Quantum-inspired cyber threat detection via entanglement-based feature fusion
|
Full Text |
Pdf
|
|
Author |
Sareddy Hemalatha and Kuruva Chandra Shekhar
|
|
e-ISSN |
1819-6608 |
|
On Pages
|
476-487
|
|
Volume No. |
21
|
|
Issue No. |
7
|
|
Issue Date |
June 10, 2026
|
|
DOI |
https://doi.org/10.59018/042656
|
|
Keywords |
quantum-inspired machine learning, intrusion detection system (IDS), variational quantum circuit (VQC), cybersecurity, CICIDS2017, hybrid quantum-classical model, entanglement-based feature fusion, network traffic analysis, cyber threat detection.
|
Abstract
The increasing sophistication of cyberattacks requires intrusion detection systems that can model nonlinear and complex network-traffic patterns beyond the capabilities of conventional machine-learning designs. This paper proposes a Hybrid Quantum-Inspired Cyber Threat Detection (Hybrid QCTD) framework that combines amplitude-based quantum feature encoding, a variational quantum circuit (VQC), and a lightweight classical classifier. Evaluated on the CICIDS2017 dataset, the proposed model achieved an accuracy of 0.9550, an F1-score of 0.8779, a precision of 0.8108, a recall of 0.9571, and an ROC-AUC of 0.9931. Random Forest and XGBoost achieved higher accuracy and F1-scores, whereas the Hybrid QCTD model maintained strong detection performance and stable training convergence. A prediction-level example further illustrates that the hybrid model may produce different decisions from the classical baselines for complex network flows; however, such examples should be interpreted together with the aggregate test metrics. These findings demonstrate the feasibility of entanglement-based feature fusion for intrusion detection and motivate further evaluation on noisy quantum hardware and resource-constrained, real-time cybersecurity systems.
Back