Pith. sign in

REVIEW 3 major objections 5 minor 36 references

RF Spectrogram Anomaly Detection with Quantum Kitchen Sinks: Architecture, Representation, and Hardware Validation

T0 review · 3 major / 5 minor · reviewed 2026-08-02 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Quantum Kitchen Sinks improve RF spectrogram anomaly detection over matched classical baselines across all evaluated representation–readout pairs, with the best configuration reaching 0.8778 test AUROC, and the improvement survives real qua

desk verdict Careful QKS benchmark for RF anomaly detection, but the 'matched' baselines don't control capacity, so the central claim is oversold. read the letter →

arxiv 2607.13897 v2 pith:45QCJAWZ submitted 2026-07-15 cs.LG

Abdallah Aaraba , Alexis Vieloszynski , Remon Polus , Ola Ahmad , Soumaya Cherkaoui This is my paper · ORCID
classification cs.LG
keywords quantummachinelearningkitchensinksanomalydetectionRFspectrogramsdatare-uploadingDCTrepresentationhybridquantum-classicalhardwarevalidation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that a hybrid quantum-classical pipeline—Quantum Kitchen Sinks with repeated data re-uploading and ring entanglement—improves RF spectrogram anomaly detection over feeding the same input representation directly to the same classical classifier. On a controlled benchmark built from real LTE signals with synthetic jamming and interference, the best configuration reaches a test AUROC of 0.8778 and F1 of 0.7995, with QKS beating its matched classical baseline on all 15 representation–readout pairs. The gains are specific to Discrete Cosine Transform inputs; raw spectrograms and PCA features do not benefit. The paper also claims the pipeline survives real quantum hardware execution with AUROC deviations below 0.013 relative to simulation. If these claims hold, QKS offers a practical way to boost simple classical detectors on structured time–frequency data without training a full quantum model.

What carries the argument

The load-bearing machinery is the multi-depth QKS featurizer: a randomly initialized quantum circuit that, for each of E episodes and D layers, projects the classical representation through random affine maps into RX/RY rotation angles on n qubits, entangles qubits with a fixed CZ ring between layers, and reads out all one- and two-body Pauli expectation values. Concatenated across episodes, this produces a high-dimensional randomized feature vector ξ(x)∈R^{E|O_n|} that a lightweight classical readout (linear SVM, logistic regression, or approximate-kernel lift) turns into a detection score. The other essential component is the DCT input representation, which selects the upper-left k_f×k_t b

What would settle it

Run the Stage 5 benchmark with a classical randomized feature baseline—e.g., random kitchen sinks or random Fourier features—whose output dimension equals E|O_n| for the QKS configuration, applied to the same DCT input and same readouts. If this equal-capacity classical model matches or exceeds the QKS AUROC of 0.8778, the quantum-specific claim is falsified. Alternatively, hold the feature dimension fixed by reducing QKS episodes until its feature count equals the direct-input dimension and check whether the gain disappears.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central finding is that multi-depth quantum kitchen sinks act as a representation enhancer rather than a universal classifier: when the input is first compressed to a small block of low-frequency DCT coefficients, a randomly initialized quantum circuit that re-uploads the data across several entangled layers produces features that a linear SVM or logistic regression can separate better than the same DCT vector read out directly. Across all fifteen finalized representation–readout comparisons on the untouched test set, the QKS pipeline outranks its matched direct-readout baseline, with the largest gains for linear readouts. The effect requires the DCT representat

Load-bearing premise

The comparison assumes that matching input representation and readout family while allowing the QKS feature vector to grow to many times the input dimension is fair; if a classical baseline with the same feature dimensionality closes the AUROC gap, the central claim that the quantum feature map causes the improvement collapses.

