pith. machine review for the scientific record. sign in

arxiv: 2604.16953 · v1 · submitted 2026-04-18 · 🪐 quant-ph · cs.AI· cs.CV· cs.LG

Recognition: unknown

Hybrid Quantum Neural Networks for Enhanced Breast Cancer Thermographic Classification: A Novel Quantum-Classical Integration Approach

Riza Alaudin Syah , Irwan Alnarus Kautsar , Gunawan Witjaksono , Haza Nuzly Bin Abdull Hamed
Authors on Pith no claims yet

Pith reviewed 2026-05-10 06:48 UTC · model grok-4.3

classification 🪐 quant-ph cs.AIcs.CVcs.LG
keywords hybrid quantum neural networksbreast cancer classificationthermographic imagesvariational quantum circuitsquantum machine learningmedical image analysisattention mechanismsconvolutional neural networks
0
0 comments X

The pith

A hybrid quantum-classical neural network outperforms classical models on breast cancer thermographic image classification.

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

The paper introduces a hybrid quantum neural network that combines a 4-qubit variational quantum circuit with classical convolutional layers and multi-head attention mechanisms. The quantum circuit uses strongly entangling layers to encode thermal features in a quantum-aware way, while the classical part performs pattern recognition and feature fusion. Experiments on breast cancer thermographic data show the hybrid model achieves higher accuracy, faster convergence, and better feature representation than state-of-the-art classical architectures. A reader would care because the approach claims to deliver quantum benefits while remaining runnable on near-term devices through simulation. If correct, it offers a concrete path to apply quantum principles to medical imaging tasks without waiting for large-scale quantum hardware.

Core claim

The central claim is that integrating parameterized quantum circuits with multi-head attention into a convolutional neural network produces superior classification of breast cancer thermograms, with the 4-qubit variational circuit and strongly entangling layers providing quantum-enhanced feature encoding that yields measurable gains in accuracy and training dynamics over purely classical baselines.

What carries the argument

The 4-qubit variational quantum circuit with strongly entangling layers for quantum-aware feature encoding, fused via multi-head attention with classical convolutional layers.

If this is right

  • The hybrid model demonstrates superior convergence dynamics compared with classical architectures during training.
  • Enhanced feature representation from the quantum component improves overall classification performance on thermographic images.
  • The architecture establishes a deployable framework for quantum-classical hybrids in medical imaging applications.
  • The approach remains computationally feasible on near-term devices when the quantum circuit is simulated classically.
  • It addresses practical challenges in quantum machine learning deployment for healthcare without requiring full fault-tolerant quantum computers.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Future tests on actual quantum hardware could reveal whether simulation already captures the full benefit or if hardware noise and entanglement provide additional gains.
  • The same hybrid pattern might transfer to other thermal or infrared medical datasets where classical CNNs currently struggle with subtle pattern differences.
  • Without ablations, the paper leaves open the possibility that classical attention alone could explain much of the reported lift, which would narrow the claimed quantum contribution.
  • Scaling the qubit count or testing on larger thermographic cohorts would clarify whether the 4-qubit design generalizes or hits a representation limit.

Load-bearing premise

The observed performance gains come specifically from the quantum circuit rather than from the added attention mechanisms, hyperparameter choices, or data preprocessing steps.

What would settle it

An ablation experiment that removes the quantum circuit, keeps the attention layers and all other classical components identical, and retrains on the same thermographic dataset, then finds no statistically significant drop in accuracy or convergence speed.

Figures

Figures reproduced from arXiv: 2604.16953 by Gunawan Witjaksono, Haza Nuzly Bin Abdull Hamed, Irwan Alnarus Kautsar, Riza Alaudin Syah.

