Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

Applied Energy

Aligning quantum kernels for detecting false data injection attacks in power systems

Bin Huang, Jianhui Wang, Xiaoge Huang

Published Jan 8, 2026• 6 citations
Publisher

Abstract

False Data Injection Attacks (FDIAs) are an elusive and impactful security threat in power systems. Existing solutions, such as those based on machine learning methods with classical kernels, suffer from weak noise resistance, extensive parameter tuning, and unstable performance in high-dimensional spaces. This work provides a novel solution to this problem from the perspective of quantum computing, specifically through the use of quantum kernel embedding (QKE). Classical kernels require costly hyperparameter searches that scale poorly with the number of measurements and often fail to separate sparse, overlapping attack patterns inherent in FDIA detection; by contrast, quantum kernel embedding leverages the exponentially large Hilbert space via shallow, fixed-depth circuits to generate highly expressive feature maps with far fewer tunable parameters, enabling more efficient and robust discrimination of subtle false-data-injection patterns. A kernel-alignment framework with a weighted loss function is proposed to tune the quantum circuit parameters so that the quantum kernel captures the imbalanced label structure and more effectively distinguishes attacked from normal data. Additionally, the Nyström approximation is extended to the quantum Hilbert space, utilizing only a subset of training data instead of the entire dataset, which enhances the scalability and computational efficiency of the method. Case studies on test systems and high-performance quantum simulators demonstrate the effectiveness of the method and evaluate its robustness under varying quantum incoherent noise.

DOI
10.1016/j.apenergy.2025.127332

Related papers

Quantum-Efficient Kernel Target Alignment

Rodrigo Coelho, Georg Kruse, Andreas Roßkopf

In recent years, quantum computers have emerged as promising candidates for implementing kernels. Quantum Embedding Kernels embed data points into quantum states and calculate their inner product in a high-dimensional Hilbert Space by computing the overlap between the resulting quantum states. Variational Quantum Circuits (VQCs) are typically used for this end, with Kernel Target Alignment (KTA) as cost function. The optimized kernels can then be deployed in Support Vector Machines (SVMs) for classification tasks. However, both classical and quantum SVMs scale poorly with increasing dataset sizes. This issue is exacerbated in quantum kernel methods, as each inner product requires a quantum circuit execution. In this paper, we investigate KTA-trained quantum embedding kernels and employ a low-rank matrix approximation, the Nyström method, to reduce the quantum circuit executions needed to construct the Kernel Matrix. We empirically evaluate the performance of our approach across various datasets, focusing on the accuracy of the resulting SVM and the reduction in quantum circuit executions. Additionally, we examine and compare the robustness of our model under different noise types, particularly coherent and depolarizing noise.

arXiv:2502.08225•International Conference on Agents and Artificial Intelligence (ICAART)•Jan 1, 2025•9 citations

Quantum Key Distribution Applications for Enhancing Smart Grid Security Against False Data Injection Attacks

Brundavanam Seshasai, Ebha Koley, Subhojit Ghosh

Cyber threats are more likely to happen to modern smart grids because they rely more on real-time sensing, communication, and automation. One of the most dangerous types of attacks is a False Data Injection Attack (FDIA). This is because it can get around normal bad-data detection systems and change how state estimation and electricity market operations work. Most of the time, existing mitigation techniques use statistical or machine learning methods to find problems. They don’t fix problems with communication. Moreover, conventional cryptographic techniques utilized for safeguarding measurement data are fundamentally limited by computational security assumptions and are increasingly threatened by the advent of quantum computing. This paper analyzes the application of Quantum Key Distribution (QKD) as an enabling technology to enhance the cyber-resilience of smart grids against FDIAs. A hybrid security framework is proposed, employing QKD-based secure key generation for message authentication on critical measurement links, thereby preventing malicious data manipulation before the state estimation process. MATPOWER and MATLAB-based simulations are used to test the suggested method on the IEEE 14-bus system. The simulation results demonstrate that authentication using QKD effectively prevents FDIA-induced alterations in voltage states and Locational Marginal Prices (LMPs), while incurring minimal computational overhead. The results show that using QKD on only the most important communication links is a cheap and scalable way to make the smart grid safer.