Editorial extensions

If this is right

  • Spectrum monitoring systems could adopt DCT preprocessing plus a shallow-to-moderate QKS feature map to boost simple detectors without training a variational quantum circuit.
  • The peak at depth 4 and the decline at depth 8 imply hardware budgets can stay small; the benefit does not require deepest circuits.
  • The small hardware–simulation AUROC gap suggests the feature geometry is robust to current-device noise, encouraging larger-scale QPU validation.
  • The DCT-specificity warns that quantum feature maps do not help on every representation; representation choice dominates the outcome.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The matched comparison is confounded by feature dimension: for dct64x64 with n=10, E=64, QKS produces 27,840 features versus 4,096 for direct readout, so part or all of the AUROC gain could come from greater classical model capacity rather than the quantum map. A classical random-feature baseline matched on dimension would settle this.
  • The DCT advantage suggests the quantum map is exploiting a compressed, nearly linear subspace; a classical kernel method (e.g., RFF) on the same DCT vector might reproduce most of the gain, though the paper's RFF readouts still show QKS ahead.
  • Hardware validation used only 4 qubits and single depth, far below the best simulated configuration; whether the 10-qubit depth-4 entangled model survives real-device noise is untested.
  • The paper makes no quantum-advantage claim; the practical value, if any, is as a low-training-cost representation enhancer, not as a fundamentally more powerful model.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes a multi-depth Quantum Kitchen Sinks (QKS) pipeline with repeated data re-uploading and ring entanglement for RF spectrogram anomaly detection. It evaluates the pipeline on a dataset built from real measured LTE signals with synthetically injected anomalies, using a five-stage validation-locked protocol that separates architecture, depth, episode budget, input representation, and classical readout. The authors report that DCT representations dominate raw and PCA inputs, that moderate-depth entangled QKS configurations perform best, and that QKS improves over direct-readout baselines across all 15 final representation–readout pairs on the held-out test set, with best test AUROC 0.8778 and best test F1 0.7995. A reduced real-hardware study on ibm_quebec shows AUROC deviations within 0.013 of simulation. The paper explicitly disclaims a universal quantum-advantage claim, but presents QKS as a representation enhancer in the DCT regime.

Significance. If the central claim is established, the paper would be a valuable benchmark for hybrid quantum-classical anomaly detection on spectrogram data. The strengths are real: the five-stage protocol is carefully designed to avoid test-set leakage, Stages 1–4 use a separate model-selection subset, Stage 5 evaluates once on untouched test data, and the paired direct-readout comparisons share data partitions, representation, and readout family. The real-device validation on ibm_quebec with the simulator re-run under identical settings is a useful realism check. The paper is also appropriately cautious in its limitations, explicitly disclaiming quantum advantage and acknowledging that stronger classical heads may reduce observed gains. However, the main empirical claim currently rests on a comparison that is not capacity-matched, so the conclusion that QKS improves over classical baselines is narrower than the abstract implies. The missing classical random-feature control is the key obstacle.

major comments (3)
  1. [§III-C, Table III] The 'matched' baseline is not capacity-matched. The QKS feature vector ξ(x) has dimension E|O_n|, while the direct readout operates on r(x) in R^d. For dct64x64, d=4096 but E|O_n|=64×435=27840; for dct128x128 it is 27840 vs 16384. A linear readout on a higher-dimensional random nonlinear lift is generally more expressive than a linear readout on the raw linear representation. For dct256x256 the QKS dimension (128×435=55680) is actually lower than the direct dimension (65536), so raw dimension alone cannot explain the gains, but the nonlinear lift nevertheless changes the hypothesis class. Because the paper claims 'only the feature map changes' (§III-C), the missing control is a classical random-feature baseline at the same output dimension and with the same readout family, e.g., random Fourier features or classical random kitchen sinks applied directly to r(x). The paper's own Limitation
  2. [§IV-B, §V(e)] The completed benchmark is a single finalized run with no seed ensemble. The affine maps W,b are sampled randomly, and Stages 1–4 use a 20% stratified model-selection subset; both the random initialization and the specific subset choice can affect which configurations advance to Stage 5. The test results in Table III and Figure 6 are therefore point estimates without uncertainty quantification. The paper's Limitations (§V(e)) acknowledge that stability across random initializations 'has not been quantified.' Given that the central claim is that QKS systematically improves over baselines across all 15 pairs, the authors should either run multiple seeds/splits and report mean±std, or substantially temper the claim. This is a correctness-risk concern, not merely a presentation issue.
  3. [§IV-C, Table IV] The real-hardware validation uses a much reduced configuration—4 qubits, depth 1, 32 episodes, Z/ZZ observables only, and about 10% of the data—whereas the main benchmark's best configurations use 10 qubits, depth 4, 64 episodes, and the full one- and two-body Pauli family. The AUROC deviations below 0.013 are thus measured for a different pipeline than the one that produces the paper's headline results. The hardware experiment is a legitimate sanity check of near-term viability, but the abstract's phrasing 'bridges two levels of realism... with AUROC deviations below 0.013 relative to simulation' overstates its scope. Please clarify that the hardware validation does not cover the main benchmark configuration.
minor comments (5)
  1. [§IV-A] Typo: 'Because The 10-qubit...' should read 'Because the 10-qubit...'.
  2. [Abstract, §IV-C] Consistent spacing for the processor name: 'ibm_quebecQuantum' should be 'ibm_quebec Quantum'.
  3. [§III-B, Eq. (13)] The notation 'E|O_n|' is used but not explicitly defined as the product E · |O_n|; consider defining it in the text to avoid ambiguity.
  4. [Figure 5] The non-DCT bars (pca32, pca128, etc.) are visually compressed and hard to read; adding numerical value labels or a table would improve clarity.
  5. [References] Reference [22] lacks venue and page information; check for completeness.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the central QKS-vs-baseline claim is a held-out empirical benchmark; the feature-dimension/capacity mismatch is a validity concern, not a circular reduction.

