Wei Liu
Research Subjects
Wei Liu has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
Xiaoliang Lv
NARI Group Corporation (State Grid Electric Power Research Institute), Nanjing, China
2 shared publicationsBaohua Zhao
Global Energy Interconnection Research Institute Co., Ltd., Beijing, China
1 shared publication- Binbin Duan
State Key Laboratory of Industrial Control Technology, Zhejiang University, Hangzhou 310027, China
1 shared publication - Bo Li
Beihang University
1 shared publication - Boyan Sun
NARI Group Corporation/State Grid Electric Power Research Institute, Nanjing 210003, China
1 shared publication - Cen Chen
State Grid Henan Electric Power Research Institute, Zhengzhou 450052, China
1 shared publication
Latest Publications
- Open Access
Dual-Graph Fusion with Mutual Knowledge Distillation for Log Anomaly Detection
Advanced persistent threats are complex, stealthy, and long-lasting attacks that can seriously compromise networked systems. System logs provide critical information for APT detection, and provenance graph analysis is commonly used for log anomaly detection. However, existing graph-based methods build provenance graphs from a single perspective, lacking adequate integration of structural and semantic information among entities. To address this issue, we propose a log anomaly detection method based on dual graph modeling and mutual knowledge distillation. The structural branch captures event interactions among entities, while the semantic branch captures associations implied by node text attributes. Relational graph convolution is applied for representation learning in both views. During training, a mutual knowledge distillation framework enables the two branches to serve as teachers and students for each other, facilitating bidirectional knowledge transfer. A reliability estimator based on noise-perturbation invariance assigns higher weights to samples with more trustworthy teacher outputs. Experiments on three public datasets demonstrate that the proposed method achieves statistically significant improvements in Recall on Wget and Cadets, and in Precision on StreamSpot, while maintaining F1-scores comparable to MAGIC across several datasets. Threshold-independent ROC-AUC evaluation further shows that the observed advantages on Wget and Cadets are also reflected in anomaly ranking performance rather than being solely attributable to threshold selection. Ablation experiments validate the contributions of dual-graph modeling, mutual distillation, and the reliability estimation mechanism.
Open AccessDual-Graph Fusion with Mutual Knowledge Distillation for Log Anomaly Detection
Advanced persistent threats are complex, stealthy, and long-lasting attacks that can seriously compromise networked systems. System logs provide critical information for APT detection, and provenance graph analysis is commonly used for log anomaly detection. However, existing graph-based methods build provenance graphs from a single perspective, lacking adequate integration of structural and semantic information among entities. To address this issue, we propose a log anomaly detection method based on dual graph modeling and mutual knowledge distillation. The structural branch captures event interactions among entities, while the semantic branch captures associations implied by node text attributes. Relational graph convolution is applied for representation learning in both views. During training, a mutual knowledge distillation framework enables the two branches to serve as teachers and students for each other, facilitating bidirectional knowledge transfer. A reliability estimator based on noise-perturbation invariance assigns higher weights to samples with more trustworthy teacher outputs. Experiments on three public datasets demonstrate that the proposed method achieves statistically significant improvements in Recall on Wget and Cadets, and in Precision on StreamSpot, while maintaining F1-scores comparable to MAGIC across several datasets. Threshold-independent ROC-AUC evaluation further shows that the observed advantages on Wget and Cadets are also reflected in anomaly ranking performance rather than being solely attributable to threshold selection. Ablation experiments validate the contributions of dual-graph modeling, mutual distillation, and the reliability estimation mechanism.
- Open Access
- 20 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 651-658
A physical signal-based anomaly detection for industrial terminal
With the development of industry, industrial control systems are widely used in critical infrastructures. As one of the most significant components, the industrial terminals include Programmable Logic Controller, Distributed Control Systems, Remote Terminal Unit, and etc. Attacks against industrial terminals can cause catastrophic damage since they are directly related to business in the field. Unfortunately, traditional protection methods, such as intrusion detection systems and antivirus software, are not applicable to industrial terminals which are resource-constrained. In this paper, we propose an anomaly detection method for industrial controllers based on physical signals. The basic idea is to detect abnormal execution in a industrial terminal through analyzing its physical signals, including system power consumption and CPU usage. To effectively analyze the physical signals, a Transformer-based neural network is trained on the time series dataset to predict the next normal physical signals. The abnormal execution is detected by comparing the predicted signals and actual ones. We evaluate the proposed method on a real world dataset collected from a smart grid terminal, and the results show that the detection accuracy is over 97%.
