Qiankun Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Qiankun Zhang has a total of 40 co-authors. Here are the top co-authors by number of shared publications:
- Zhiyi Huang
The University of Hong Kong
4 shared publications - Bin Yuan
School of Cyber Science and Engineering, Huazhong University of Science and Technology, Wuhan, China
3 shared publications - Xianjun Deng
School of Cyber Science and Engineering, Huazhong University of Science and Technology, Wuhan, China
3 shared publications - Aocheng Shen
Huazhong University of Science and Technology
2 shared publications - Deqing Zou
Huazhong University of Science and Technology, Wuhan, China
2 shared publications - Di Yuan
School of Cyber Science and Engineering, Huazhong University of Science and Technology, Wuhan, China
2 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 8479-8488
TextAE: A Benchmark and Flow-Matching Baseline for Text-Controlled Layered Animation
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 8479-8488
Open AccessTextAE: A Benchmark and Flow-Matching Baseline for Text-Controlled Layered Animation
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 7166-7175
TempJail: Temporal Jailbreak Attacks against Image-to-Video Generation Models
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 7166-7175
Open AccessTempJail: Temporal Jailbreak Attacks against Image-to-Video Generation Models
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 777-785
Unsupervised Subgraph Anomaly Detection Based on Pattern Collaboration
Subgraph Anomaly Detection (SAD) is crucial for identifying groups that deviate from the regular pattern within graphs, which benefits different domains such as financial fraud and network security. However, current studies rely on traditional node detection methods and fixed sampling strategies of subgraph structures, which makes it difficult to learn the pattern collaboration behavior of subgraphs. To address this limitation, this paper proposes a novel unsupervised framework named PC-SAD. The PC-SAD framework first employs an improved Graph AutoEncoder to identify core anomaly nodes by capturing multi-scale neighborhood information. Starting from these core anomaly nodes, we sample candidate subgraphs with path, tree, and cyclic structures, and enhance them according to the characteristics of the subgraph structures. Subsequently, candidate subgraphs are fed into the proposed Pattern Collaboration-based Graph Contrastive Learning method to generate collaborative pattern embeddings, thereby distinguishing anomaly subgraphs. The experimental results show that PC-SAD outperforms the state-of-the-art baseline methods on four benchmark datasets, which proves that PC-SAD is an effective solution to detect anomaly subgraphs.
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 777-785
Open AccessUnsupervised Subgraph Anomaly Detection Based on Pattern Collaboration
Subgraph Anomaly Detection (SAD) is crucial for identifying groups that deviate from the regular pattern within graphs, which benefits different domains such as financial fraud and network security. However, current studies rely on traditional node detection methods and fixed sampling strategies of subgraph structures, which makes it difficult to learn the pattern collaboration behavior of subgraphs. To address this limitation, this paper proposes a novel unsupervised framework named PC-SAD. The PC-SAD framework first employs an improved Graph AutoEncoder to identify core anomaly nodes by capturing multi-scale neighborhood information. Starting from these core anomaly nodes, we sample candidate subgraphs with path, tree, and cyclic structures, and enhance them according to the characteristics of the subgraph structures. Subsequently, candidate subgraphs are fed into the proposed Pattern Collaboration-based Graph Contrastive Learning method to generate collaborative pattern embeddings, thereby distinguishing anomaly subgraphs. The experimental results show that PC-SAD outperforms the state-of-the-art baseline methods on four benchmark datasets, which proves that PC-SAD is an effective solution to detect anomaly subgraphs.
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 5322-5333
Automated Model Selection for Multivariate Time Series Forecasting
Accurate multivariate time series forecasting (MTSF) is critical for intelligent web services in Web of Things. When confronted with unseen multivariate time series (MTS), the industry typically invests significant time and resources in training multiple models to identify the optimal model for deployment. This paper proposes a novel, efficient, and scalable MTSF model selection method that directly selects suitable MTSF methods based on data characteristics without extensive model training. Model selection is a core component of AutoML, which has made significant progress in recent years. However, existing methods incur high operational costs and cannot be directly applied to MTSF tasks. Moreover, there is a lack of a comprehensive and cohesive public time series library for MTSF model selection. To address these challenges, we compile the first large heterogeneous labeled MTSF model selection dataset, called the ModelPile, which covers 41 mainstream datasets across 11 domains. We then propose AutoMTSF, a large model-enabled model selection method that transforms the MTSF model selection problem into a time series classification problem and utilizes the ModelPile to unlock large-scale multi-dataset training. AutoMTSF first uses the pre-trained large model to encode raw MTS. Given the coarse-grained limitations of large model encoding, Recursive Temporal Pattern Feature (RTPF) is proposed to capture both fine-grained and global temporal feature evolution, thereby effectively mapping data characteristics to the MTSF method space. Experiments comparing AutoMTSF with 2 baselines, 17 MTSF methods, and 4 large time series models show that AutoMTSF outperforms state-of-the-art methods while maintaining comparable execution time. This work represents a critical step in validating the accuracy and efficiency of large model-enabled classification for MTSF.
