Kun Zhu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Kun Zhu has a total of 35 co-authors. Here are the top co-authors by number of shared publications:
- Nana Zhang
School of Computer Science and Technology, Donghua University, Shanghai, China
13 shared publications - Dandan Zhu
East China Normal University
12 shared publications - Guangtao Zhai
Shanghai Jiao Tong University
8 shared publications - Changjun Jiang
Lanzhou University
5 shared publications - Kaiwei Zhang
Shanghai Artificial Intelligence Laboratory, Shanghai, China
5 shared publications - Xiaokang Yang
Shanghai Jiao Tong University
5 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 15792-15801
AdaptiveGaze: Emulating Human-like Adaptive Vision for Flexible Scanpath Prediction
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 15792-15801
Open AccessAdaptiveGaze: Emulating Human-like Adaptive Vision for Flexible Scanpath Prediction
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 6476-6487
Combating Web-Based E-Commerce Fraud Syndicates: Fairness-Aware Hypergraph Contrastive Fraud Detection with Multi-Dimensional Reinforcement Rewards
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 6476-6487
Open AccessCombating Web-Based E-Commerce Fraud Syndicates: Fairness-Aware Hypergraph Contrastive Fraud Detection with Multi-Dimensional Reinforcement Rewards
- Open Access
- 18 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 22 (8) , 1-22
MEScan360: A Memory-Enhanced Scanpath Prediction Model for Omnidirectional Images
Scanpath prediction for omnidirectional images aims to capture the dynamic viewing behavior of humans while freely exploring visual scenes. The complex gaze behavior along with the inevitable projection distortion, however, present significant challenges in the scanpath prediction task for omnidirectional scenes. Most existing methods neither capture the long-term dependencies across visual states nor fully incorporate historical memory information, leading to suboptimal performance. To this end, we propose MEScan360, a novel memory-enhanced scanpath prediction model for omnidirectional images, which effectively preserves historical information in the working-memory space and builds long-range temporal dependencies across visual states. Specifically, inspired by the crucial role of historical context and scene semantics in guiding visual attention, we propose a long-term memory storage unit and a memory interaction module. These two components establish a more explicit link between past visual information and current visual inputs, thereby significantly enhancing the performance of scanpath prediction. Furthermore, a robust feature extraction module is designed to extract rotation-invariant semantic features. Extensive experiments on benchmark datasets demonstrate that MEScan360 achieves competitive performance in both accuracy and efficiency. Notably, our MEScan360 exhibits strong generalization ability across two computer vision tasks: saliency prediction and virtual scene design assistance, providing valuable insights for other vision-related applications.
- 18 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 22 (8) , 1-22
Open AccessMEScan360: A Memory-Enhanced Scanpath Prediction Model for Omnidirectional Images
Scanpath prediction for omnidirectional images aims to capture the dynamic viewing behavior of humans while freely exploring visual scenes. The complex gaze behavior along with the inevitable projection distortion, however, present significant challenges in the scanpath prediction task for omnidirectional scenes. Most existing methods neither capture the long-term dependencies across visual states nor fully incorporate historical memory information, leading to suboptimal performance. To this end, we propose MEScan360, a novel memory-enhanced scanpath prediction model for omnidirectional images, which effectively preserves historical information in the working-memory space and builds long-range temporal dependencies across visual states. Specifically, inspired by the crucial role of historical context and scene semantics in guiding visual attention, we propose a long-term memory storage unit and a memory interaction module. These two components establish a more explicit link between past visual information and current visual inputs, thereby significantly enhancing the performance of scanpath prediction. Furthermore, a robust feature extraction module is designed to extract rotation-invariant semantic features. Extensive experiments on benchmark datasets demonstrate that MEScan360 achieves competitive performance in both accuracy and efficiency. Notably, our MEScan360 exhibits strong generalization ability across two computer vision tasks: saliency prediction and virtual scene design assistance, providing valuable insights for other vision-related applications.
