Li Zeng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Li Zeng has a total of 25 co-authors. Here are the top co-authors by number of shared publications:
- Hangtao Zhang
Huazhong University of Science and Technology, Wuhan, China
2 shared publications - Yanchun Li
Xiangtan University
2 shared publications - Chao Chen
RMIT University
1 shared publication - Dongsu Shen
School of Computer Science, Xiangtan University, Xiangtan, 411105, China; Hunan International Scientific and Technological Cooperation Base of Intelligent network, Xiangtan, 411105, China; Key Laboratory of Hunan Province for Internet of Things and Information Security, Xiangtan, 411105, China
1 shared publication - Guosheng Han
School of Mathematics and Computational Science, Xiangtan University, Xiangtan, China
1 shared publication - Jialin Li
School of Mathematics and Computational Science, Xiangtan University, Xiangtan, China
1 shared publication
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 14022-14031
GhostPrompt: Cross-Image Adversarial Prompt for Vision-Language Models
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 14022-14031
Open AccessGhostPrompt: Cross-Image Adversarial Prompt for Vision-Language Models
- 2 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 23 (1) , 250-262
Fine-Grained Poisoning Framework Against Federated Learning
Federated learning (FL) is one of the most widely used distributed machine learning frameworks. However, FL is susceptible to poisoning attacks that can degrade the quality of the global model. Recent studies on fine-grained poisoning attacks highlight a strategic shift where attackers no longer prioritize maximal disruption of the global model, but instead control the degree of model poisoning to maintain stealth and avoid detection. However, research on fine-grained poisoning is still in the infant stage. Numerous fundamental questions have yet to be addressed, including its underlying mechanisms and optimization strategies. To this end, we introduces FGP, the first comprehensive framework for Fine-Grained Poisoning on FL, which allows adversaries to precisely manipulate the global model by strategically inducing accurate and stealthy sub-optimal solution. Fundamentally, FGP innovatively formalizes fine-grained attacks as an optimization problem to minimize the distance between the current global model and the adversary's target (a sub-optimal solution). It then employs a real-time search strategy to dynamically refine the malicious model updates in each round. To ensure optimal attack performance, we further introduce a novel topology-based approach as the error feedback. Additionally, we present a formal convergence analysis of our attacks. Armed with FGP, we conduct a comprehensive evaluation of FL's robustness against fine-grained poisoning across diverse settings. Results demonstrate that FGP significantly outperforms the prior work, achieving an average $6.5\times$ higher attack accuracy.
- 2 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 23 (1) , 250-262
Fine-Grained Poisoning Framework Against Federated Learning
Federated learning (FL) is one of the most widely used distributed machine learning frameworks. However, FL is susceptible to poisoning attacks that can degrade the quality of the global model. Recent studies on fine-grained poisoning attacks highlight a strategic shift where attackers no longer prioritize maximal disruption of the global model, but instead control the degree of model poisoning to maintain stealth and avoid detection. However, research on fine-grained poisoning is still in the infant stage. Numerous fundamental questions have yet to be addressed, including its underlying mechanisms and optimization strategies. To this end, we introduces FGP, the first comprehensive framework for Fine-Grained Poisoning on FL, which allows adversaries to precisely manipulate the global model by strategically inducing accurate and stealthy sub-optimal solution. Fundamentally, FGP innovatively formalizes fine-grained attacks as an optimization problem to minimize the distance between the current global model and the adversary's target (a sub-optimal solution). It then employs a real-time search strategy to dynamically refine the malicious model updates in each round. To ensure optimal attack performance, we further introduce a novel topology-based approach as the error feedback. Additionally, we present a formal convergence analysis of our attacks. Armed with FGP, we conduct a comprehensive evaluation of FL's robustness against fine-grained poisoning across diverse settings. Results demonstrate that FGP significantly outperforms the prior work, achieving an average $6.5\times$ higher attack accuracy.
- 24 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Services Computing
- Vol. 18 (5) , 3262-3275
Task Offloading and Resource Scheduling in Mobile Edge-Cloud Computing Based on Edge Competition and Task Prediction
In the emerging cloud-edge-end computing networks, edge servers possess more constrained resources and face greater task offloading pressure than centralized cloud servers due to the surge in mobile applications and data. Concurrently, the presence of multiple edge service providers introduces additional challenges, including competition among servers, disordered resource pricing, and a lack of coordination in edge and cloud resource allocation. To address these issues, we propose a novel approach aimed at optimizing task deployment, resource pricing, and system coordination. First, we develop a competitiveness model to facilitate efficient edge-side task allocation while addressing the challenges of resource pricing under competitive conditions. Second, we design a transformer-based task prediction model to enhance the accuracy of resource demand forecasting, thereby enabling more effective edge-cloud resource allocation. To achieve these objectives, the system's interaction is structured into two distinct stages. This division simplifies the problem-solving process and ensures that the long-term goal of maximizing benefits for all stakeholders—edge service providers, cloud providers, and end-users—is achieved. The proposed solution not only improves task offloading efficiency and resource utilization, but also promotes fair competition and pricing transparency across the system.