full rationale

The paper's central claim is an empirical result, not a derivation that reduces to its inputs. The five-stage protocol explicitly keeps the test split untouched until the final stage: Stages 1-4 operate only on model-selection and validation subsets, and Stage 5 'retrains the finalists on the full raw training split and evaluates them once on the untouched raw test split.' The reported test AUROC/F1 values are thus out-of-sample measurements, not fitted quantities renamed as predictions. Hyperparameters such as sigma=2 and the final DCT configurations are selected on validation data, not on the test set. The 'matched direct-readout baseline' is defined as the same readout family applied directly to r(x) instead of to xi(x); this is a controlled comparison, even though the QKS feature vector has a different dimensionality and nonlinear character. That difference is a capacity/validity confound, and the paper itself flags it: the Limitations state that 'stronger classical heads may reduce or eliminate the observed QKS gains' and that only 'one finalized run rather than a broad seed ensemble' was performed. A possible confound is not circularity unless the result is equivalent to the input by construction, which is not the case here. The self-citations in the introduction ([18], [19], [21], [22]) are related-work references for quantum machine learning interest; they are not used to justify the central benchmark claim, and no uniqueness theorem or prior-work-derived ansatz is imported as a load-bearing premise. The hardware validation compares real-device AUROC to a simulator rerun under identical settings, and the deviations are measured, not fitted. Overall, no circular step meeting the required evidentiary standard can be identified.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central claim is an empirical benchmark rather than a derivation, so the ledger is dominated by model-selection degrees of freedom and domain assumptions. The main uncharged burden is the assumption that a same-family/same-representation readout is a fair 'matched' baseline despite large feature-dimension differences, plus the single-run/no-error-bars design that the authors explicitly acknowledge.

free parameters (5)
  • sigma (angle-map weight scale) = 2
    Paper states sigma=2 'was found empirically to yield good results' (Sec. III-B); all W entries are N(0, sigma^2), so this scale shapes feature geometry.
  • Finalist QKS architecture (n, E, D, entanglement, DCT size) = n=10, E=64, D=4, ent=True, dct64x64/dct128x128/dct256x256
    Chosen in Stages 1-4 by validation AUROC/F1 on an ~20% training subset; test results depend on these selections.
  • F1 threshold = 0.5
    F1 is computed by thresholding scores at 0.5 (Stage 5 protocol); a different threshold would change F1.
  • Kernel lift dimension for RFF/Nystrom readouts = 1024
    phi_RFF and phi_Nys are fixed to R^1024 before the final linear layer (Sec. III-C), capping classical nonlinear capacity in those baselines.
  • Model-selection subset fraction = ~20% of training split
    Stages 1-4 use a stratified 20% sample of the training split; the composition of this subset influences which configurations advance.
assumptions (5)
  • domain assumption Simulation expectation values are treated as exact (noiseless statevector).
    QKS features in simulation are expectation values of Pauli observables; the paper does not state whether shot noise is simulated, so hardware-vs-simulator deviations mix noise with backend differences.
  • domain assumption WASD real LTE IQ measurements are representative sub-6 GHz cellular signals.
    The data side uses real measured LTE from [26] and assumes relevance to current LTE/5G NR shared bands.
  • domain assumption Synthetic anomalies (chirp, barrage, frequency-hopping) adequately represent the threats of interest.
    Only three synthetic anomaly types at JSR -10 to 5 dB are considered; generalization to unseen interference is untested.
  • ad hoc to paper A 20% stratified sample preserves the task's discriminative structure.
    All architecture selection is done on this subset; if unrepresentative, selected configurations and test results could shift.
  • ad hoc to paper Equal readout family and input representation define a 'matched' baseline.
    The paper claims 'only the feature map changes' between direct readout and QKS, but feature dimensionality also changes by up to roughly 7x, so this equality assumption is load-bearing.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RF Spectrogram Anomaly Detection with Quantum Kitchen Sinks: Architecture, Representation, and Hardware Validation." pith.science (2026). https://pith.science/paper/45QCJAWZ

