pith. sign in

arxiv: 2605.18540 · v1 · pith:S6YR3ST6new · submitted 2026-05-18 · 🪐 quant-ph

Discovering Data Encoding Strategies for Quantum-Classical Neural Networks Using Monte Carlo Tree Search

Lena Tokuhiro , Amine Bentellis , Jeanette Miriam Lorenz This is my paper

Pith reviewed 2026-05-20 10:26 UTC · model grok-4.3

classification 🪐 quant-ph
keywords quantum machine learningdata encodingMonte Carlo Tree Searchquantum-classical networksfeature mapseffective rankmedical imaginghybrid quantum-classical models
0
0 comments X

The pith

Monte Carlo Tree Search discovers data encoding circuits that outperform standard strategies in quantum-classical neural networks on medical imaging tasks.

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

The authors explore ways to automatically find good methods for turning classical data into quantum states for use in hybrid quantum-classical models. They apply Monte Carlo Tree Search to search through possible encoding circuits in a setup that uses a fixed quantum layer for feature extraction followed by a classical neural network classifier. Tests on two medical imaging datasets show these searched encodings work better than usual fixed choices like angle or amplitude encoding. The study also finds that a measure called the effective rank of the quantum feature maps predicts which encodings will perform well, allowing quicker identification of good candidates.

Core claim

The paper establishes that Monte Carlo Tree Search can be used to find data encoding circuits for a quantum-classical convolutional neural network that lead to higher classification accuracy on medical imaging data compared to commonly used encodings. Additionally, the effective rank of the feature maps generated by these circuits shows a correlation with performance that can serve as a threshold to speed up the discovery process.

What carries the argument

Monte Carlo Tree Search over the space of possible quantum data encoding circuits, guided by the effective rank of the resulting feature maps as a performance indicator.

Load-bearing premise

The observed correlation between the effective rank of feature maps and classification performance is consistent enough to act as a reliable threshold across various datasets and model configurations without dataset-specific tuning.

What would settle it

Evaluating the discovered encodings and the effective rank threshold on an additional dataset different from the two medical imaging ones used, to check if high-rank encodings consistently yield better performance without recalibrating the threshold.

Figures

Figures reproduced from arXiv: 2605.18540 by Amine Bentellis, Jeanette Miriam Lorenz, Lena Tokuhiro.

Figure 1
Figure 1. Figure 1: Architecture sketch of the CNN and QCCNN for the [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Data encoding circuits with the highest validation AUC identified through MCTS for BreastMNIST (a, c) and [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Comparison of data encoding strategies on BreastMNIST. The MCTS-derived 2 [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Performance evaluation on (a) BreastMNIST and (b) PneumoniaMNIST datasets. Training metrics (top row of each [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Accuracy and AUC performance on the (a) BreastM [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Feature maps produced by 2×2 patch-sized quantum encodings on BreastMNIST. Circuits C1, C400, and C700 represent the best, medium, and worst performing encodings from the search, respectively. Higher-ranked circuits produce more diverse feature representations. Table I: Performance comparison of selected encoding circuits and classical baselines on the 128×128 BreastMNIST vali￾dation dataset. Results are a… view at source ↗
Figure 9
Figure 9. Figure 9: Normalized effective rank against validation AUC [PITH_FULL_IMAGE:figures/full_fig_p009_9.png] view at source ↗
Figure 8
Figure 8. Figure 8: Distribution of Fourier coefficients in the complex plane [PITH_FULL_IMAGE:figures/full_fig_p009_8.png] view at source ↗
Figure 10
Figure 10. Figure 10: Percentage of true top-10% encodings (by AUC) [PITH_FULL_IMAGE:figures/full_fig_p010_10.png] view at source ↗
read the original abstract

Quantum machine learning (QML) has attracted considerable research interest, yet whether it offers practical benefits over classical approaches remains an open question. The choice of data encoding significantly influences QML performance, but why certain encodings outperform others remains poorly understood. We employ Monte Carlo Tree Search (MCTS) to discover optimal data encoding circuits for a quantum-classical convolutional neural network (QCCNN) combining a non-variational quantum block for feature extraction with a classical classifier. Evaluating on two medical imaging datasets, the discovered circuits outperform commonly used encoding strategies while showing competitive results compared to purely classical counterparts. We further analyze metrics to identify predictors of encoding performance. Entanglement capability and Fourier decomposition provide minimal insight, whereas the effective rank of the feature maps exhibits meaningful correlation and can serve as a threshold criterion to accelerate the search for high-performing encodings. Our findings provide both a practical method for encoding discovery and new insights into what makes data encodings effective in QML.