Figure 1
Figure 1. Figure 1: Augmented training samples from the breast cancer thermography dataset. Top row: normal tissue samples (label 0) with applied data augmentations [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: (Left) Training dynamics showing HQNN convergence with 98.11% [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: (Top) Confusion matrix showing high precision for both normal [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Quantum circuit architecture with 4 qubits, 2 variational layers, and entangling CNOT gates for enhanced feature correlation learning. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
read the original abstract

Breast cancer diagnosis through thermographic image analysis remains a critical challenge in medical AI, with classical deep learning approaches facing limitations in complex thermal pattern classification tasks. This paper presents a novel Hybrid Quantum Neural Network (HQNN) architecture that integrates quantum computing principles with classical convolutional neural networks for enhanced breast cancer classification. Our approach employs parameterized quantum circuits with multi-head attention mechanisms for quantum-aware feature encoding, coupled with classical convolutional layers for comprehensive pattern recognition. The quantum component utilizes a 4qubit variational circuit with strongly entangling layers, while the classical component incorporates advanced attention mechanisms for feature fusion. Experimental validation on breast cancer thermographic data demonstrates substantial performance improvements over state-of-the-art classical architectures, with the quantum-enhanced approach exhibiting superior convergence dynamics and enhanced feature representation capabilities. Our findings provide evidence for quantum advantage in medical image classification through classical simulation, establishing a framework for quantum-classical hybrid systems in healthcare applications. The methodology addresses key challenges in quantum machine learning deployment while maintaining computational feasibility on near-term quantum devices.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes a Hybrid Quantum Neural Network (HQNN) that integrates a 4-qubit variational quantum circuit with strongly entangling layers and multi-head attention mechanisms into a classical convolutional neural network backbone for classifying breast cancer thermographic images. It claims that this architecture delivers substantial performance improvements over state-of-the-art classical models, superior convergence, and enhanced feature representation, thereby furnishing evidence of quantum advantage even when the quantum component is classically simulated.

Significance. If the claimed gains were substantiated by quantitative metrics, baselines, ablations, and statistical tests, the work would offer a concrete example of a near-term hybrid quantum-classical pipeline for medical imaging and could help define practical integration strategies for variational circuits with attention-based classical modules. At present the absence of any numerical evidence or isolation of the quantum contribution limits the result to a conceptual proposal rather than a demonstrated advance.

major comments (2)
  1. [Abstract] Abstract: the assertion of 'substantial performance improvements over state-of-the-art classical architectures' and 'quantum advantage' is made without any reported accuracy, AUC, F1 scores, dataset size, baseline details, error bars, or statistical significance tests. The central experimental claim therefore cannot be evaluated from the text.
  2. [Abstract] Abstract: the architecture description includes classical multi-head attention, convolutional layers, and feature fusion in addition to the 4-qubit variational circuit; all experiments are performed by classical simulation. No ablation that removes or replaces only the quantum circuit (while holding attention, preprocessing, and hyperparameters fixed) is described, so any observed gain cannot be attributed to quantum entanglement rather than added classical capacity.
minor comments (2)
  1. [Abstract] Abstract: '4qubit' should be written '4-qubit'.
  2. [Abstract] Abstract: the precise form of the 'strongly entangling layers' and the dimension of the quantum feature map are not stated; these details should appear in the methods section.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their careful reading and constructive feedback on our manuscript. We address the major comments point by point below, indicating the revisions we will implement.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the assertion of 'substantial performance improvements over state-of-the-art classical architectures' and 'quantum advantage' is made without any reported accuracy, AUC, F1 scores, dataset size, baseline details, error bars, or statistical significance tests. The central experimental claim therefore cannot be evaluated from the text.

    Authors: We agree that the abstract, standing alone, should contain the key quantitative results to substantiate the stated claims. We will revise the abstract to report the specific accuracy, AUC, F1 scores, dataset size, baseline comparisons, error bars from repeated runs, and references to the statistical significance tests performed in the experimental section. revision: yes

  2. Referee: [Abstract] Abstract: the architecture description includes classical multi-head attention, convolutional layers, and feature fusion in addition to the 4-qubit variational circuit; all experiments are performed by classical simulation. No ablation that removes or replaces only the quantum circuit (while holding attention, preprocessing, and hyperparameters fixed) is described, so any observed gain cannot be attributed to quantum entanglement rather than added classical capacity.