@misc{pith2026260713897,
  author       = {Pith},
  title        = {Pith review of: RF Spectrogram Anomaly Detection with Quantum Kitchen Sinks: Architecture, Representation, and Hardware Validation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/45QCJAWZ}},
  note         = {Machine review of arXiv:2607.13897}
}
read the original abstract

The broadcast nature of wireless channels exposes radio-frequency (RF) networks to anomalous and malicious transmissions, making anomaly detection a fundamental requirement for secure spectrum management. Quantum Kitchen Sinks (QKS) offer a lightweight hybrid quantum feature map suitable for near-term quantum devices, yet their behavior on structured signal data remains poorly understood. In this paper, we extend the standard QKS template with multi-depth data re-uploading and ring entanglement, and evaluate the resulting pipeline on controlled RF spectrogram anomaly detection. We introduce a validation-locked five-stage ablation protocol that systematically separates the effects of shallow architecture, re-uploading depth, episode budget, input representation, and classical readout. Across the completed benchmark, Discrete Cosine Transform (DCT) representations consistently dominate raw and Principal Component Analysis (PCA) inputs, moderate-depth entangled QKS configurations form the strongest operating regime, and QKS improves over matched classical direct-readout baselines across all evaluated representation-readout pairs on the held-out test set, with the best configuration reaching a test Area Under the Receiver Operating Characteristic curve (AUROC) of 0.8778 and a test F1 of 0.7995. The study bridges two levels of realism: real measured sub-6\,GHz cellular signals on the data side and real-device validation on the ibm_quebec Quantum Processing Unit (QPU) on the computing side, with AUROC deviations below 0.013 relative to simulation. These results provide a practical, reproducible framework for deploying QKS-based anomaly detection in wireless networks.

Figures

Figures reproduced from arXiv: 2607.13897 by the authors.

Figure 1
Figure 1. Diagram of the data acquisition setup. It depicts the measurement of LTE signals, with simulated anomalous signals [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Example RF spectrograms from the dataset used in this study: (a) normal spectrogram; (b)–(d) spectrograms illustrating [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Illustrative 3-qubit multi-depth QKS episode schematic showing only the first and final layer-specific RX-then-RY uploads. CZ-ring entanglers are inserted only between successive upload layers, so the displayed pattern is layer 1 upload → CZ ring → · · · → CZ ring → layer D upload. The special case D = 1 recovers the shallow template. III. QKS PIPELINE AND ABLATION PROTOCOL A. Input representations Let x ∈ R 400×400… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Validation-stage coreset results used to select finalists across Stages 1–3 of the ablation protocol. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Stage 4 comparison across input representations. Each bar reports the validation AUROC obtained by applying the two [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Stage 5 paired test AUROC comparison between matched direct-readout baselines and QKS pipelines. Each row [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 1 linked inside Pith

  1. [1]

    Intelligent Jamming Strategies for Secure Spectrum Sharing Systems,

    X. Jiang, P. Li, B. Li, Y . Zou, and R. Wang, “Intelligent Jamming Strategies for Secure Spectrum Sharing Systems,”IEEE Trans. Commun., vol. 70, no. 2, pp. 1153–1167, 2022

    2022

  2. [2]

    A Survey on Wireless Security: Technical Challenges, Recent Advances, and Future Trends,

    Y . Zou, J. Zhu, X. Wang, and L. Hanzo, “A Survey on Wireless Security: Technical Challenges, Recent Advances, and Future Trends,”Proc. IEEE, vol. 104, no. 9, pp. 1727–1765, 2016