- 26 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. PP (4) , 1-16
HFDFM: A Heterogeneous Credit Card Fraud Detection Model Based on Federated Learning With Membership Privacy
Credit card fraud brings serious losses to both cardholders and card issuers. To reduce losses caused by fraudulent behaviors, banking institutions establish credit card fraud detection (CCFD) models to identify potential fraudulent behaviors. To develop more effective fraud detection models, banking institutions need to collaborate on model training. Federated learning (FL) enables collaboration on fraud detection model training without exchanging data between banking institutions. Nevertheless, the distribution of transaction data in the real-world is heterogeneous among banking institutions, which may lead to convergence issues in global fraud detection models. Furthermore, the behavior and weights of the model may implicitly contain the cardholders’ personal information, which makes existing federated models prone to transaction data leakage. In this article, we propose a heterogeneous credit card fraud detection model based on federated learning with membership privacy, called HFDFM. To protect the sensitive information of cardholders, we design a novel mechanism that ensures training-data confidentiality by minimizing the accuracy of the best black-box membership inference attack (MIA) against the model. Unlike previous FL frameworks that either optimize for client drift or for membership privacy, HFDFM simultaneously mitigates drift and provides certified membership privacy through a single min–max game. Additionally, we use the control variable to rectify the client drift in its local update in the scenario of heterogeneous data. Extensive experimental results on three mainstream real-world transaction datasets demonstrate that the proposed HFDFM has advantages in utility compared to 11 SOTA baselines, and the proposed HFDFM can mitigate the risks of MIAs (near random guess).
- 26 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. PP (4) , 1-16
HFDFM: A Heterogeneous Credit Card Fraud Detection Model Based on Federated Learning With Membership Privacy
Credit card fraud brings serious losses to both cardholders and card issuers. To reduce losses caused by fraudulent behaviors, banking institutions establish credit card fraud detection (CCFD) models to identify potential fraudulent behaviors. To develop more effective fraud detection models, banking institutions need to collaborate on model training. Federated learning (FL) enables collaboration on fraud detection model training without exchanging data between banking institutions. Nevertheless, the distribution of transaction data in the real-world is heterogeneous among banking institutions, which may lead to convergence issues in global fraud detection models. Furthermore, the behavior and weights of the model may implicitly contain the cardholders’ personal information, which makes existing federated models prone to transaction data leakage. In this article, we propose a heterogeneous credit card fraud detection model based on federated learning with membership privacy, called HFDFM. To protect the sensitive information of cardholders, we design a novel mechanism that ensures training-data confidentiality by minimizing the accuracy of the best black-box membership inference attack (MIA) against the model. Unlike previous FL frameworks that either optimize for client drift or for membership privacy, HFDFM simultaneously mitigates drift and provides certified membership privacy through a single min–max game. Additionally, we use the control variable to rectify the client drift in its local update in the scenario of heterogeneous data. Extensive experimental results on three mainstream real-world transaction datasets demonstrate that the proposed HFDFM has advantages in utility compared to 11 SOTA baselines, and the proposed HFDFM can mitigate the risks of MIAs (near random guess).
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 742-753
STG-DGR: Fraud Detection on Streaming Transaction Graphs with Diffusion-based Generative Replay
Fraud detection on streaming transaction graphs (STGs) faces challenges on the catastrophic forgetting of previously learned fraud patterns when adapting to evolving patterns. Although some Graph Continual Learning (GCL) approaches mitigate this issue by storing and revisiting historical samples, practical storage constraints prevent them from fully preserving previous patterns. In this work, we propose STG-DGR, a streaming GNN model with diffusion-based generative replay that generates synthetic samples to retain previously learned patterns without storing real samples. The generation of replay samples for STGs faces two key challenges: (1) Heterogeneity challenge of generating STG samples with discrete adjacency table, user features, transaction features, and transaction timestamps. (2) Dependency challenge of capturing bottom-up dependencies across layers in STG samples. To address these challenges, STG-DGR integrates two novel components: (1) a Computational Subgraph Processor (CSP) that transforms heterogeneous STG samples into well-organized hierarchical subgraphs, and (2) a Diffusion-based Subgraph Generator (DSG) that captures the bottom-up dependencies using a novel Transformer-based Hierarchical Denoising Network (THDN), and generates synthetic replay samples that preserve these dependencies. Extensive experiments on four streaming fraud detection datasets demonstrate STG-DGR's superiority in reducing forgetting and improving accuracy over nineteen state-of-the-art baselines.
