Quantum AI for Cybersecurity: A hybrid Quantum-Classical models for attack path analysis

TL;DR

This work tackles attack path analysis under data scarcity by proposing a hybrid quantum–classical pipeline that maps classical cybersecurity features into quantum embeddings via angle encoding and variational circuits, then classifies with a classical SVM. Classical models outperform in large-data settings, but quantum embeddings show superior attack recall and improved separability when data are limited, highlighting potential representational advantages of quantum feature spaces. The study provides a reproducible framework using CPU-based simulators and identifies practical constraints—such as limited qubits and circuit depth—that shape current performance and guide future quantum hardware-ready research. Overall, the paper demonstrates a plausible, though not yet superior, role for quantum embeddings in cybersecurity analytics and outlines concrete avenues for scaling and refining quantum approaches as hardware advances.

Abstract

Modern cyberattacks are increasingly complex, posing significant challenges to classical machine learning methods, particularly when labeled data is limited and feature interactions are highly non-linear. In this study we investigates the potential of hybrid quantum-classical learning to enhance feature representations for intrusion detection and explore possible quantum advantages in cybersecurity analytics. Using the UNSW-NB15 dataset, network traffic is transformed into structured feature vectors through classical preprocessing and normalization. Classical models, including Logistic Regression and Support Vector Machines with linear and RBF kernels, are evaluated on the full dataset to establish baseline performance under large-sample conditions. Simultaneously, a quantum-enhanced pipeline maps classical features into variational quantum circuits via angle encoding and entangling layers, executed on a CPU-based quantum simulator, with resulting quantum embeddings classified using a classical SVM. Experiments show that while classical models achieve higher overall accuracy with large datasets, quantum-enhanced representations demonstrate superior attack recall and improved class separability when data is scarce, suggesting that quantum feature spaces capture complex correlations inaccessible to shallow classical models. These results highlight the potential of quantum embeddings to improve generalization and representation quality in cybersecurity tasks and provide a reproducible framework for evaluating quantum advantages as quantum hardware and simulators continue to advance.

Figures (4)

• Figure 1: Comparison of classical and hybrid quantum--classical workflows for cybersecurity attack path analysis.
• Figure 2: Comparison of classical machine learning models performance on the UNSW-NB15 dataset in terms of Accuracy, Precision, Recall, and F1-score. Similar outcomes are observed with Logistic Regression, Linear SVM and RBF SVM and higher recall and F1 scores are obtained with SVM based models.
• Figure 3: Quantitative performance assessment of the Quantum ML-SVM model in terms of accuracy, precision, recall, and F1-score.
• Figure 4: Confusion matrix of the model prediction and the True label in terms of quantum SVM classification.