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 / 1 minor

Summary. The manuscript applies Monte Carlo Tree Search (MCTS) to discover data-encoding circuits for a quantum-classical convolutional neural network (QCCNN) that pairs a non-variational quantum feature extractor with a classical classifier. On two medical-imaging datasets the discovered encodings are reported to outperform standard encoding strategies while remaining competitive with purely classical baselines. The authors further evaluate several candidate predictors of encoding quality and conclude that the effective rank of the resulting feature maps correlates with classification accuracy and can therefore be used as a cheap threshold to prune the MCTS search tree.

Significance. If the empirical claims are placed on firmer statistical footing and the rank-based filter is shown to generalize, the work supplies both a concrete search procedure for QML encodings and a potentially useful diagnostic for circuit quality. The explicit comparison against classical counterparts on real medical data is a positive feature; the identification of a low-cost proxy metric could reduce the computational burden of future encoding searches.

major comments (2)
  1. [Results] Results section (performance tables and figures): the reported outperformance on the two datasets is presented without error bars, without the number of independent runs, and without statistical significance tests against the chosen baselines. In the absence of these quantities it is impossible to judge whether the observed gains are reliable or could be explained by post-hoc selection of the displayed circuits.
  2. [Analysis of metrics] Analysis of metrics (effective-rank subsection): the proposal that effective rank can serve as a fixed threshold criterion rests on the correlation observed in the two medical-imaging datasets only. No cross-dataset or cross-architecture validation is provided to show that the same numeric cutoff remains predictive when the data distribution or the downstream classical head changes; without such evidence the claimed acceleration benefit is not yet load-bearing.
minor comments (1)
  1. [Abstract] The abstract states that entanglement capability and Fourier decomposition yield minimal insight, yet the main text does not detail the precise definitions or computational procedures used for these two metrics.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments on our manuscript. We address each major comment point by point below, indicating the changes we will make to strengthen the presentation and clarify the scope of our claims.

read point-by-point responses

  1. Referee: [Results] Results section (performance tables and figures): the reported outperformance on the two datasets is presented without error bars, without the number of independent runs, and without statistical significance tests against the chosen baselines. In the absence of these quantities it is impossible to judge whether the observed gains are reliable or could be explained by post-hoc selection of the displayed circuits.

    Authors: We agree that the current presentation of results lacks the statistical details needed to fully assess reliability. The experiments underlying the reported accuracies were performed with multiple random seeds, but the manuscript omitted the exact count of independent runs and any error bars or significance testing. In the revised version we will update the Results section to report the number of independent runs, add error bars (standard deviation) to all tables and figures, and include paired statistical tests (e.g., Wilcoxon signed-rank) against the baseline encodings. These additions will allow readers to evaluate whether the observed improvements are statistically meaningful. revision: yes

  2. Referee: [Analysis of metrics] Analysis of metrics (effective-rank subsection): the proposal that effective rank can serve as a fixed threshold criterion rests on the correlation observed in the two medical-imaging datasets only. No cross-dataset or cross-architecture validation is provided to show that the same numeric cutoff remains predictive when the data distribution or the downstream classical head changes; without such evidence the claimed acceleration benefit is not yet load-bearing.

