and hacking frequently happen on Ethereum and have caused serious losses. Therefore, it is necessary to classify Ethereum accounts in order to better identify those involved in illegal transactions and analyze the behavior patterns of different classes of accounts. In this paper, we construct an Ethereum transaction network based on transaction records and find that this network is with heterophily. However, most of the current work on account classification ignores the role of this heterophily information. We first figure out that the heterophily information of the neighborhood may also be beneficial for the final predictions. Based on this, we propose a new graph neural network (GNN) model, named BPA-GNN, which incorporates both homophilic and heterophilic information into the neighborhood aggregations. Specifically, BPA-GNN consists of three main modules including bi-path neighbor sampling, separated neighborhood aggregation, and attention-based node representation learning. Comprehensive experiments on a real Ethereum transaction dataset demonstrate the state-of-the-art performance of BPA-GNN, showing that the model can effectively extract and utilize neighborhood information to improve the distinguishability of node representations. As an effective solution for Ethereum account de-anonymization, BPA-GNN can help identify illegal activities and promote the healthy development of the Ethereum ecosystem.
- 14 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Emerging and Selected Topics in Circuits and Systems
- Vol. 13 (3) , 829-840
Both Homophily and Heterophily Matter: Bi-Path Aware Graph Neural Network for Ethereum Account Classification
In recent years, the cryptocurrency market has been booming with an ever-increasing market capitalization. However, due to the anonymity of blockchain technology, this market has become a hotbed of financial crimes. As the largest blockchain platform supporting smart contracts, financial crimes including scams and hacking frequently happen on Ethereum and have caused serious losses. Therefore, it is necessary to classify Ethereum accounts in order to better identify those involved in illegal transactions and analyze the behavior patterns of different classes of accounts. In this paper, we construct an Ethereum transaction network based on transaction records and find that this network is with heterophily. However, most of the current work on account classification ignores the role of this heterophily information. We first figure out that the heterophily information of the neighborhood may also be beneficial for the final predictions. Based on this, we propose a new graph neural network (GNN) model, named BPA-GNN, which incorporates both homophilic and heterophilic information into the neighborhood aggregations. Specifically, BPA-GNN consists of three main modules including bi-path neighbor sampling, separated neighborhood aggregation, and attention-based node representation learning. Comprehensive experiments on a real Ethereum transaction dataset demonstrate the state-of-the-art performance of BPA-GNN, showing that the model can effectively extract and utilize neighborhood information to improve the distinguishability of node representations. As an effective solution for Ethereum account de-anonymization, BPA-GNN can help identify illegal activities and promote the healthy development of the Ethereum ecosystem.
- 15 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. 10 (6) , 3512-3523
Hide and Seek: An Adversarial Hiding Approach Against Phishing Detection on Ethereum
With the wide application and development of blockchain technology, the past years have witnessed the emergence of various cybercrimes, which have caused a huge amount of economic loss. Among them, phishing scams on the blockchain are regarded as a serious threat to the trading security of the blockchain ecosystem. By modeling the transaction data of blockchain as a network, a series of graph-based phishing detection frameworks have been proposed. Enlightened by adversarial attacks of graph data, we propose to verify the robustness of current phishing detection frameworks under intentional attackers aiming to hide phishing behaviors. In this study, we first propose a general phishing detection framework based on feature engineering and then propose a phishing hiding framework combing the greedy selection mechanism with four phishing hiding strategies to measure the robustness of the proposed general detection models. Extensive experiments evaluate the detective performance of the phishing detection model and its robustness against the hiding framework. The experimental results indicate that the detective model based on feature engineering is rather fragile under adversarial attacks.
- 15 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. 10 (6) , 3512-3523
Hide and Seek: An Adversarial Hiding Approach Against Phishing Detection on Ethereum
With the wide application and development of blockchain technology, the past years have witnessed the emergence of various cybercrimes, which have caused a huge amount of economic loss. Among them, phishing scams on the blockchain are regarded as a serious threat to the trading security of the blockchain ecosystem. By modeling the transaction data of blockchain as a network, a series of graph-based phishing detection frameworks have been proposed. Enlightened by adversarial attacks of graph data, we propose to verify the robustness of current phishing detection frameworks under intentional attackers aiming to hide phishing behaviors. In this study, we first propose a general phishing detection framework based on feature engineering and then propose a phishing hiding framework combing the greedy selection mechanism with four phishing hiding strategies to measure the robustness of the proposed general detection models. Extensive experiments evaluate the detective performance of the phishing detection model and its robustness against the hiding framework. The experimental results indicate that the detective model based on feature engineering is rather fragile under adversarial attacks.
- 22 March 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems II: Express Briefs
- Vol. 69 (5) , 2523-2527
Analyzing Robustness of Complex Networks Against Incomplete Information
Large scale networked systems are playing an indispensable role in modern society, and thus the robustness of these systems against random failures or malicious attacks has become a critical research issue. As a major threat to network robustness, cascading failures have attracted increasing research attention in the past decades. Previous studies have put forward many heuristic methods to investigate the network robustness against cascading failure problems. However, most of them assume that the attackers can obtain the complete topology information of the network, which may not be available in practice. To tackle this problem, we use the link prediction methods to restore the missing information (i.e., the topological structure of networks) of the network first, and then utilize the predicted information to further help distinguish the critical nodes in the network systems from the attacker’s perspective. Simulation results on both synthetic and real-world networks have demonstrated the effectiveness of the proposed method.
- 22 March 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems II: Express Briefs
- Vol. 69 (5) , 2523-2527
Analyzing Robustness of Complex Networks Against Incomplete Information
Large scale networked systems are playing an indispensable role in modern society, and thus the robustness of these systems against random failures or malicious attacks has become a critical research issue. As a major threat to network robustness, cascading failures have attracted increasing research attention in the past decades. Previous studies have put forward many heuristic methods to investigate the network robustness against cascading failure problems. However, most of them assume that the attackers can obtain the complete topology information of the network, which may not be available in practice. To tackle this problem, we use the link prediction methods to restore the missing information (i.e., the topological structure of networks) of the network first, and then utilize the predicted information to further help distinguish the critical nodes in the network systems from the attacker’s perspective. Simulation results on both synthetic and real-world networks have demonstrated the effectiveness of the proposed method.