Quantum networked systems introduce a new class of threats that classical intrusion detection systems (IDS) were not designed to handle, as adversaries can manipulate quantum circuits at the gate, pulse, and measurement levels while also targeting the classical control plane by intercepting encrypted IDS communications for future decryption. This paper addresses this dual threat model by proposing a two-layer defense architecture that jointly secures both the quantum data plane and the IDS infrastructure. The first layer uses machine learning models trained on the Quantum Attack Dataset (QAD) to detect circuit-level attacks under both attack-balanced and heavily skewed class distributions. The second layer protects the IDS control plane with post-quantum cryptographic (PQC) primitives, specifically ML-KEM for key encapsulation and ML-DSA for authentication, mitigating exposure to future decryption attacks. Experimental results across multiple deployment conditions show that tree-based models achieve strong detection performance but degrade under class-prior shift, with pulse-level attacks remaining the most challenging to detect, highlighting limitations of current feature representations. To address this, a confidence-aware policy maps detections to monitoring, verification, and isolation actions, although the verification tier is rarely exercised under the evaluated distributions. The results show that detection and cryptographic protection must be designed together to achieve robust intrusion detection in quantum networked environments.
Quantum networked systems introduce a new class of threats that classical intrusion detection systems (IDS) were not designed to handle, as adversaries can manipulate quantum circuits at the gate, pulse, and measurement levels while also targeting the classical control plane by intercepting encrypted IDS communications for future decryption. This paper addresses this dual threat model by proposing a two-layer defense architecture that jointly secures both the quantum data plane and the IDS infrastructure. The first layer uses machine learning models trained on the Quantum Attack Dataset (QAD) to detect circuit-level attacks under both attack-balanced and heavily skewed class distributions. The second layer protects the IDS control plane with post-quantum cryptographic (PQC) primitives, specifically ML-KEM for key encapsulation and ML-DSA for authentication, mitigating exposure to future decryption attacks. Experimental results across multiple deployment conditions show that tree-based models achieve strong detection performance but degrade under class-prior shift, with pulse-level attacks remaining the most challenging to detect, highlighting limitations of current feature representations. To address this, a confidence-aware policy maps detections to monitoring, verification, and isolation actions, although the verification tier is rarely exercised under the evaluated distributions. The results show that detection and cryptographic protection must be designed together to achieve robust intrusion detection in quantum networked environments.
Purpose This study aims to develop and validate a quantum-enhanced behavioral security framework that integrates a lightweight quantum obfuscation layer into classical intrusion detection systems (IDSs). The primary objective is to harden these models against post-quantum threats, specifically model extraction and feature inversion attacks, while maintaining high detection accuracy across both traditional machine learning and deep learning architectures. Design/methodology/approach A quantum feature obfuscation layer was designed using 6-qubit parameterized circuits with angle encoding and variational ansatz to nonlinearly transform behavioral features. This layer was integrated with classical classifiers (random forest, support vector machine, logistic regression) and a deep learning model multi-layer perceptron (MLP). The framework was evaluated on UNSW-NB15 and NSL-KDD datasets using Qiskit Aer simulators in Google Colab, measuring accuracy, precision, recall, F1-score, model extraction error (MEE) and feature inversion error (FIE). Findings Quantum-assisted models maintained detection performance comparable to classical baselines, with accuracy degradation of 0.7% across all architectures. Security resilience significantly improved, with a 3–4× increase in both MEE and FIE, indicating substantially enhanced resistance to model theft and feature reconstruction. The framework demonstrated improved noise tolerance under Gaussian perturbations. The MLP achieved 92.5% accuracy (classical) versus 91.8% (quantum-assisted), with 3.3× and 3.5× improvements in MEE and FIE, respectively. Research limitations/implications The evaluation was