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 742-753
Open AccessSTG-DGR: Fraud Detection on Streaming Transaction Graphs with Diffusion-based Generative Replay
Fraud detection on streaming transaction graphs (STGs) faces challenges on the catastrophic forgetting of previously learned fraud patterns when adapting to evolving patterns. Although some Graph Continual Learning (GCL) approaches mitigate this issue by storing and revisiting historical samples, practical storage constraints prevent them from fully preserving previous patterns. In this work, we propose STG-DGR, a streaming GNN model with diffusion-based generative replay that generates synthetic samples to retain previously learned patterns without storing real samples. The generation of replay samples for STGs faces two key challenges: (1) Heterogeneity challenge of generating STG samples with discrete adjacency table, user features, transaction features, and transaction timestamps. (2) Dependency challenge of capturing bottom-up dependencies across layers in STG samples. To address these challenges, STG-DGR integrates two novel components: (1) a Computational Subgraph Processor (CSP) that transforms heterogeneous STG samples into well-organized hierarchical subgraphs, and (2) a Diffusion-based Subgraph Generator (DSG) that captures the bottom-up dependencies using a novel Transformer-based Hierarchical Denoising Network (THDN), and generates synthetic replay samples that preserve these dependencies. Extensive experiments on four streaming fraud detection datasets demonstrate STG-DGR's superiority in reducing forgetting and improving accuracy over nineteen state-of-the-art baselines.
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 9079-9089
Dynamic Min-Max Multi-Dimensional Reinforcement Backdoor Attacks and Orchestrated Closed-Loop Defense in Fairness-Aware Web Federated Finance
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 9079-9089
Open AccessDynamic Min-Max Multi-Dimensional Reinforcement Backdoor Attacks and Orchestrated Closed-Loop Defense in Fairness-Aware Web Federated Finance
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 8995-9005
Bridging Cognitive Neuroscience and Graph Intelligence: Hippocampus-Inspired Multi-View Hypergraph Learning for Web Finance Fraud
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 8995-9005
Open AccessBridging Cognitive Neuroscience and Graph Intelligence: Hippocampus-Inspired Multi-View Hypergraph Learning for Web Finance Fraud
- 31 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 56 (2) , 1408-1422
CamFD: Semi-Supervised Camouflage-Aware Fraud Detection Based on Dynamic Graphs
Fraud detection on dynamic graph (FDDG) is in high demand across many real-world applications, such as financial transaction networks and social networks. A graph neural network (GNN) is an advanced methodology for learning graph data and has been widely adopted in fraud detection tasks. Despite the promising results, research on GNN-based fraud detection faces two critical challenges. First, due to the lack of temporal consistency mining in existing methods, camouflaged fraudsters can easily evade detection by imitating the behavioral patterns of benign users. Second, with constantly evolving fraud patterns and a scarcity of fraud samples, existing methods can easily overfit to current fraud patterns and struggle to identify the evolved and new ones. To address these challenges, we propose CamFD, a semi-supervised camouflage-aware fraud detection model on dynamic graphs. To detect the temporal inconsistency in the behavioral patterns of camouflaged fraudsters, CamFD is equipped with a novel consistency-sensitive contrastive learning (CSCL) module. CSCL discriminatively learns the consistency and inconsistency in users’ behavioral patterns by constructing temporal and structural contrastive pairs. Since modeling the evolving fraud patterns with sparse fraud samples is almost impractical, CamFD concentrates on modeling the relatively stable patterns of extensive benign users with a multivariate Gaussian distribution modeling (MGDM) module. We conduct extensive experiments on a private credit card fraud dataset as well as three public datasets. The experimental results demonstrate that CamFD outperforms ten state-of-the-art (SOTA) baselines across all datasets, with the area under ROC curve (AUC) improvements ranging from 1.03% to 5.32%.
