Evaluating the potential of quantum machine learning in cybersecurity: A case-study on PCA-based intrusion detection systems

2025-02-01

Computers & Security

AI Summary

This work investigates the potential of quantum machine learning (QML) to enhance classical intrusion detection systems (IDS) in cybersecurity, specifically within a principal component analysis (PCA)-based detection pipeline. We propose the first integration of quantum variational PCA (QV-PCA) into a conventional IDS workflow, introducing a dual encoding scheme—combining amplitude and angle encodings—tailored to network traffic features, and establishing a unified quantum–classical comparative evaluation framework. Experiments on the NSL-KDD dataset using Qiskit simulation demonstrate that QV-PCA achieves a 4.7× acceleration in feature compression while preserving 92.3% detection accuracy; under simulated noise, F1-score variation remains within ±1.2%, indicating preliminary robustness. Our key contribution is the novel, architecture-aware fusion of QV-PCA with operational IDS design, delivering a reproducible methodology and empirical benchmark for deploying QML in resource-constrained security applications.

Technology Category

• Machine Learning: Quantum Machine Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Humans and AI: Human-in-the-loop Machine Learning

Application Category

• Security and Privacy: Large-scale security measurements
• Web Mining and Content Analysis: Machine learning and data science for the Web
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Problem

• Quantum Machine Learning in Cybersecurity
• PCA-based Intrusion Detection Systems
• Quantum Advantage in ML Applications

Innovation

• Quantum Machine Learning
• PCA-based Intrusion Detection
• Fault-tolerant QML Algorithms