IEEE International Conference on Sustainable Energy and Future Electric Transportation (SEFET)•Jul 1, 2026•0 citations

Towards Practical Quantum Kernels for Network Intrusion Detection

Mary L. Cotrupi, Brian Callahan

With cyber-attacks becoming increasingly sophisticated, modern network intrusion detection systems (NIDSs) are relying on machine learning (ML) methods for their flexibility in detecting subtle anomalous patterns in huge amounts of network data. However, classical ML methods such as support vector machines (SVMs) often rely on the conversion of low-dimensional data into a high-dimensional space, creating complex linear systems that are time-consuming to evaluate on large data inputs such as network flow logs. We propose addressing this limitation by employing a hybrid quantum-classical ML model to leverage quantum computing's (QC's) superiority in high-dimensional areas. We constructed a quantum kernel with an SVM model and evaluated it on four different network attacks from a modern intrusion detection dataset. Results reveal an average hardware accuracy rate of 85% with noticeably small deviations between runs, suggesting that quantum kernels may be a noise-resistant solution. We evaluated these results alongside classical and noiseless quantum simulator benchmarks.

Proceedings of the AAAI Symposium Series•Nov 23, 2025•5 citations

QKIDS: Quantum Kernel Learning for Network Intrusion Detection Systems

Santanu Roy, R. S. Ramya, Abhinava P +3 more

Network intrusion detection is very critical for cybersecurity in modern infrastructure in order to maintain secure digital systems. This research paper presents QKIDS (Quantum Kernel Learning for Intrusion Detection System), a comparative study on the quantum kernel methods against classical machine learning approaches on the NSL-KDD dataset. We implement a variational quantum classifier using parameterized quantum circuits with angle encoding and entanglement-based feature maps, demonstrating performance competitiveness can be achieved by quantum kernel methods with classical baselines including Support Vector Machines. Through systematic qubit ablation studies, we establish that quantum kernels can match classical models on real-world cybersecurity data while revealing current practical limitations including computational overhead. The proposed quantum approach demonstrates competitive de-taction performance comparable to the best classical methods, with identical attack detection rates on critical security metrics. Our experimental setup with stratified train-test splitting and mutual information-based feature selection establishes practical baselines for quantum machine learning in network security applications. What makes our framework new is that it uses quantum kernel methods that use high-dimensional Hilbert space transformations to find complicated, non-linear patterns in data on network traffic that classical kernels might not be able to represent well. Our variational quantum circuits encode data into a quantum state space through parameterized rotation gates and controlled entanglement operations. The proposed approach is different from classical SVMs, which work in original feature space. This could help separate normal and malicious network behaviors better.

•Jul 1, 2026•0 citations

Investigating Quantum Kernel Feature Maps and Kernel Alignment for Multivariate Spatio-Temporal Pattern Classification

Vishwambhar Pathak, Nisheeth Joshi, Prabhat Mahanti

AI-driven solutions and simulations often involve spatio-temporal multivariate signals. However, such real-life signals are mostly nonlinear, noisy, and high-dimensional. Fetching sparse and subtle spatio-temporal patterns from them for descriptive and predictive applications poses challenge to traditional machine learning models. With recent practicable advancements in quantum computing technologies, quantum kernel methods have been exploited to map classical data into high-dimensional Hilbert spaces which transform complex relationships into separable patterns. In this work we investigate the effectiveness of quantum kernel methods for the case of EEG signal classification and compares their performance with classical radial basis function (RBF) kernels. Quantum encoding was performed using angle-encoding and necessary circuit elements including swap-test. The investigation was conducted using two sets of synthetic EEG datasets: (1) simple sinusoidal signals and (2) nonlinear signals with phase and amplitude variability. Principal Component Analysis was used to reduce dimensionality before kernel construction. For contrasting the Classical vs Quantum workflows, feature maps produced by Classical and Quantum kernel matrices were computed and fed to Support Vector Machine (SVM) classifier. The workflow then evaluated Kernel-target alignment for estimation of the correspondence between kernel similarity structures and class labels. For simple sinusoidal signals, training with the classical kernels achieved accuracies up to 0.967 for 8 principal component features as compared to 0.86 accuracy with the quantum kernels. However, for nonlinear EEG signals, while both workflows gave lower accuracies, still quantum kernels yielded slightly better accuracy 0.46 as compared to 0.40 by the classical kernels. The kernel-target alignment was mostly better with classical kernel, with slightly better for quantum kernel in one case. The slight improvement with quantum kernel in non-linear signal case suggests that quantum kernels may provide advantages in modelling complex nonlinear relationships. Limited alignment and limited accuracy improvement with the quantum kernel indicate need for further improvement in quantum-feature-map design to achieve quantum advantage.