    Authors: The referee correctly identifies that the observed correlation between effective rank and classification accuracy is demonstrated only on the two medical-imaging datasets used in this study. We do not claim or provide evidence that a single numeric cutoff generalizes across arbitrary data distributions or classical heads. In the revision we will add an explicit limitations paragraph in the Analysis of metrics subsection stating that the threshold may require recalibration for new tasks and that broader validation lies outside the scope of the present work. We will also tone down the language around the acceleration benefit to reflect that it is currently supported only within the medical-imaging setting examined here. revision: partial

Circularity Check

0 steps flagged

No circularity: empirical MCTS search and rank correlation validated on external datasets

full rationale

The manuscript describes an empirical workflow: Monte Carlo Tree Search explores encoding circuits for a fixed QCCNN architecture, performance is measured directly on two held-out medical imaging datasets, and the effective rank of the resulting feature maps is computed post hoc and shown to correlate with accuracy. These quantities are obtained from external data and standard matrix-rank definitions rather than from any self-referential equation or self-citation chain that would make the reported correlation or performance advantage tautological by construction. No fitted parameter is relabeled as a prediction, no uniqueness theorem is invoked, and the central claims remain falsifiable against independent benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on standard quantum-circuit assumptions and empirical evaluation; no new physical entities are postulated and the search hyperparameters are treated as engineering choices rather than fitted constants.

free parameters (1)
  • MCTS exploration hyperparameters
    Chosen by the authors to balance exploration of the encoding circuit space; their specific values affect which circuits are returned.
axioms (1)
  • domain assumption A non-variational quantum block can extract useful features from encoded data for a downstream classical classifier
    Invoked in the definition of the QCCNN architecture.

pith-pipeline@v0.9.0 · 5696 in / 1172 out tokens · 46075 ms · 2026-05-20T10:26:35.009764+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

46 extracted references · 46 canonical work pages · 2 internal anchors

  1. [1]

    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, no. 3, p. 032 430, 2021

    work page 2021

  2. [2]

    Tschöpe, V

    M. Tschöpe, V . F. Rey, S. P. Sanon, P. Lukowicz, N. Palaiodimopoulos, and M. Kiefer-Emmanouilidis, Boosting classification with quantum-inspired augmen- tations, 2025

    work page 2025

  3. [3]

    Quantum-classical convolutional neural networks in radiological image classification,

    A. Matic, M. Monnet, J. M. Lorenz, B. Schachtner, and T. Messerer, “Quantum-classical convolutional neural networks in radiological image classification,” in2022 IEEE International Conference on Quantum Computing and Engineering (QCE), IEEE, 2022, pp. 56–66

    work page 2022

  4. [4]

    Understanding the effects of data encoding on quantum-classical convolutional neural net- works,

    M. Monnet, N. Chaabani, T.-A. Dr ˘agan, B. Schachtner, and J. M. Lorenz, “Understanding the effects of data encoding on quantum-classical convolutional neural net- works,” in2024 IEEE International Conference on Quantum Computing and Engineering (QCE), IEEE, vol. 1, 2024, pp. 1436–1446

    work page 2024

  5. [5]

    Hybrid classical-quantum autoencoder for anomaly detection,

    A. Sakhnenko, C. O’Meara, K. J. B. Ghosh, C. B. Mendl, G. Cortiana, and J. Bernabé-Moreno, “Hybrid classical-quantum autoencoder for anomaly detection,” Quantum Machine Intelligence, vol. 4, no. 2, p. 27, Sep. 2022,ISSN: 2524-4914.DOI: 10 . 1007 / s42484 - 022 - 00075-z

    work page 2022

  6. [6]

    Hybrid quantum–classical convolutional neural networks with privacy quantum computing,

    S. Huang, Y . Chang, Y . Lin, and S. Zhang, “Hybrid quantum–classical convolutional neural networks with privacy quantum computing,”Quantum Science and Technology, vol. 8, no. 2, p. 025 015, 2023

    work page 2023

  7. [7]

    Generalization bounds in hybrid quantum-classical machine learning models,

    T. Wu, A. Bentellis, A. Sakhnenko, and J. M. Lorenz, “Generalization bounds in hybrid quantum-classical machine learning models,”Phys. Rev. A, vol. 113, p. 022 425, 2 Feb. 2026.DOI: 10.1103/6ckt-hlh8

    work page doi:10.1103/6ckt-hlh8 2026

  8. [8]