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 5322-5333
Open AccessAutomated Model Selection for Multivariate Time Series Forecasting
Accurate multivariate time series forecasting (MTSF) is critical for intelligent web services in Web of Things. When confronted with unseen multivariate time series (MTS), the industry typically invests significant time and resources in training multiple models to identify the optimal model for deployment. This paper proposes a novel, efficient, and scalable MTSF model selection method that directly selects suitable MTSF methods based on data characteristics without extensive model training. Model selection is a core component of AutoML, which has made significant progress in recent years. However, existing methods incur high operational costs and cannot be directly applied to MTSF tasks. Moreover, there is a lack of a comprehensive and cohesive public time series library for MTSF model selection. To address these challenges, we compile the first large heterogeneous labeled MTSF model selection dataset, called the ModelPile, which covers 41 mainstream datasets across 11 domains. We then propose AutoMTSF, a large model-enabled model selection method that transforms the MTSF model selection problem into a time series classification problem and utilizes the ModelPile to unlock large-scale multi-dataset training. AutoMTSF first uses the pre-trained large model to encode raw MTS. Given the coarse-grained limitations of large model encoding, Recursive Temporal Pattern Feature (RTPF) is proposed to capture both fine-grained and global temporal feature evolution, thereby effectively mapping data characteristics to the MTSF method space. Experiments comparing AutoMTSF with 2 baselines, 17 MTSF methods, and 4 large time series models show that AutoMTSF outperforms state-of-the-art methods while maintaining comparable execution time. This work represents a critical step in validating the accuracy and efficiency of large model-enabled classification for MTSF.
- Open Access
- 13 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 13, 6685-6705
DroneTest-Copilot: LLM-Powered Automated Detection and Repair of Flaws in Drone Autotest Suites
In the landscape of the low-altitude economy, ensuring the reliability of drone flight control software is of utmost importance. ArduPilot's autotest is an automated testing framework integrated into its continuous integration (CI) pipeline, which is crucial for validating autopilot functionality. However, numerous logic defects exist in its test cases, where the intended workflow doesn't match the code implementation. This paper presents DroneTest-Copilot (DTC), a novel framework that leverages large language models (LLMs) to automate the validation and correction of autotest test cases. The system consists of four components: Preprocessor, Test Case Processor, Flaw Corrector, and Vulnerability Tester. It extracts relevant context from pull requests, uses LLMs to detect flaws such as parameter mismatch, API misuse, execution order errors, and logic errors, and then generates corrected test code. We also conduct a detailed survey and comparative analysis of existing related works to highlight the novelty and positioning of DTC. Evaluations on a dataset of patched and new test case pull requests show that DTC can correctly identify 87.23% of introduced defects in the verification set and discover previously unknown vulnerabilities in the new test case set. DTC substantially reduces manual effort in maintaining autotest suites and uncovers subtle flaws in test case code, enhancing the reliability of drone flight control software.
- 13 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 13, 6685-6705
Open AccessDroneTest-Copilot: LLM-Powered Automated Detection and Repair of Flaws in Drone Autotest Suites
In the landscape of the low-altitude economy, ensuring the reliability of drone flight control software is of utmost importance. ArduPilot's autotest is an automated testing framework integrated into its continuous integration (CI) pipeline, which is crucial for validating autopilot functionality. However, numerous logic defects exist in its test cases, where the intended workflow doesn't match the code implementation. This paper presents DroneTest-Copilot (DTC), a novel framework that leverages large language models (LLMs) to automate the validation and correction of autotest test cases. The system consists of four components: Preprocessor, Test Case Processor, Flaw Corrector, and Vulnerability Tester. It extracts relevant context from pull requests, uses LLMs to detect flaws such as parameter mismatch, API misuse, execution order errors, and logic errors, and then generates corrected test code. We also conduct a detailed survey and comparative analysis of existing related works to highlight the novelty and positioning of DTC. Evaluations on a dataset of patched and new test case pull requests show that DTC can correctly identify 87.23% of introduced defects in the verification set and discover previously unknown vulnerabilities in the new test case set. DTC substantially reduces manual effort in maintaining autotest suites and uncovers subtle flaws in test case code, enhancing the reliability of drone flight control software.