- 24 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Services Computing
- Vol. 18 (5) , 3262-3275
Task Offloading and Resource Scheduling in Mobile Edge-Cloud Computing Based on Edge Competition and Task Prediction
In the emerging cloud-edge-end computing networks, edge servers possess more constrained resources and face greater task offloading pressure than centralized cloud servers due to the surge in mobile applications and data. Concurrently, the presence of multiple edge service providers introduces additional challenges, including competition among servers, disordered resource pricing, and a lack of coordination in edge and cloud resource allocation. To address these issues, we propose a novel approach aimed at optimizing task deployment, resource pricing, and system coordination. First, we develop a competitiveness model to facilitate efficient edge-side task allocation while addressing the challenges of resource pricing under competitive conditions. Second, we design a transformer-based task prediction model to enhance the accuracy of resource demand forecasting, thereby enabling more effective edge-cloud resource allocation. To achieve these objectives, the system's interaction is structured into two distinct stages. This division simplifies the problem-solving process and ensures that the long-term goal of maximizing benefits for all stakeholders—edge service providers, cloud providers, and end-users—is achieved. The proposed solution not only improves task offloading efficiency and resource utilization, but also promotes fair competition and pricing transparency across the system.
- 18 July 2025
- Published by Elsevier in Neural Networks
- Vol. 192, 107877
Adversarial purification with one-step guided diffusion model
- 18 July 2025
- Published by Elsevier in Neural Networks
- Vol. 192, 107877
Adversarial purification with one-step guided diffusion model
- Open Access
- 22 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-10
GMF-MGCN-LDA: Prediction of IncRNA-disease association based on novel generalized matrix factorization and graph neural networks
A growing number of long non-coding RNAs have been found to be associated with various diseases, highlighting the significance of predicting lncRNA-disease associations (LDAs) for disease diagnosis and treatment. In light of the time-consuming and laborious nature of traditional biological experiments, numerous algorithms have been developed to predict unknown LDAs. To predict potential LDAs, we propose a model called GMF-MGCN-LDA, that leverages the advantages of graph convolutional networks and generalized matrix factorization. Firstly, in order to extract high-order nonlinear feature of lncRNA and disease, the hypergraph regularization term is introduced into the generalized matrix factorization framework to learn latent feature of lncRNA and disease. Next, in order to enrich the feature of lncRNA and disease, the graph convolutional networks is employed to learn nonlinear features from diverse similarity graph views, followed by the attention mechanism that adaptively assigns weights to these features. Finally, the convolutional neural networks is utilized to integrate features from each view. The features learned by generalized matrix factorization and convolutional neural networks are concatenated to obtain the final lncRNA and disease feature. Based on known LDAs, the two final features are further concatenated and merged into one feature as input to the CatBoost classifier to output the predicted association probability of lncRNA-disease pairs. The experimental results not only demonstrate the effectiveness of the components in GMF-MGCN-LDA but also establish its performance is better than or comparable to other models. Furthermore, the case study underscores the model's reliable prediction capability.
- 22 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-10
Open AccessGMF-MGCN-LDA: Prediction of IncRNA-disease association based on novel generalized matrix factorization and graph neural networks
A growing number of long non-coding RNAs have been found to be associated with various diseases, highlighting the significance of predicting lncRNA-disease associations (LDAs) for disease diagnosis and treatment. In light of the time-consuming and laborious nature of traditional biological experiments, numerous algorithms have been developed to predict unknown LDAs. To predict potential LDAs, we propose a model called GMF-MGCN-LDA, that leverages the advantages of graph convolutional networks and generalized matrix factorization. Firstly, in order to extract high-order nonlinear feature of lncRNA and disease, the hypergraph regularization term is introduced into the generalized matrix factorization framework to learn latent feature of lncRNA and disease. Next, in order to enrich the feature of lncRNA and disease, the graph convolutional networks is employed to learn nonlinear features from diverse similarity graph views, followed by the attention mechanism that adaptively assigns weights to these features. Finally, the convolutional neural networks is utilized to integrate features from each view. The features learned by generalized matrix factorization and convolutional neural networks are concatenated to obtain the final lncRNA and disease feature. Based on known LDAs, the two final features are further concatenated and merged into one feature as input to the CatBoost classifier to output the predicted association probability of lncRNA-disease pairs. The experimental results not only demonstrate the effectiveness of the components in GMF-MGCN-LDA but also establish its performance is better than or comparable to other models. Furthermore, the case study underscores the model's reliable prediction capability.