    2016

  3. [3]

    RF Jamming Dataset: A Wireless Spectral Scan Approach for Malicious Interference Detection,

    A. S. Ali, W. T. Lunardi, G. Singh, L. Bariah, M. Baddeley, M. A. Lopez, J.-P. Giacalone, and S. Muhaidat, “RF Jamming Dataset: A Wireless Spectral Scan Approach for Malicious Interference Detection,”IEEE Commun. Mag., vol. 62, no. 11, pp. 114–120, 2024

    2024

  4. [4]

    LTE/LTE-A Jamming, Spoofing, and Sniffing: Threat Assessment and Mitigation,

    M. Lichtman, R. P. Jover, M. Labib, R. Rao, V . Marojevic, and J. H. Reed, “LTE/LTE-A Jamming, Spoofing, and Sniffing: Threat Assessment and Mitigation,”IEEE Commun. Mag., vol. 54, no. 4, pp. 54–61, 2016

    2016

  5. [5]

    A Communications Jamming Taxonomy,

    M. Lichtman, J. D. Poston, S. Amuru, C. Shahriar, T. C. Clancy, R. M. Buehrer, and J. H. Reed, “A Communications Jamming Taxonomy,” IEEE Secur. Priv., vol. 14, no. 1, pp. 47–54, 2016

    2016

  6. [6]

    Jamming Attacks and Anti-Jamming Strategies in Wireless Networks: A Comprehensive Survey,

    H. Pirayesh and H. Zeng, “Jamming Attacks and Anti-Jamming Strategies in Wireless Networks: A Comprehensive Survey,”IEEE Commun. Surv. Tutor., vol. 24, no. 2, pp. 767–809, 2022

    2022

  7. [7]

    Dynamic Spectrum Anti-Jamming Communications: Challenges and Opportunities,

    X. Wang, J. Wang, Y . Xu, J. Chen, L. Jia, X. Liu, and Y . Yang, “Dynamic Spectrum Anti-Jamming Communications: Challenges and Opportunities,” IEEE Commun. Mag., vol. 58, no. 2, pp. 79–85, 2020

    2020

  8. [8]

    Anomaly Detection in Wireless Sensor Networks,

    S. Rajasegarar, C. Leckie, and M. Palaniswami, “Anomaly Detection in Wireless Sensor Networks,”IEEE Wireless Commun., vol. 15, no. 4, pp. 34–40, 2008

    2008

Show all 36 references
  1. [9]

    A Survey of Spectrum Sensing Algorithms for Cognitive Radio Applications,

    T. Yucek and H. Arslan, “A Survey of Spectrum Sensing Algorithms for Cognitive Radio Applications,”IEEE Commun. Surv. Tutor., vol. 11, no. 1, pp. 116–130, 2009

    2009

  2. [10]

    Digital Twin of the Radio Environment: A Novel Approach for Anomaly Detection in Wireless Networks,

    A. Krause, M. D. Khursheed, P. Schulz, F. Burmeister, and G. Fettweis, “Digital Twin of the Radio Environment: A Novel Approach for Anomaly Detection in Wireless Networks,” in2023 IEEE Globecom Workshops (GC Wkshps). IEEE, 2023, pp. 1307–1312

    2023

  3. [11]

    Jamming Signals Classification Using Convolutional Neural Network,

    Z. Wu, Y . Zhao, Z. Yin, and H. Luo, “Jamming Signals Classification Using Convolutional Neural Network,” in2017 IEEE International Symposium on Signal Processing and Information Technology (ISSPIT). IEEE, 2017, pp. 062–067

    2017

  4. [12]

    A neural network approach for wireless spectrum anomaly detection in 5G-unlicensed network,

    H. Xu, X. Ma, C. Wang, X. Wang, C. Xu, F. Gao, and L. Kong, “A neural network approach for wireless spectrum anomaly detection in 5G-unlicensed network,”CCF Transactions on Pervasive Computing and Interaction, vol. 4, no. 4, pp. 465–473, 2022

    2022

  5. [13]