    Authors: We acknowledge the importance of isolating the quantum circuit's contribution. We will add an ablation study to the revised manuscript in which the 4-qubit variational circuit is replaced by a classical layer with matched parameter count and expressivity, while holding the multi-head attention, CNN backbone, preprocessing pipeline, and all hyperparameters fixed. The results of this controlled comparison will be reported to clarify the source of any observed gains. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical claims rest on reported experiments rather than self-referential derivation

full rationale

The paper presents a hybrid quantum-classical architecture, describes its components (4-qubit variational circuit with entangling layers plus classical attention and CNN layers), and reports experimental performance on thermographic breast cancer data. No load-bearing mathematical derivation or uniqueness theorem is invoked that reduces by construction to fitted parameters or self-citations; the central claim of performance improvement is framed as an empirical outcome from training and evaluation. The explicit note that experiments use classical simulation of the quantum circuit is a methodological statement, not a definitional loop that equates the claimed quantum advantage to its own inputs. Absence of ablation studies concerns experimental isolation and is a validity issue rather than circularity in any derivation chain.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The abstract relies on standard assumptions of variational quantum circuits and classical deep learning without stating new axioms or free parameters explicitly. No invented physical entities are introduced.

free parameters (1)
  • variational circuit parameters
    The 4-qubit circuit contains trainable rotation angles whose values are optimized during training; these are free parameters fitted to the thermography data.
axioms (2)
  • standard math Parameterized quantum circuits can be efficiently simulated classically for small qubit counts
    Implicit in the statement that experiments were performed via classical simulation.
  • domain assumption Thermographic images contain extractable features that benefit from quantum feature encoding
    Assumed when the quantum circuit is inserted for feature encoding.

pith-pipeline@v0.9.0 · 5506 in / 1456 out tokens · 25444 ms · 2026-05-10T06:48:25.638359+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

18 extracted references · 1 canonical work pages

  1. [1]

    Global cancer statistics 2020: Globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries,

    H. Sung, J. Ferlay, R. L. Siegel, M. Laversanne, I. Soerjomataram, A. Je- mal, and F. Bray, “Global cancer statistics 2020: Globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries,” CA: a cancer journal for clinicians , vol. 71, no. 3, pp. 209–249, 2021

    2020

  2. [2]

    Cancer statistics, 2022,

    R. L. Siegel, K. D. Miller, H. E. Fuchs, and A. Jemal, “Cancer statistics, 2022,” CA: a cancer journal for clinicians , vol. 72, no. 1, pp. 7–33, 2022

    2022

  3. [3]

    Diagnostic concordance among pathologists interpreting breast biopsy specimens,

    J. G. Elmore, G. M. Longton, P. A. Carney, B. M. Geller, T. Onega, A. N. Tosteson, H. D. Nelson, M. S. Pepe, K. H. Allison, S. J. Schnitt et al. , “Diagnostic concordance among pathologists interpreting breast biopsy specimens,” Jama, vol. 313, no. 11, pp. 1122–1132, 2015

    2015

  4. [4]

    A survey on deep learning in medical image analysis,

    G. Litjens, T. Kooi, B. E. Bejnordi, A. A. A. Setio, F. Ciompi, M. Ghafoorian, J. A. Van Der Laak, B. Van Ginneken, and C. I. S´anchez, “A survey on deep learning in medical image analysis,” Medical image analysis, vol. 42, pp. 60–88, 2017

    2017

  5. [5]

    Dermatologist-level classification of skin cancer with deep neural networks,

    A. Esteva, B. Kuprel, R. A. Novoa, J. Ko, S. M. Swetter, H. M. Blau, and S. Thrun, “Dermatologist-level classification of skin cancer with deep neural networks,” nature, vol. 542, no. 7639, pp. 115–118, 2017