conducted entirely on quantum simulators, not capturing real hardware noise, decoherence and fidelity limitations. The threat model assumes secrecy of quantum circuit parameters, representing a form of model-level security through obscurity. Scalability is constrained by exponential simulation costs, limiting current experiments to 6 qubits. Only feedforward neural networks (MLP) were tested; advanced architectures like CNNs, LSTMs and transformers require future validation. The framework’s resilience against adversaries with partial knowledge of the ansatz structure remains unexplored. Practical implications The framework provides a pragmatic, classifier-agnostic defense layer deployable on freely accessible cloud platforms (Google Colab) without specialized quantum hardware. With only 15–25 ms inference overhead and 40–60% training overhead, it offers viable post-quantum hardening for security-critical applications. Resource-constrained environments such as edge computing nodes and IoT deployments can benefit from this approach. The demonstration that meaningful security gains (3–4× improvement) are achievable using only 6 qubits lowers the barrier for organizations to adopt quantum-assisted security measures today. Social implications As quantum computing threatens classical cryptography, protecting behavioral analytics becomes crucial for critical infrastructure, financial systems and healthcare networks. This research contributes to building resilient cyber defense mechanisms that protect sensitive data and privacy even when encryption is compromised. By democratizing access to quantum-enhanced security through cloud-based simulation platforms, the framework helps bridge the gap between institutions with varying resources, promoting more equitable cybersecurity preparedness for the post-quantum era across both developed and developing nations. Originality/value This work shifts the focus of quantum-enhanced security from pure cryptographic replacement to model-hardening, harnessing quantum state complexity as a defensive mechanism rather than pursuing quantum advantage for classification speed. It introduces a novel quantum obfuscation layer specifically designed for IDS, validated across both traditional ML and deep learning architectures. The consistent security improvements (3–4×) across classifier types confirm the approach is classifier-agnostic. The Colab-based implementation ensures reproducibility and accessibility, providing a practical foundation for future quantum-aware cyber defense research and deployment.
The advent of cryptographically relevant quantum computers (CRQCs) poses a fundamental threat to public-key cryptography securing modern digital communications. The most immediate danger is not direct decryption but the harvest-now-decrypt-later (HNDL) attack, wherein adversaries passively collect and archive encrypted traffic today for retrospective decryption once a CRQC becomes operational. This threat is structurally difficult to counter because the collection phase generates no host- or account-level signals detectable by conventional intrusion-detection systems.Simultaneously, migration to NIST-standardized post-quantum cryptography (PQC) specifically ML-KEM (FIPS 203), ML-DSA (FIPS 204), and SLH-DSA (FIPS 205) introduces new vulnerabilities, including chosen-ciphertext side-channel attacks, fault-injection attacks, and protocol downgrade attacks arising during hybrid deployment periods.Existing security tooling addresses cryptographic inventory, encrypted-traffic anomaly detection, and PQC readiness as isolated problems, leaving a critical operational gap. To address this, we propose QCAD (Quantum Cryptographic Attack Detection), a unified multi-plane detection framework integrating four complementary detection planes: (A) flow-metadata behavioral analytics targeting HNDL collection indicators; (B) host- and protocol-level telemetry for PQC implementation flaws and downgrade attacks; (C) continuous crypto-agility monitoring with per-asset Mosca-inequality risk scoring; and (D) quantum-vulnerable deception canaries. These planes feed a unified correlation layer that maps observations to MITRE ATT&CK techniques and produces risk-ranked alerts.Evaluation against synthetic HNDL and PQC-attack benchmarks demonstrates that cross-plane fusion achieves a recall of 0.97 at a 5% false-positive rate (F1 = 0.90, ROC-AUC = 0.99), substantially outperforming any individual plane, confirming that no single observable captures all attack families.A key boundary condition remains: HNDL detection is inherently a probabilistic behavioral-inference problem, and interception occurring on transit infrastructure beyond the defender's perimeter may remain undetectable regardless of sensor deployment.