    Unsupervised Wireless Spectrum Anomaly Detection With Interpretable Features,

    S. Rajendran, W. Meert, V . Lenders, and S. Pollin, “Unsupervised Wireless Spectrum Anomaly Detection With Interpretable Features,”IEEE Trans. Cogn. Commun. Netw., vol. 5, no. 3, pp. 637–647, 2019

    2019

  6. [14]

    Deep Predictive Coding Neural Network for RF Anomaly Detection in Wireless Networks,

    N. Tandiya, A. Jauhar, V . Marojevic, and J. H. Reed, “Deep Predictive Coding Neural Network for RF Anomaly Detection in Wireless Networks,” in2018 IEEE International Conference on Communications Workshops (ICC Workshops). IEEE, 2018, pp. 1–6

    2018

  7. [15]

    Using Deep Convolutional Neural Network to Recognize LTE Uplink Interference,

    J. Ren, X. Zhang, and Y . Xin, “Using Deep Convolutional Neural Network to Recognize LTE Uplink Interference,” in2019 IEEE Wireless Communications and Networking Conference (WCNC). IEEE, 2019, pp. 1–6

    2019

  8. [16]

    A Radio Anomaly Detection Algorithm Based on Modified Generative Adversarial Network,

    X. Zhou, J. Xiong, X. Zhang, X. Liu, and J. Wei, “A Radio Anomaly Detection Algorithm Based on Modified Generative Adversarial Network,” IEEE Wireless Commun. Lett., vol. 10, no. 7, pp. 1552–1556, 2021

    2021

  9. [17]

    Wireless Sensing in Artificial Intelligence of Things: A General Quantum Machine Learning Framework,

    P. Liao, X. Wang, Y . Shan, L. An, and S. Mao, “Wireless Sensing in Artificial Intelligence of Things: A General Quantum Machine Learning Framework,”IEEE Netw., vol. 39, no. 3, pp. 207–214, 2025

    2025

  10. [18]

    QuaCK-TSF: Quantum- classical kernelized time series forecasting,

    A. Aaraba, S. Cherkaoui, O. Ahmad, J.-F. Laprade, O. Nahman- Lévesque, A. Vieloszynski, and S. Wang, “QuaCK-TSF: Quantum- classical kernelized time series forecasting,” in2024 IEEE International Conference on Quantum Computing and Engineering (QCE), vol. 01, 2024, pp. 1628–1638

    2024

  11. [19]

    LatentQGAN: A hybrid QGAN with classical convolutional autoencoder,

    A. Vieloszynski, S. Cherkaoui, O. Ahmad, J.-F. Laprade, O. Nahman- Lévesque, A. Aaraba, and S. Wang, “LatentQGAN: A hybrid QGAN with classical convolutional autoencoder,” in2024 IEEE 10th World Forum on Internet of Things (WF-IoT), 2024, pp. 1–7

    2024

  12. [20]

    Quantum Machine Learning for Anomaly Detection: The Future of Smarter and Safer IoT Networks,

    I. Ahmad, L. Yang, Y . Alkhrijah, A. Almadhor, M. A. Alawad, L. Peng, and P.-H. Ho, “Quantum Machine Learning for Anomaly Detection: The Future of Smarter and Safer IoT Networks,”IEEE Netw., vol. 40, no. 2, pp. 30–38, 2026

    2026

  13. [21]

    Learning Gaussian processes with randomized quantum local kernels,

    A. Aaraba, S. Cherkaoui, O. Ahmad, J.-F. Laprade, and S. Wang, “Learning Gaussian processes with randomized quantum local kernels,” in 2026 International Conference on Quantum Communications, Networking, and Computing (QCNC), 2026, pp. 660–667

    2026

  14. [22]

    Multivariate time series forecasting with gate-based quantum reservoir computing on NISQ hardware,

    W. Hamhoum, S. Cherkaoui, J.-F. Laprade, O. Ahmed, and S. Wang, “Multivariate time series forecasting with gate-based quantum reservoir computing on NISQ hardware,” 2025

    2025

  15. [23]

    Quantum kitchen sinks: An algorithm for machine learning on near-term quantum computers,

    C. M. Wilson, J. S. Otterbach, N. Tezak, R. S. Smith, G. E. Crooks, and M. P. da Silva, “Quantum kitchen sinks: An algorithm for machine learning on near-term quantum computers,” 2018. [Online]. Available: https://arxiv.org/abs/1806.08321

