Researchers publish hybrid quantum computing and deep learning framework for cybersecurity threat detection

Researchers from Saudi Arabia and Malaysia have published a hybrid cybersecurity framework combining deep learning and quantum computing. By integrating CNNs for spatial features, RNNs for temporal patterns, and Variational Quantum Circuits for classification, the model addresses traditional computational limits. Tested on the CICIDS-2017 dataset, the framework achieved 93% accuracy and low false-positive rates in a virtual environment, marking a step toward real-time threat detection.

Hybrid quantum-deep learning framework

• ▪Researchers published a hybrid deep learning and quantum computing framework designed to enhance cybersecurity threat detection and analysis on July 31, 2026.
• ▪The research was supported by a grant from the National Cybersecurity Authority in the Kingdom of Saudi Arabia under the Cybersecurity Research and Innovation Pioneers Initiative.
• ▪The hybrid framework was developed by researchers from the University of Bisha, Taibah University, INTI-IU, and Universiti Tenaga Nasional.

CNN spatial feature extraction

• ▪Traditional convolutional neural networks are effective for detecting network anomalies but face challenges with high-dimensionality networks and computational performance.
• ▪The hybrid framework employs convolutional neural networks to extract spatial features from network traffic data.

RNN temporal pattern detection

• ▪The hybrid framework utilizes recurrent neural networks to leverage temporal patterns for cyber threat detection.
• ▪Traditional recurrent neural networks struggle to adapt to new cyber attacks and handle high-dimensionality networks on their own.

Variational Quantum Circuits classification

• ▪The hybrid framework integrates Variational Quantum Circuits to perform quantum-enhanced classification of cybersecurity threats.
• ▪Quantum computing provides parallel computing and superior optimization advantages that complement traditional deep learning methods in the framework.

CICIDS-2017 dataset preprocessing

• ▪The hybrid framework is built using the CICIDS-2017 dataset to extract features during an extensive preprocessing phase.
• ▪Preprocessing of the CICIDS-2017 dataset involves normalization, feature selection, feature time structuring, and feature balancing using over-sampling and under-sampling techniques.

Cybersecurity threat detection performance

• ▪Evaluation of the hybrid model demonstrated high performance across multiple metrics, including recall, F1 score, and ROC-AUC area.
• ▪The hybrid model achieved a 93% accuracy rate and low false-positive rates across different categories of cyberattacks in an experimental virtual environment.

1 source