Intrusion detection is an essential aspect of cybersecurity in fast-changing networked systems, but traditional machine learning-based IDS cannot scale or use high-dimensional traffic data easily. In this paper, a hybrid quantumclassical intrusion detection system (HQ-IDS) has been introduced, and it uses variational quantum circuits to enhance the learning and classification. Classical preprocessing techniques such as standardization and principal component analysis (PCA) are used to down sample and reduce features, which are then angle-encoded into quantum states and processed through parameterized quantum circuits with entangling layers. Binary classification is performed using quantum measurements combined with a classical readout layer, and the model is trained end-to-end via gradient-based optimization. Experimental results on synthetic intrusion data show that the proposed HQ-IDS achieves competitive accuracy while remaining scalable to noisy intermediate-scale quantum (NISQ) devices, emphasizing the possibility of hybrid quantum machine learning for future network security applications.
Quantum computing introduces a new class of cyber threats that challenge existing detection paradigms. While significant research has focused on post-quantum cryptography (PQC) and the long-term risk of cryptographic breakage, far less attention has been given to the detection of quantum-enabled attacks during the pre-cryptographically-relevant quantum (pre-CRQC) era. Current intrusion detection systems, network monitors, and cryptographic telemetry tools lack the conceptual models and operational indicators needed to identify adversaries who leverage quantum acceleration, quantum-optimized reconnaissance, or hybrid-mode downgrade strategies. This paper proposes QEADT-1, a systems-level taxonomy for detecting or monitoring quantum-enabled attack patterns against classical and PQC-transitioning infrastructure during the pre-CRQC era, addressing a gap in the literature, which has focused predominantly on prevention through PQC migration and on QKD or quantum-hardware concerns rather than operational detection. We classify six major attack classes, identify their underlying mechanisms, and map each to proposed observable indicators and required telemetry sources. A SOC-oriented detection matrix is introduced to operationalize the taxonomy as a synthesized detection model, followed by a systems architecture for quantum-attack monitoring that is designed to integrate with ongoing PQC migration. This work provides a foundational framework for organizations seeking to detect quantum-enabled threats before cryptographically relevant quantum computers emerge.
Threat detection models in cybersecurity must keep up with shifting traffic, strict feature budgets, and noisy hardware, yet even strong classical systems still miss rare or borderline attacks when the data distribution drifts. Small, near-term quantum processors are now available, but existing work rarely shows whether quantum components can improve end-to-end detection under these unstable, resource constrained conditions rather than just adding complexity. We address this gap with a hybrid architecture that uses a compact multilayer perceptron to compress security data and then routes a few features to 2-4 qubit quantum heads implemented as quantum support vector machines and variational circuits. Under matched preprocessing and training budgets, we benchmark these hybrids against tuned classical baselines on two security tasks, network intrusion detection on NSL-KDD and spam filtering on Ling-Spam datasets, and then deploy the best 4-qubit quantum SVM to an IBM Quantum device with noise-aware execution (readout mitigation and dynamical decoupling). Across both datasets, shallow quantum heads consistently match, and on difficult near-boundary cases modestly reduce, missed attacks and false alarms relative to classical models using the same features. Hardware results track simulator behavior closely enough that the remaining gap is dominated by device noise rather than model design. Taken together, the study shows that even on small, noisy chips, carefully engineered quantum components can already function as competitive, budget-aware elements in practical threat detection pipelines.
A vital component in ensuring network security is an Intrusion Detection System (IDS), and classical Machine Learning (ML) models often struggle to keep pace with sophisticated cyber threats. Quantum Computing, a probabilistic computational class utilising the principles of quantum mechanics, has shown its potential to address some of these challenges through quantum parallelism and entanglement. In this work, we explore the use of Quantum Machine Learning (QML), specifically Quantum Convolutional Neural Network (QCNN) and Quantum Support Vector Machine (QSVM), to determine the benefits of using quantum kernels over classical kernels. We investigate multiple ansatz designs, and entanglement patterns, that generate sufficient representations of vectors in the Hilbert space to distinguish normal from malicious activity. The results have shown improvements in both models compared to their classical counterparts, with QSVM being the best performing when circuit block entanglement is used.