- 31 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 56 (2) , 1408-1422
CamFD: Semi-Supervised Camouflage-Aware Fraud Detection Based on Dynamic Graphs
Fraud detection on dynamic graph (FDDG) is in high demand across many real-world applications, such as financial transaction networks and social networks. A graph neural network (GNN) is an advanced methodology for learning graph data and has been widely adopted in fraud detection tasks. Despite the promising results, research on GNN-based fraud detection faces two critical challenges. First, due to the lack of temporal consistency mining in existing methods, camouflaged fraudsters can easily evade detection by imitating the behavioral patterns of benign users. Second, with constantly evolving fraud patterns and a scarcity of fraud samples, existing methods can easily overfit to current fraud patterns and struggle to identify the evolved and new ones. To address these challenges, we propose CamFD, a semi-supervised camouflage-aware fraud detection model on dynamic graphs. To detect the temporal inconsistency in the behavioral patterns of camouflaged fraudsters, CamFD is equipped with a novel consistency-sensitive contrastive learning (CSCL) module. CSCL discriminatively learns the consistency and inconsistency in users’ behavioral patterns by constructing temporal and structural contrastive pairs. Since modeling the evolving fraud patterns with sparse fraud samples is almost impractical, CamFD concentrates on modeling the relatively stable patterns of extensive benign users with a multivariate Gaussian distribution modeling (MGDM) module. We conduct extensive experiments on a private credit card fraud dataset as well as three public datasets. The experimental results demonstrate that CamFD outperforms ten state-of-the-art (SOTA) baselines across all datasets, with the area under ROC curve (AUC) improvements ranging from 1.03% to 5.32%.
- 12 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 48 (4) , 4257-4274
Developing Evolving Adaptability in Biological Intelligence: A Novel Biologically-Inspired Continual Learning Model for Video Saliency Prediction
In the era of deep learning, video saliency prediction task still remains major challenge due to the issue of catastrophic forgetting during feature learning. Most prior works commonly employ generative replay strategies to generate pseudo-samples from previous tasks, enabling them to recall the data distribution. However, scaling up generative replay to accommodate class-incremental and task-incremental settings poses challenges, as generated data with low quality can severely deteriorate performance. Additionally, existing advances mainly focus on preserving memory stability to alleviate catastrophic forgetting, but they remain difficult to flexibly adapt to incremental changes in dynamic scenes. To achieve a better balance between memory stability and learning plasticity, we propose a novel biologically-inspired continual learning (BICL) model tailored to effectively predict human attention in dynamic scenes while mitigate catastrophic forgetting. In particular, inspired by the function of the hippocampus in the human neural system, we elaborately design a visual saliency memory bank module to explicitly store and retrieve representative features from previous tasks. Furthermore, drawing inspiration from the Drosophila$\gamma$ MB system, we propose an active forgetting strategy equipped with multiple parallel adaptive learner modules, which can appropriately attenuate old memories in parameter distribution to enhance learning plasticity to adapt to new tasks, and accordingly to ensure compatibility among multiple learners. Notably, without compromising the performance of old tasks, our proposed model can achieve a better trade-off between memory stability and learning plasticity. Through extensive experiments on several benchmark datasets, our model not only enhances performance in task-incremental settings, but also potentially provides deep insights into neurological adaptive mechanisms.- 12 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 48 (4) , 4257-4274
Developing Evolving Adaptability in Biological Intelligence: A Novel Biologically-Inspired Continual Learning Model for Video Saliency Prediction
In the era of deep learning, video saliency prediction task still remains major challenge due to the issue of catastrophic forgetting during feature learning. Most prior works commonly employ generative replay strategies to generate pseudo-samples from previous tasks, enabling them to recall the data distribution. However, scaling up generative replay to accommodate class-incremental and task-incremental settings poses challenges, as generated data with low quality can severely deteriorate performance. Additionally, existing advances mainly focus on preserving memory stability to alleviate catastrophic forgetting, but they remain difficult to flexibly adapt to incremental changes in dynamic scenes. To achieve a better balance between memory stability and learning plasticity, we propose a novel biologically-inspired continual learning (BICL) model tailored to effectively predict human attention in dynamic scenes while mitigate catastrophic forgetting. In particular, inspired by the function of the hippocampus in the human neural system, we elaborately design a visual saliency memory bank module to explicitly store and retrieve representative features from previous tasks. Furthermore, drawing inspiration from the Drosophila$\gamma$ MB system, we propose an active forgetting strategy equipped with multiple parallel adaptive learner modules, which can appropriately attenuate old memories in parameter distribution to enhance learning plasticity to adapt to new tasks, and accordingly to ensure compatibility among multiple learners. Notably, without compromising the performance of old tasks, our proposed model can achieve a better trade-off between memory stability and learning plasticity. Through extensive experiments on several benchmark datasets, our model not only enhances performance in task-incremental settings, but also potentially provides deep insights into neurological adaptive mechanisms.