•May 1, 2026•0 citations

Parameterised Quantum SVM with Data-Driven Entanglement for Zero-Day Exploit Detection

Steven Jabulani Nhlapo, Elodie Ngoie Mutombo, Mike Nkongolo

Zero-day attacks pose a persistent threat to computing infrastructure by exploiting previously unknown software vulnerabilities that evade traditional signature-based network intrusion detection systems (NIDSs). To address this limitation, machine learning (ML) techniques offer a promising approach for enhancing anomaly detection in network traffic. This study evaluates several ML models on a labeled network traffic dataset, with a focus on zero-day attack detection. Ensemble learning methods, particularly eXtreme gradient boosting (XGBoost), achieved perfect classification, identifying all 6231 zero-day instances without false positives and maintaining efficient training and prediction times. While classical support vector machines (SVMs) performed modestly at 64% accuracy, their performance improved to 98% with the use of the borderline synthetic minority oversampling technique (SMOTE) and SMOTE + edited nearest neighbours (SMOTEENN). To explore quantum-enhanced alternatives, a quantum SVM (QSVM) is implemented using three-qubit and four-qubit quantum circuits simulated on the aer_simulator_statevector. The QSVM achieved high accuracy (99.89%) and strong F1-scores (98.95%), indicating that nonlinear quantum feature maps (QFMs) can increase sensitivity to zero-day exploit patterns. Unlike prior work that applies standard quantum kernels, this study introduces a parameterised quantum feature encoding scheme, where each classical feature is mapped using a nonlinear function tuned by a set of learnable parameters. Additionally, a sparse entanglement topology is derived from mutual information between features, ensuring a compact and data-adaptive quantum circuit that aligns with the resource constraints of noisy intermediate-scale quantum (NISQ) devices. Our contribution lies in formalising a quantum circuit design that enables scalable, expressive, and generalisable quantum architectures tailored for zero-day attack detection. This extends beyond conventional usage of QSVMs by offering a principled approach to quantum circuit construction for cybersecurity. While these findings are obtained via noiseless simulation, they provide a theoretical proof of concept for the viability of quantum ML (QML) in network security. Future work should target real quantum hardware execution and adaptive sampling techniques to assess robustness under decoherence, gate errors, and dynamic threat environments.