- Open Access
- 2 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 1256-1270
Membership Inference Attacks against Vision Transformers: Mosaic MixUp Training to the Defense
Vision transformers (ViTs) have demonstrated great success in various fundamental CV tasks, mainly benefiting from their self-attention-based transformer architectures, and the paradigm of pre-training followed by fine-tuning. However, such advantages may lead to significant data privacy risks, such as membership inference attacks (MIAs), which remain unclear. This paper presents the first comprehensive study on MIAs and corresponding defenses against ViTs. Our first contribution is a rollout-attention-based MIA method (RAMIA), based on an experimental observation that the attention, more precisely the rollout attention, behaves disproportionately for members and non-members. We evaluate RAMIA on the standard ViT architecture proposed by Google (ICLR 2021), achieving high accuracy, precision, and recall performance. Further, inspired by another experimental observation on a strong connection between positional embeddings (PEs) and attentions, we propose a novel framework for training ViTs, named Mosaic MixUp Training (MMUT), as a defense against RAMIA. Intuitively, MMUT mixes up private images and public ones at a patch level, and mosaics the corresponding PEs with a global learnable mosaic embedding. Our empirical results show MMUT achieves a much better accuracy-privacy trade-off than some common defense mechanisms. Extensive experiments are conducted to rigorously evaluate both RAMIA and MMUT.
- 2 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 1256-1270
Open AccessMembership Inference Attacks against Vision Transformers: Mosaic MixUp Training to the Defense
Vision transformers (ViTs) have demonstrated great success in various fundamental CV tasks, mainly benefiting from their self-attention-based transformer architectures, and the paradigm of pre-training followed by fine-tuning. However, such advantages may lead to significant data privacy risks, such as membership inference attacks (MIAs), which remain unclear. This paper presents the first comprehensive study on MIAs and corresponding defenses against ViTs. Our first contribution is a rollout-attention-based MIA method (RAMIA), based on an experimental observation that the attention, more precisely the rollout attention, behaves disproportionately for members and non-members. We evaluate RAMIA on the standard ViT architecture proposed by Google (ICLR 2021), achieving high accuracy, precision, and recall performance. Further, inspired by another experimental observation on a strong connection between positional embeddings (PEs) and attentions, we propose a novel framework for training ViTs, named Mosaic MixUp Training (MMUT), as a defense against RAMIA. Intuitively, MMUT mixes up private images and public ones at a patch level, and mosaics the corresponding PEs with a global learnable mosaic embedding. Our empirical results show MMUT achieves a much better accuracy-privacy trade-off than some common defense mechanisms. Extensive experiments are conducted to rigorously evaluate both RAMIA and MMUT.
- 20 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Applied Superconductivity
- Vol. 34 (8) , 1-5
IDM-Net: A Multi-Task Supported Encoder-Decoder Framework for Magnetic Field Inverse Design
We propose an end-to-end framework for inversely designing permanent magnets named IDM-Net. It utilizes a fundamental encoder-decoder architecture to handle multiple tasks. In more detail, the encoder is responsible for deeply extracting features of the magnetic field and categorizing different types of magnet shapes. The decoder focuses on inferring significant properties of each specific type of magnet shape, including size, position, and magnetization intensity. Such architecture breaks the critical limitation of designing only a single type of magnet in literature. Further, it allows for flexible choices of encoders' networks, such as convolutional neural networks (CNNs) or transformers, which are widely used in various computer vision tasks. Our experimental results demonstrate that the ResNet50-based and ViT-B/16-based IDM-Nets achieve accuracies of 93.8% and 91.4% in magnet shapes classification and errors of 0.31% and 0.33% in predicting magnetic properties, respectively.