    Generalization in quantum machine learnin g from few training data

    M. C. Caro et al., “Generalization in quantum ma- chine learning from few training data,”Nature Com- munications, vol. 13, no. 1, p. 4919, Aug. 2022, arXiv:2111.05292 [quant-ph],ISSN: 2041-1723.DOI: 10.1038/s41467-022-32550-3

    work page doi:10.1038/s41467-022-32550-3 2022

  9. [9]

    The effective rank: A measure of effective dimensionality,

    O. Roy and M. Vetterli, “The effective rank: A measure of effective dimensionality,” in2007 15th European signal processing conference, IEEE, 2007, pp. 606–610

    work page 2007

  10. [10]

    Quantum convolu- tional neural networks,

    I. Cong, S. Choi, and M. D. Lukin, “Quantum convolu- tional neural networks,”Nature Physics, vol. 15, no. 12, pp. 1273–1278, 2019

    work page 2019

  11. [11]

    Quantum classical hybrid convolu- tional neural networks for breast cancer diagnosis,

    Q. Xiang et al., “Quantum classical hybrid convolu- tional neural networks for breast cancer diagnosis,” Scientific Reports, vol. 14, no. 1, p. 24 699, Oct. 21, 2024,ISSN: 2045-2322.DOI: 10 . 1038 / s41598 - 024 - 74778-7

    work page 2024

  12. [12]

    Digital-analog quantum convolutional neural networks for image classification,

    A. Simen et al., “Digital-analog quantum convolutional neural networks for image classification,”Physical Re- view Research, vol. 6, no. 4, p. L042060, 2024

    work page 2024

  13. [13]

    The power of quantum neural net- works,

    A. Abbas, D. Sutter, C. Zoufal, A. Lucchi, A. Figalli, and S. Woerner, “The power of quantum neural net- works,”Nature Computational Science, vol. 1, no. 6, pp. 403–409, Jun. 2021,ISSN: 2662-8457.DOI: 10 . 1038/s43588-021-00084-1

    work page 2021

  14. [14]

    Subtleties in the trainability of quantum machine learning models,

    S. Thanasilp, S. Wang, N. A. Nghiem, P. Coles, and M. Cerezo, “Subtleties in the trainability of quantum machine learning models,”Quantum Machine Intelli- gence, vol. 5, no. 1, p. 21, May 15, 2023,ISSN: 2524- 4914.DOI: 10.1007/s42484-023-00103-6

    work page doi:10.1007/s42484-023-00103-6 2023

  15. [15]

    Challenges and opportunities in quantum machine learning,

    M. Cerezo, G. Verdon, H.-Y . Huang, L. Cincio, and P. J. Coles, “Challenges and opportunities in quan- tum machine learning,”Nature Computational Science, vol. 2, no. 9, pp. 567–576, Sep. 2022,ISSN: 2662-8457. DOI: 10.1038/s43588-022-00311-3

    work page doi:10.1038/s43588-022-00311-3 2022

  16. [16]

    Cost function dependent barren plateaus in shallow parametrized quantum circuits,

    M. Cerezo, A. Sone, T. V olkoff, L. Cincio, and P. J. Coles, “Cost function dependent barren plateaus in shallow parametrized quantum circuits,”Nature Com- munications, vol. 12, no. 1, Mar. 2021,ISSN: 2041- 1723.DOI: 10.1038/s41467-021-21728-w

    work page doi:10.1038/s41467-021-21728-w 2021

  17. [17]

    Supervised learning with quantum- enhanced feature spaces,

    V . Havlí ˇcek et al., “Supervised learning with quantum- enhanced feature spaces,”Nature, vol. 567, no. 7747, pp. 209–212, 2019

    work page 2019

  18. [18]

    Better than classical? the subtle artofbenchmarkingquantummachinelearningmodels

    J. Bowles, S. Ahmed, and M. Schuld. “Better than classical? The subtle art of benchmarking quantum machine learning models. ”[Online]. Available: http : //arxiv.org/abs/2403.07059 pre-published

    work page arXiv

  19. [19]