Computers•Aug 15, 2025•6 citations
False Data Injection Attacks (FDIAs) are an elusive and impactful security threat in power systems. Existing solutions, such as those based on machine learning methods with classical kernels, suffer from weak noise resistance, extensive parameter tuning, and unstable performance in high-dimensional spaces. This work provides a novel solution to this problem from the perspective of quantum computing, specifically through the use of quantum kernel embedding (QKE). Classical kernels require costly hyperparameter searches that scale poorly with the number of measurements and often fail to separate sparse, overlapping attack patterns inherent in FDIA detection; by contrast, quantum kernel embedding leverages the exponentially large Hilbert space via shallow, fixed-depth circuits to generate highly expressive feature maps with far fewer tunable parameters, enabling more efficient and robust discrimination of subtle false-data-injection patterns. A kernel-alignment framework with a weighted loss function is proposed to tune the quantum circuit parameters so that the quantum kernel captures the imbalanced label structure and more effectively distinguishes attacked from normal data. Additionally, the Nyström approximation is extended to the quantum Hilbert space, utilizing only a subset of training data instead of the entire dataset, which enhances the scalability and computational efficiency of the method. Case studies on test systems and high-performance quantum simulators demonstrate the effectiveness of the method and evaluate its robustness under varying quantum incoherent noise.
In recent years, quantum computers have emerged as promising candidates for implementing kernels. Quantum Embedding Kernels embed data points into quantum states and calculate their inner product in a high-dimensional Hilbert Space by computing the overlap between the resulting quantum states. Variational Quantum Circuits (VQCs) are typically used for this end, with Kernel Target Alignment (KTA) as cost function. The optimized kernels can then be deployed in Support Vector Machines (SVMs) for classification tasks. However, both classical and quantum SVMs scale poorly with increasing dataset sizes. This issue is exacerbated in quantum kernel methods, as each inner product requires a quantum circuit execution. In this paper, we investigate KTA-trained quantum embedding kernels and employ a low-rank matrix approximation, the Nyström method, to reduce the quantum circuit executions needed to construct the Kernel Matrix. We empirically evaluate the performance of our approach across various datasets, focusing on the accuracy of the resulting SVM and the reduction in quantum circuit executions. Additionally, we examine and compare the robustness of our model under different noise types, particularly coherent and depolarizing noise.
Cyber threats are more likely to happen to modern smart grids because they rely more on real-time sensing, communication, and automation. One of the most dangerous types of attacks is a False Data Injection Attack (FDIA). This is because it can get around normal bad-data detection systems and change how state estimation and electricity market operations work. Most of the time, existing mitigation techniques use statistical or machine learning methods to find problems. They don’t fix problems with communication. Moreover, conventional cryptographic techniques utilized for safeguarding measurement data are fundamentally limited by computational security assumptions and are increasingly threatened by the advent of quantum computing. This paper analyzes the application of Quantum Key Distribution (QKD) as an enabling technology to enhance the cyber-resilience of smart grids against FDIAs. A hybrid security framework is proposed, employing QKD-based secure key generation for message authentication on critical measurement links, thereby preventing malicious data manipulation before the state estimation process. MATPOWER and MATLAB-based simulations are used to test the suggested method on the IEEE 14-bus system. The simulation results demonstrate that authentication using QKD effectively prevents FDIA-induced alterations in voltage states and Locational Marginal Prices (LMPs), while incurring minimal computational overhead. The results show that using QKD on only the most important communication links is a cheap and scalable way to make the smart grid safer.
With cyber-attacks becoming increasingly sophisticated, modern network intrusion detection systems (NIDSs) are relying on machine learning (ML) methods for their flexibility in detecting subtle anomalous patterns in huge amounts of network data. However, classical ML methods such as support vector machines (SVMs) often rely on the conversion of low-dimensional data into a high-dimensional space, creating complex linear systems that are time-consuming to evaluate on large data inputs such as network flow logs. We propose addressing this limitation by employing a hybrid quantum-classical ML model to leverage quantum computing's (QC's) superiority in high-dimensional areas. We constructed a quantum kernel with an SVM model and evaluated it on four different network attacks from a modern intrusion detection dataset. Results reveal an average hardware accuracy rate of 85% with noticeably small deviations between runs, suggesting that quantum kernels may be a noise-resistant solution. We evaluated these results alongside classical and noiseless quantum simulator benchmarks.