- 20 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Applied Superconductivity
- Vol. 34 (8) , 1-5
IDM-Net: A Multi-Task Supported Encoder-Decoder Framework for Magnetic Field Inverse Design
We propose an end-to-end framework for inversely designing permanent magnets named IDM-Net. It utilizes a fundamental encoder-decoder architecture to handle multiple tasks. In more detail, the encoder is responsible for deeply extracting features of the magnetic field and categorizing different types of magnet shapes. The decoder focuses on inferring significant properties of each specific type of magnet shape, including size, position, and magnetization intensity. Such architecture breaks the critical limitation of designing only a single type of magnet in literature. Further, it allows for flexible choices of encoders' networks, such as convolutional neural networks (CNNs) or transformers, which are widely used in various computer vision tasks. Our experimental results demonstrate that the ResNet50-based and ViT-B/16-based IDM-Nets achieve accuracies of 93.8% and 91.4% in magnet shapes classification and errors of 0.31% and 0.33% in predicting magnetic properties, respectively.
- 4 September 2024
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Computing
- Vol. 53 (5) , 1217-1256
Online Primal Dual Meets Online Matching with Stochastic Rewards: Configuration LP to the Rescue
Mehta and Panigrahi ( FOCS 2012, IEEE, Piscataway, NJ, 2012, pp. 728–737) introduce the problem of online matching with stochastic rewards, where edges are associated with success probabilities and a match succeeds with the probability of the corresponding edge. It is one of the few online matching problems that have defied the randomized online primal dual framework by Devanur, Jain, and Kleinberg ( SODA 2013, SIAM, Philadelphia, 2013, pp. 101–107) thus far. This paper unlocks the power of randomized online primal dual in online matching with stochastic rewards by employing the configuration linear program rather than the standard matching linear program used in previous works. Our main result is a 0.572 competitive algorithm for the case of vanishing and unequal probabilities, improving the best previous bound of 0.534 by Mehta, Waggoner, and Zadimoghaddam ( SODA 2015, SIAM, Philadelphia, 2015, pp. 1388–1404) and, in fact, is even better than the best previous bound of 0.567 by Mehta and Panigrahi ( FOCS 2012, IEEE, Piscataway, NJ, 2012, pp. 728–737) for the more restricted case of vanishing and equal probabilities. For vanishing and equal probabilities, we get a better competitive ratio of 0.576. Our results further generalize to the vertex-weighted case due to the intrinsic robustness of the randomized online primal dual analysis.
- 4 September 2024
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Computing
- Vol. 53 (5) , 1217-1256
Online Primal Dual Meets Online Matching with Stochastic Rewards: Configuration LP to the Rescue
Mehta and Panigrahi ( FOCS 2012, IEEE, Piscataway, NJ, 2012, pp. 728–737) introduce the problem of online matching with stochastic rewards, where edges are associated with success probabilities and a match succeeds with the probability of the corresponding edge. It is one of the few online matching problems that have defied the randomized online primal dual framework by Devanur, Jain, and Kleinberg ( SODA 2013, SIAM, Philadelphia, 2013, pp. 101–107) thus far. This paper unlocks the power of randomized online primal dual in online matching with stochastic rewards by employing the configuration linear program rather than the standard matching linear program used in previous works. Our main result is a 0.572 competitive algorithm for the case of vanishing and unequal probabilities, improving the best previous bound of 0.534 by Mehta, Waggoner, and Zadimoghaddam ( SODA 2015, SIAM, Philadelphia, 2015, pp. 1388–1404) and, in fact, is even better than the best previous bound of 0.567 by Mehta and Panigrahi ( FOCS 2012, IEEE, Piscataway, NJ, 2012, pp. 728–737) for the more restricted case of vanishing and equal probabilities. For vanishing and equal probabilities, we get a better competitive ratio of 0.576. Our results further generalize to the vertex-weighted case due to the intrinsic robustness of the randomized online primal dual analysis.
- 17 June 2024
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Computing
- Vol. 53 (3) , 701-763
AdWords in a Panorama
Three decades ago, Karp, Vazirani, and Vazirani [ Proceedings of the 22 nd Annual ACM Symposium on Theory of Computing, 1990, pp. 352–358] defined the online matching problem and gave an optimal [Formula: see text]-competitive algorithm. Fifteen years later, Mehta et al. [ J. ACM, 54 (2007), pp. 22:1–22:19] introduced the first generalization called AdWords driven by online advertising and obtained the optimal [Formula: see text] competitive ratio in the special case of small bids. It has been open ever since whether there is an algorithm for general bids better than the 0.5-competitive greedy algorithm. This paper presents a 0.5016-competitive algorithm for AdWords, answering this open question on the positive end. The algorithm builds on several ingredients, including a combination of the online primal dual framework and the configuration linear program of matching problems recently explored by Huang and Zhang [ Proceedings of the 52 nd ACM Symposium on Theory of Computing, 2020], a novel formulation of AdWords which we call the panorama view, and a generalization of the online correlated selection by Fahrbach et al. [ Proceedings of the 61 st Annual IEEE Symposium on Foundations of Computer Science, 2020], which we call the panoramic online correlated selection.