- 20 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 651-658
Open AccessA physical signal-based anomaly detection for industrial terminal
With the development of industry, industrial control systems are widely used in critical infrastructures. As one of the most significant components, the industrial terminals include Programmable Logic Controller, Distributed Control Systems, Remote Terminal Unit, and etc. Attacks against industrial terminals can cause catastrophic damage since they are directly related to business in the field. Unfortunately, traditional protection methods, such as intrusion detection systems and antivirus software, are not applicable to industrial terminals which are resource-constrained. In this paper, we propose an anomaly detection method for industrial controllers based on physical signals. The basic idea is to detect abnormal execution in a industrial terminal through analyzing its physical signals, including system power consumption and CPU usage. To effectively analyze the physical signals, a Transformer-based neural network is trained on the time series dataset to predict the next normal physical signals. The abnormal execution is detected by comparing the predicted signals and actual ones. We evaluate the proposed method on a real world dataset collected from a smart grid terminal, and the results show that the detection accuracy is over 97%.
- Open Access
- 9 October 2023
- Published by MDPI AG in Electronics
- Vol. 12 (19) , 4182
Survey on Application of Trusted Computing in Industrial Control Systems
The Fourth Industrial Revolution, also known as Industrial 4.0, has greatly accelerated inter-connectivity and smart automation in industrial control systems (ICSs), which has introduced new challenges to their security. With the fast growth of the Internet of Things and the advent of 5G/6G, the collaboration of Artificial Intelligence (Al) and the Internet of Things (loT) in ICSs has also introduced lots of security issues as it highly relies on advanced communication and networking techniques. Frequent ICS security incidents have demonstrated that attackers have the ability to stealthily breach the current system defenses and cause catastrophic effects to ICSs. Thankfully, trusted computing technology, which has been a popular research topic in the field of information security in recent years, offers distinct advantages when applied to ICSs. In this paper, we first analyze the vulnerabilities of ICSs and the limitations of existing protection technologies. Then, we introduce the concept of trusted computing and present a security framework for ICSs based on Trusted Computing 3.0. Finally, we discuss potential future research directions.
- 9 October 2023
- Published by MDPI AG in Electronics
- Vol. 12 (19) , 4182
Open AccessSurvey on Application of Trusted Computing in Industrial Control Systems
The Fourth Industrial Revolution, also known as Industrial 4.0, has greatly accelerated inter-connectivity and smart automation in industrial control systems (ICSs), which has introduced new challenges to their security. With the fast growth of the Internet of Things and the advent of 5G/6G, the collaboration of Artificial Intelligence (Al) and the Internet of Things (loT) in ICSs has also introduced lots of security issues as it highly relies on advanced communication and networking techniques. Frequent ICS security incidents have demonstrated that attackers have the ability to stealthily breach the current system defenses and cause catastrophic effects to ICSs. Thankfully, trusted computing technology, which has been a popular research topic in the field of information security in recent years, offers distinct advantages when applied to ICSs. In this paper, we first analyze the vulnerabilities of ICSs and the limitations of existing protection technologies. Then, we introduce the concept of trusted computing and present a security framework for ICSs based on Trusted Computing 3.0. Finally, we discuss potential future research directions.
- Open Access
- 9 October 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 6, 57778-57786
Active Defense Technology of Power Monitoring System With Adaptive Features
Internet of Things has found a lot of applications in power systems. However, with the increased the sensing, networking, and control capabilities, the security issues have become even more urgent at the same time. In this paper, a TMAC model based on trusted mandatory access control is proposed by studying the security situation of a power monitoring system. The model has self-learning characteristic and can realize the automatic escalation of the global security strategy based on intelligent agent, so as to build the safe immune ability and active defense system for the power monitoring. This paper introduces the key technology of the TMAC security model, formalizes some of the work, and finally tests its credibility and effectiveness for the typical application scenario of the power monitoring system. Through the study of this technology, the power monitoring system is further equipped with an immune ability against virus Trojan and hacker attacks, especially for application scenarios, such as substation, power plant, and master station.
- 9 October 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 6, 57778-57786
Open AccessActive Defense Technology of Power Monitoring System With Adaptive Features
Internet of Things has found a lot of applications in power systems. However, with the increased the sensing, networking, and control capabilities, the security issues have become even more urgent at the same time. In this paper, a TMAC model based on trusted mandatory access control is proposed by studying the security situation of a power monitoring system. The model has self-learning characteristic and can realize the automatic escalation of the global security strategy based on intelligent agent, so as to build the safe immune ability and active defense system for the power monitoring. This paper introduces the key technology of the TMAC security model, formalizes some of the work, and finally tests its credibility and effectiveness for the typical application scenario of the power monitoring system. Through the study of this technology, the power monitoring system is further equipped with an immune ability against virus Trojan and hacker attacks, especially for application scenarios, such as substation, power plant, and master station.
- 28 September 2018
- Published by Springer Nature in Journal of Signal Processing Systems
- Vol. 91 (10) , 1127-1136
Research on Microkernel-based Power Dedicated Secure Operating System
- 28 September 2018
- Published by Springer Nature in Journal of Signal Processing Systems
- Vol. 91 (10) , 1127-1136
Research on Microkernel-based Power Dedicated Secure Operating System