Network intrusion detection is very critical for cybersecurity in modern infrastructure in order to maintain secure digital systems. This research paper presents QKIDS (Quantum Kernel Learning for Intrusion Detection System), a comparative study on the quantum kernel methods against classical machine learning approaches on the NSL-KDD dataset. We implement a variational quantum classifier using parameterized quantum circuits with angle encoding and entanglement-based feature maps, demonstrating performance competitiveness can be achieved by quantum kernel methods with classical baselines including Support Vector Machines. Through systematic qubit ablation studies, we establish that quantum kernels can match classical models on real-world cybersecurity data while revealing current practical limitations including computational overhead. The proposed quantum approach demonstrates competitive de-taction performance comparable to the best classical methods, with identical attack detection rates on critical security metrics. Our experimental setup with stratified train-test splitting and mutual information-based feature selection establishes practical baselines for quantum machine learning in network security applications. What makes our framework new is that it uses quantum kernel methods that use high-dimensional Hilbert space transformations to find complicated, non-linear patterns in data on network traffic that classical kernels might not be able to represent well. Our variational quantum circuits encode data into a quantum state space through parameterized rotation gates and controlled entanglement operations. The proposed approach is different from classical SVMs, which work in original feature space. This could help separate normal and malicious network behaviors better.
AI-driven solutions and simulations often involve spatio-temporal multivariate signals. However, such real-life signals are mostly nonlinear, noisy, and high-dimensional. Fetching sparse and subtle spatio-temporal patterns from them for descriptive and predictive applications poses challenge to traditional machine learning models. With recent practicable advancements in quantum computing technologies, quantum kernel methods have been exploited to map classical data into high-dimensional Hilbert spaces which transform complex relationships into separable patterns. In this work we investigate the effectiveness of quantum kernel methods for the case of EEG signal classification and compares their performance with classical radial basis function (RBF) kernels. Quantum encoding was performed using angle-encoding and necessary circuit elements including swap-test. The investigation was conducted using two sets of synthetic EEG datasets: (1) simple sinusoidal signals and (2) nonlinear signals with phase and amplitude variability. Principal Component Analysis was used to reduce dimensionality before kernel construction. For contrasting the Classical vs Quantum workflows, feature maps produced by Classical and Quantum kernel matrices were computed and fed to Support Vector Machine (SVM) classifier. The workflow then evaluated Kernel-target alignment for estimation of the correspondence between kernel similarity structures and class labels. For simple sinusoidal signals, training with the classical kernels achieved accuracies up to 0.967 for 8 principal component features as compared to 0.86 accuracy with the quantum kernels. However, for nonlinear EEG signals, while both workflows gave lower accuracies, still quantum kernels yielded slightly better accuracy 0.46 as compared to 0.40 by the classical kernels. The kernel-target alignment was mostly better with classical kernel, with slightly better for quantum kernel in one case. The slight improvement with quantum kernel in non-linear signal case suggests that quantum kernels may provide advantages in modelling complex nonlinear relationships. Limited alignment and limited accuracy improvement with the quantum kernel indicate need for further improvement in quantum-feature-map design to achieve quantum advantage.
Zero-day attacks pose a persistent threat to computing infrastructure by exploiting previously unknown software vulnerabilities that evade traditional signature-based network intrusion detection systems (NIDSs). To address this limitation, machine learning (ML) techniques offer a promising approach for enhancing anomaly detection in network traffic. This study evaluates several ML models on a labeled network traffic dataset, with a focus on zero-day attack detection. Ensemble learning methods, particularly eXtreme gradient boosting (XGBoost), achieved perfect classification, identifying all 6231 zero-day instances without false positives and maintaining efficient training and prediction times. While classical support vector machines (SVMs) performed modestly at 64% accuracy, their performance improved to 98% with the use of the borderline synthetic minority oversampling technique (SMOTE) and SMOTE + edited nearest neighbours (SMOTEENN). To explore quantum-enhanced alternatives, a quantum SVM (QSVM) is implemented using three-qubit and four-qubit quantum circuits simulated on the aer_simulator_statevector. The QSVM achieved high accuracy (99.89%) and strong F1-scores (98.95%), indicating that nonlinear quantum feature maps (QFMs) can increase sensitivity to zero-day exploit patterns. Unlike prior work that applies standard quantum kernels, this study introduces a parameterised quantum feature encoding scheme, where each classical feature is mapped using a nonlinear function tuned by a set of learnable parameters. Additionally, a sparse entanglement topology is derived from mutual information between features, ensuring a compact and data-adaptive quantum circuit that aligns with the resource constraints of noisy intermediate-scale quantum (NISQ) devices. Our contribution lies in formalising a quantum circuit design that enables scalable, expressive, and generalisable quantum architectures tailored for zero-day attack detection. This extends beyond conventional usage of QSVMs by offering a principled approach to quantum circuit construction for cybersecurity. While these findings are obtained via noiseless simulation, they provide a theoretical proof of concept for the viability of quantum ML (QML) in network security. Future work should target real quantum hardware execution and adaptive sampling techniques to assess robustness under decoherence, gate errors, and dynamic threat environments.