- 17 June 2024
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Computing
- Vol. 53 (3) , 701-763
AdWords in a Panorama
Three decades ago, Karp, Vazirani, and Vazirani [ Proceedings of the 22 nd Annual ACM Symposium on Theory of Computing, 1990, pp. 352–358] defined the online matching problem and gave an optimal [Formula: see text]-competitive algorithm. Fifteen years later, Mehta et al. [ J. ACM, 54 (2007), pp. 22:1–22:19] introduced the first generalization called AdWords driven by online advertising and obtained the optimal [Formula: see text] competitive ratio in the special case of small bids. It has been open ever since whether there is an algorithm for general bids better than the 0.5-competitive greedy algorithm. This paper presents a 0.5016-competitive algorithm for AdWords, answering this open question on the positive end. The algorithm builds on several ingredients, including a combination of the online primal dual framework and the configuration linear program of matching problems recently explored by Huang and Zhang [ Proceedings of the 52 nd ACM Symposium on Theory of Computing, 2020], a novel formulation of AdWords which we call the panorama view, and a generalization of the online correlated selection by Fahrbach et al. [ Proceedings of the 61 st Annual IEEE Symposium on Foundations of Computer Science, 2020], which we call the panoramic online correlated selection.
- Open Access
- 10 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 21 (3) , 3636-3655
Toward Automated Attack Discovery in SDN Controllers Through Formal Verification
Software-defined Network (SDN), presented to be a novel architecture of network because of its separation of data plane and control plane, brings centralization and extensibility to network management as well as new attacks that exploit the flexibility of SDN. OpenFlow, which is the protocol that is applied by the majority of SDN, leads to the widely used definition of the communication between the controller and the switch resulting in similar implementations regardless of different vendors. In this paper, we focus on the mechanisms of packet processing and topology discovery and their fundamental weaknesses caused by general implementations or device limitations. Despite the common vulnerabilities, the universal standard mechanisms of basic function in SDN also enlighten us to present an automated attack discovery method based on the formal verification with a generic model of SDN system. We describe the abstraction of the SDN components, their key functions, and communications along with the malicious operations that could be executed by malicious hosts and malicious switches and translate them into a formal model of the SDN system. The formal verification carried on with the assertion representing the security properties derived from the common vulnerabilities of the SDN system reports the potential attack paths each of which shows an attack process. Our evaluation shows that our method can discover feasible attack paths efficiently and effectively, with 23 attacks being identified, among which 2 are new. We further demonstrate the practicality of the 2 new attacks.
- 10 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 21 (3) , 3636-3655
Open AccessToward Automated Attack Discovery in SDN Controllers Through Formal Verification
Software-defined Network (SDN), presented to be a novel architecture of network because of its separation of data plane and control plane, brings centralization and extensibility to network management as well as new attacks that exploit the flexibility of SDN. OpenFlow, which is the protocol that is applied by the majority of SDN, leads to the widely used definition of the communication between the controller and the switch resulting in similar implementations regardless of different vendors. In this paper, we focus on the mechanisms of packet processing and topology discovery and their fundamental weaknesses caused by general implementations or device limitations. Despite the common vulnerabilities, the universal standard mechanisms of basic function in SDN also enlighten us to present an automated attack discovery method based on the formal verification with a generic model of SDN system. We describe the abstraction of the SDN components, their key functions, and communications along with the malicious operations that could be executed by malicious hosts and malicious switches and translate them into a formal model of the SDN system. The formal verification carried on with the assertion representing the security properties derived from the common vulnerabilities of the SDN system reports the potential attack paths each of which shows an attack process. Our evaluation shows that our method can discover feasible attack paths efficiently and effectively, with 23 attacks being identified, among which 2 are new. We further demonstrate the practicality of the 2 new attacks.