    2018 arXiv

  16. [24]

    Variational quantum algorithms,

    M. Cerezo, A. Arrasmith, R. Babbush, S. C. Benjamin, S. Endo, K. Fujii, J. R. McClean, K. Mitarai, X. Yuan, L. Cincio, and P. J. Coles, “Variational quantum algorithms,”Nature Reviews Physics, vol. 3, no. 9, pp. 625–644, 2021

    2021

  17. [25]

    A Systematic Review on Quantum Machine Learning Applications in Classification,

    E. Mohammadisavadkoohi, N. Shafiabady, and J. Vakilian, “A Systematic Review on Quantum Machine Learning Applications in Classification,” IEEE Trans. Artif. Intell., vol. 7, no. 1, pp. 4–21, 2026

    2026

  18. [26]

    Wireless Anomaly Signal Dataset (W ASD): An Open Dataset for Wireless Cellular Spectrum Monitoring and Anomaly Detection,

    J. Kim, H. Kim, and B. Kim, “Wireless Anomaly Signal Dataset (W ASD): An Open Dataset for Wireless Cellular Spectrum Monitoring and Anomaly Detection,”IEEE Access, vol. 12, pp. 196 240–196 248, 2024

    2024

  19. [27]

    Spectrum Anomaly Detection Using Deep Neural Networks: A Wireless Signal Perspective,

    J. Kim, S. Oh, H. Kim, and B. Kim, “Spectrum Anomaly Detection Using Deep Neural Networks: A Wireless Signal Perspective,”IEEE Access, vol. 13, pp. 154 948–154 959, 2025

    2025

  20. [28]

    From 5G to 6G Networks: A Survey on AI-Based Jamming and Interference Detection and Mitigation,

    P. Lohan, B. Kantarci, M. A. Ferrag, N. Tihanyi, and Y . Shi, “From 5G to 6G Networks: A Survey on AI-Based Jamming and Interference Detection and Mitigation,”IEEE Open J. Commun. Soc., vol. 5, pp. 3920–3974, 2024

    2024

  21. [29]

    Low Computational Enhancement of STFT-Based Parameter Estimation,

    B. Kim, S.-H. Kong, and S. Kim, “Low Computational Enhancement of STFT-Based Parameter Estimation,”IEEE J. Sel. Top. Signal Process., vol. 9, no. 8, pp. 1610–1619, 2015

    2015

  22. [30]

    The JPEG Still Picture Compression Standard,

    G. Wallace, “The JPEG Still Picture Compression Standard,”IEEE Trans. Consum. Electron., vol. 38, no. 1, pp. xviii–xxxiv, 1992

    1992

  23. [31]

    Data re-uploading for a universal quantum classifier,

    A. Pérez-Salinas, A. Cervera-Lierta, E. Gil-Fuster, and J. I. Latorre, “Data re-uploading for a universal quantum classifier,”Quantum, vol. 4, p. 226, 2020

    2020

  24. [32]

    Effect of data encoding on the expressive power of variational quantum-machine-learning models,

    M. Schuld, R. Sweke, and J. J. Meyer, “Effect of data encoding on the expressive power of variational quantum-machine-learning models,” Physical Review A, vol. 103, p. 032430, 2021

    2021

  25. [33]

    Support-Vector Networks,

    C. Cortes and V . Vapnik, “Support-Vector Networks,”Machine Learning, vol. 20, no. 3, pp. 273–297, 1995

    1995

  26. [34]

    The Regression Analysis of Binary Sequences,

    D. R. Cox, “The Regression Analysis of Binary Sequences,”Journal of the Royal Statistical Society: Series B (Methodological), vol. 20, no. 2, pp. 215–232, 1958

    1958

  27. [35]

    Random Features for Large-Scale Kernel Machines,

    A. Rahimi and B. Recht, “Random Features for Large-Scale Kernel Machines,” inAdvances in Neural Information Processing Systems 20. Curran Associates, Inc., 2007, pp. 1177–1184

    2007

  28. [36]

    Using the Nyström Method to Speed Up Kernel Machines,

    C. K. I. Williams and M. Seeger, “Using the Nyström Method to Speed Up Kernel Machines,” inAdvances in Neural Information Processing Systems 13. MIT Press, 2001, pp. 682–688

    2001

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.