Limin Sun
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
Limin Sun has a total of 296 co-authors. Here are the top co-authors by number of shared publications:
- Hong Li
Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China
29 shared publications - Hongsong Zhu
Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China
21 shared publications - Shichao Lv
Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China
18 shared publications - Zhi Li
Institute of Information Engineering, Chinese Academy of Sciences
18 shared publications - Zhiqiang Shi
Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China
17 shared publications - Dongliang Fang
Chinese Academy of Sciences
15 shared publications
Latest Publications
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. 22 (11) , 9729-9740
Detecting Counterfeits Programmable Logic Controllers via Electrical Fingerprint
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. 22 (11) , 9729-9740
Detecting Counterfeits Programmable Logic Controllers via Electrical Fingerprint
- 19 July 2026
- Published by Springer Nature
- p. 38-49
AI-based Cyberspace Mapping: Inference of HTTP(S) Message Ownership Information Based on LLM Fine-Tuning
- 19 July 2026
- Published by Springer Nature
- p. 38-49
AI-based Cyberspace Mapping: Inference of HTTP(S) Message Ownership Information Based on LLM Fine-Tuning
- 2 July 2026
- Published by Springer Nature
- p. 204-222
LLM-Augmented Semantic Analysis for Hidden Web Interface Discovery in IoT Devices
- 2 July 2026
- Published by Springer Nature
- p. 204-222
LLM-Augmented Semantic Analysis for Hidden Web Interface Discovery in IoT Devices
- 3 June 2026
- Published by Springer Nature
- p. 264-279
DepFieldGen: Research on Dependent Field Generation for Secure ICS Protocols
- 3 June 2026
- Published by Springer Nature
- p. 264-279
DepFieldGen: Research on Dependent Field Generation for Secure ICS Protocols
- Open Access
- 27 April 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
Fixseeker : An Empirical Driven Graph-based Approach for Detecting Silent Vulnerability Fixes in Open Source Software
Open source software (OSS) vulnerabilities pose significant security risks to downstream applications. While vulnerability databases provide valuable information for mitigation, many security patches are released silently in new commits of OSS repositories without explicit indications of their security impact. This makes it challenging for software maintainers and users to detect and address these vulnerabilities. There are a few approaches for detecting vulnerability-fixing commits (VFCs), but most of these approaches leverage commit messages, which would miss silent VFCs. On the other hand, there are some approaches for detecting silent VFCs based on code change patterns, but they often fail to characterize vulnerability fix patterns, thereby lacking effectiveness. For example, some approaches analyze each hunk in known VFCs, in isolation, to learn vulnerability fix patterns; but vulnerability fixes are often associated with multiple hunks, in which cases correlations of code changes across those hunks are essential for characterizing the vulnerability fixes. To address these problems, we first conduct a large-scale empirical study on 11,900 VFCs across six programming languages, in which we found that over 70% of VFCs involve multiple hunks with various types of correlations. Based on our findings, we propose Fixseeker , a graph-based approach that extracts the various correlations between code changes at the hunk level to detect silent vulnerability fixes. Our evaluation demonstrates that Fixseeker outperforms state-of-the-art approaches across multiple programming languages, achieving a high F1 score of 0.818 on average in balanced datasets and consistently improving F2 score, AUC-ROC, and AUC-PR scores by 10.64%, 5.34%, and 10.34% on imbalanced datasets compared to the best baseline methods. Our evaluation also indicates the generality of Fixseeker across different vulnerability types and repository sizes.
- 27 April 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
Open AccessFixseeker : An Empirical Driven Graph-based Approach for Detecting Silent Vulnerability Fixes in Open Source Software
Open source software (OSS) vulnerabilities pose significant security risks to downstream applications. While vulnerability databases provide valuable information for mitigation, many security patches are released silently in new commits of OSS repositories without explicit indications of their security impact. This makes it challenging for software maintainers and users to detect and address these vulnerabilities. There are a few approaches for detecting vulnerability-fixing commits (VFCs), but most of these approaches leverage commit messages, which would miss silent VFCs. On the other hand, there are some approaches for detecting silent VFCs based on code change patterns, but they often fail to characterize vulnerability fix patterns, thereby lacking effectiveness. For example, some approaches analyze each hunk in known VFCs, in isolation, to learn vulnerability fix patterns; but vulnerability fixes are often associated with multiple hunks, in which cases correlations of code changes across those hunks are essential for characterizing the vulnerability fixes. To address these problems, we first conduct a large-scale empirical study on 11,900 VFCs across six programming languages, in which we found that over 70% of VFCs involve multiple hunks with various types of correlations. Based on our findings, we propose Fixseeker , a graph-based approach that extracts the various correlations between code changes at the hunk level to detect silent vulnerability fixes. Our evaluation demonstrates that Fixseeker outperforms state-of-the-art approaches across multiple programming languages, achieving a high F1 score of 0.818 on average in balanced datasets and consistently improving F2 score, AUC-ROC, and AUC-PR scores by 10.64%, 5.34%, and 10.34% on imbalanced datasets compared to the best baseline methods. Our evaluation also indicates the generality of Fixseeker across different vulnerability types and repository sizes.