    2017

  6. [6]

    Efficientnet: Rethinking model scaling for con- volutional neural networks,

    M. Tan and Q. Le, “Efficientnet: Rethinking model scaling for con- volutional neural networks,” in International conference on machine learning. PMLR, 2019, pp. 6105–6114

    2019

  7. [7]

    Quantum machine learning,

    J. Biamonte, P. Wittek, N. Pancotti, P. Rebentrost, N. Wiebe, and S. Lloyd, “Quantum machine learning,” Nature, vol. 549, no. 7671, pp. 195–202, 2017

    2017

  8. [8]

    An introduction to quantum machine learning,

    M. Schuld, I. Sinayskiy, and F. Petruccione, “An introduction to quantum machine learning,” Contemporary Physics, vol. 56, no. 2, pp. 172–185, 2015

    2015

  9. [9]

    Classification with Quantum Neural Networks on Near Term Processors

    E. Farhi and H. Neven, “Classification with quantum neural networks on near term processors,” arXiv preprint arXiv:1802.06002 , 2018

    work page Pith review arXiv 2018

  10. [10]

    Quantum machine learning in feature hilbert spaces,

    M. Schuld and N. Killoran, “Quantum machine learning in feature hilbert spaces,” Physical review letters, vol. 122, no. 4, p. 040504, 2019

    2019

  11. [11]

    Quantum algorithms for su- pervised and unsupervised machine learning,

    S. Lloyd, M. Mohseni, and P. Rebentrost, “Quantum algorithms for su- pervised and unsupervised machine learning,” Nature Photonics, vol. 8, no. 2, pp. 129–136, 2014

    2014

  12. [12]

    Quantum convolutional neural networks for medical image classification,

    Y . Li, R. Zhou, R. Xu, J. Luo, and W. Hu, “Quantum convolutional neural networks for medical image classification,” Physical Review A , vol. 104, no. 3, p. 032418, 2021

    2021

  13. [13]

    A dataset for breast cancer histopathological image classification,

    F. A. Spanhol, L. S. Oliveira, C. Petitjean, and L. Heutte, “A dataset for breast cancer histopathological image classification,” IEEE Transactions on Biomedical Engineering , vol. 63, no. 7, pp. 1455–1462, 2016

    2016

  14. [14]

    Classification of breast cancer histology images using convolutional neural networks,

    T. Ara ´ujo, G. Aresta, E. Castro, J. Rouco, P. Aguiar, C. Eloy, A. Pol´onia, and A. Campilho, “Classification of breast cancer histology images using convolutional neural networks,” PloS one, vol. 12, no. 6, p. e0177544, 2017

    2017

  15. [15]

    A comparative study of deep learning architectures on melanoma detection,

    S. H. Kassani, P. H. Kassani, M. J. Wesolowski, K. A. Schneider, and R. Deters, “A comparative study of deep learning architectures on melanoma detection,” Tissue and Cell , vol. 58, pp. 76–83, 2019

    2019

  16. [16]

    Quantum computing models for artificial neural networks,

    S. Mangini, F. Tacchino, D. Gerace, D. Bajoni, and C. Macchiavello, “Quantum computing models for artificial neural networks,” EPL (Eu- rophysics Letters), vol. 134, no. 1, p. 10002, 2021

    2021

  17. [17]

    Advances in quantum computing,

    S. Garg and G. Ramakrishnan, “Advances in quantum computing,” Communications of the ACM , vol. 63, no. 5, pp. 42–49, 2020

    2020

  18. [18]

    Breast cancer detection using ther- mography,

    M. Thilak, “Breast cancer detection using ther- mography,” https://www.kaggle.com/datasets/thilak02/ breast-cancer-detection-using-thermography, 2023, accessed: 2024

    2023