- Open Access
- 23 April 2026
- Published by Springer Nature in Cybersecurity
- Vol. 9 (1) , 158
Resource-based online orchestration for multi-domain collaborative analysis of network encrypted traffic
In recent years, cloud-edge-end collaborative federated learning frameworks have been widely used in many scenarios and achieved good results. However, with the complexity of application requirements, the problems of device heterogeneity and data heterogeneity become more prominent. Traditional frameworks often face challenges such as uneven allocation of computational resources and inefficient training when dealing with these problems. To address this problem, this paper proposes a novel federated learning framework for multi-domain collaborative analysis of networked encrypted flows. First, we split the model training tasks, intelligently assign part of the model training tasks to terminal devices based on their performance, while the remaining model training tasks that require more computational resources are handed over to edge servers. Second, we introduce a resource scheduling scheme among edge servers to reasonably allocate model training tasks and fully utilize resources. Finally, high quality global models are obtained through a weight-enabled global model aggregation scheme. Experiments show that our proposed scheme can effectively address the impact of device heterogeneity and data heterogeneity in encrypted traffic identification in cross-domain networks, and improve the training efficiency and model performance of the overall system while ensuring data privacy and security.
- 23 April 2026
- Published by Springer Nature in Cybersecurity
- Vol. 9 (1) , 158
Open AccessResource-based online orchestration for multi-domain collaborative analysis of network encrypted traffic
In recent years, cloud-edge-end collaborative federated learning frameworks have been widely used in many scenarios and achieved good results. However, with the complexity of application requirements, the problems of device heterogeneity and data heterogeneity become more prominent. Traditional frameworks often face challenges such as uneven allocation of computational resources and inefficient training when dealing with these problems. To address this problem, this paper proposes a novel federated learning framework for multi-domain collaborative analysis of networked encrypted flows. First, we split the model training tasks, intelligently assign part of the model training tasks to terminal devices based on their performance, while the remaining model training tasks that require more computational resources are handed over to edge servers. Second, we introduce a resource scheduling scheme among edge servers to reasonably allocate model training tasks and fully utilize resources. Finally, high quality global models are obtained through a weight-enabled global model aggregation scheme. Experiments show that our proposed scheme can effectively address the impact of device heterogeneity and data heterogeneity in encrypted traffic identification in cross-domain networks, and improve the training efficiency and model performance of the overall system while ensuring data privacy and security.
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2824-2834
Breaking Cross-modal Alignment in Embodied Intelligence: A Multimodal Adversarial Attack Framework for Vision-Language-Action Models
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2824-2834
Open AccessBreaking Cross-modal Alignment in Embodied Intelligence: A Multimodal Adversarial Attack Framework for Vision-Language-Action Models
- Top Cited Papers
- 21 March 2026
- Published by Elsevier in Computer Networks
- Vol. 282
HoneyGPT: Breaking the trilemma in honeypots with large language models
- 21 March 2026
- Published by Elsevier in Computer Networks
- Vol. 282
Top Cited PapersHoneyGPT: Breaking the trilemma in honeypots with large language models
- 26 February 2026
- Published by Elsevier in Computer Networks
- Vol. 279
NetID-GPT: Adapting large language models for large-scale internet-connected device identification
- 26 February 2026
- Published by Elsevier in Computer Networks
- Vol. 279
NetID-GPT: Adapting large language models for large-scale internet-connected device identification
- 19 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 21 (15566013) , 2403-2415
LLMBA: Efficient Behavior Analytics via Large Pretrained Models in Zero Trust Networks
Guided by the principle of “Never Trust, Always Verify”, Zero Trust Architecture (ZTA) mandates continuous monitoring and analysis of users and entities, highlighting the critical role of behavior analytics. However, the growing volume of audit data and its complex contextual information render many existing behavior analytics methods insufficient. Moreover, most approaches rely on high-quality labeled data for supervised training, limiting their effectiveness against previously unseen malicious behaviors. To address these challenges, we propose the Large Language Model for Behavior Analytics (LLMBA) framework. LLMBA leverages a Large Language Model (LLM) to analyze behavioral patterns of internal users and entities, capitalizing on the LLM’s strong ability to model sequential data. We introduce a multi-level behavior encoding scheme to capture both contextual and temporal information from behavior records, producing rich input representations for the LLM-enhanced model. The LLM is fine-tuned using self-supervised learning, enabling the detection of unknown malicious behaviors. To reduce the computational and storage overhead inherent in LLMs, we apply knowledge distillation to compress the model while maintaining high detection performance. Extensive experiments on the CERT Insider Threat dataset demonstrate that LLMBA outperforms state-of-the-art baselines in detection accuracy. Furthermore, the compressed student model achieves superior performance compared with existing methods under comparable runtime constraints, making LLMBA highly suitable for real-world deployment.
- 19 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 21 (15566013) , 2403-2415
LLMBA: Efficient Behavior Analytics via Large Pretrained Models in Zero Trust Networks
Guided by the principle of “Never Trust, Always Verify”, Zero Trust Architecture (ZTA) mandates continuous monitoring and analysis of users and entities, highlighting the critical role of behavior analytics. However, the growing volume of audit data and its complex contextual information render many existing behavior analytics methods insufficient. Moreover, most approaches rely on high-quality labeled data for supervised training, limiting their effectiveness against previously unseen malicious behaviors. To address these challenges, we propose the Large Language Model for Behavior Analytics (LLMBA) framework. LLMBA leverages a Large Language Model (LLM) to analyze behavioral patterns of internal users and entities, capitalizing on the LLM’s strong ability to model sequential data. We introduce a multi-level behavior encoding scheme to capture both contextual and temporal information from behavior records, producing rich input representations for the LLM-enhanced model. The LLM is fine-tuned using self-supervised learning, enabling the detection of unknown malicious behaviors. To reduce the computational and storage overhead inherent in LLMs, we apply knowledge distillation to compress the model while maintaining high detection performance. Extensive experiments on the CERT Insider Threat dataset demonstrate that LLMBA outperforms state-of-the-art baselines in detection accuracy. Furthermore, the compressed student model achieves superior performance compared with existing methods under comparable runtime constraints, making LLMBA highly suitable for real-world deployment.