    Fourier Fingerprints of Ansatzes in Quantum Machine Learning

    M. Strobl, M. E. Sahin, L. van der Horst, E. Kuehn, A. Streit, and B. Jaderberg. “Fourier Fingerprints of Ansatzes in Quantum Machine Learning. ”[Online]. Available: http : / / arxiv . org / abs / 2508 . 20868 pre- published

    work page

  20. [20]

    Wiedmann, M

    M. Wiedmann, M. Periyasamy, and D. D. Scherer. “Fourier Analysis of Variational Quantum Circuits for Supervised Learning. ”[Online]. Available: http://arxiv. org/abs/2411.03450 pre-published

    work page arXiv

  21. [21]

    Quantum Circuit Architecture Optimization for Vari- ational Quantum Eigensolver via Monte Carlo Tree Search,

    F.-X. Meng, Z.-T. Li, X.-T. Yu, and Z.-C. Zhang, “Quantum Circuit Architecture Optimization for Vari- ational Quantum Eigensolver via Monte Carlo Tree Search,”IEEE Transactions on Quantum Engineering, vol. 2, pp. 1–10, 2021,ISSN: 2689-1808.DOI: 10.1109/ TQE.2021.3119010

    work page arXiv 2021

  22. [22]

    Iterative qubit coupled cluster approach with efficient screening of generators,

    I. G. Ryabinkin, R. A. Lang, S. N. Genin, and A. F. Izmaylov, “Iterative qubit coupled cluster approach with efficient screening of generators,”Journal of chemical theory and computation, vol. 16, no. 2, pp. 1055–1063, 2020

    work page 2020

  23. [23]

    Reinforcement Learning for Variational Quantum Circuits Design

    S. Foderà, G. Turati, R. Nembrini, M. F. Dacrema, and P. Cremonesi. “Reinforcement Learning for Variational Quantum Circuits Design. ”[Online]. Available: http : //arxiv.org/abs/2409.05475 pre-published

    work page arXiv

  24. [24]

    Automated Design of Structured Varia- tional Quantum Circuits with Reinforcement Learning

    G. Turati, S. Foderà, R. Nembrini, M. F. Dacrema, and P. Cremonesi. “Automated Design of Structured Varia- tional Quantum Circuits with Reinforcement Learning. ”[Online]. Available: http://arxiv.org/abs/2507.16001 pre-published

    work page arXiv

  25. [25]

    Quantum Embedding Search for Quantum Machine Learning,

    N. Nguyen and K.-C. Chen, “Quantum Embedding Search for Quantum Machine Learning,”IEEE Access, 11 vol. 10, pp. 41 444–41 456, 2022,ISSN: 2169-3536. DOI: 10.1109/ACCESS.2022.3167398

    work page doi:10.1109/access.2022.3167398 2022

  26. [26]

    Reinforcement learning-based architecture search for quantum machine learning,

    F. Rapp, D. A. Kreplin, M. F. Huber, and M. Roth, “Reinforcement learning-based architecture search for quantum machine learning,”Machine Learning: Science and Technology, vol. 6, no. 1, p. 015 041, Feb. 2025, ISSN: 2632-2153.DOI: 10.1088/2632-2153/adaf75

    work page doi:10.1088/2632-2153/adaf75 2025

  27. [27]

    Automatic design of quantum feature maps,

    S. Altares-López, A. Ribeiro, and J. J. García-Ripoll, “Automatic design of quantum feature maps,”Quantum Science and Technology, vol. 6, no. 4, p. 045 015, Oct. 1, 2021,ISSN: 2058-9565.DOI: 10 . 1088 / 2058 - 9565 / ac1ab1

    work page 2021

  28. [28]

    Quantum- DARTS: Differentiable Quantum Architecture Search for Variational Quantum Algorithms,

    W. Wu, G. Yan, X. Lu, K. Pan, and J. Yan, “Quantum- DARTS: Differentiable Quantum Architecture Search for Variational Quantum Algorithms,” inProceedings of the 40th International Conference on Machine Learn- ing, PMLR, Jul. 3, 2023, pp. 37 745–37 764

    work page 2023

  29. [29]

    Josephson junctions and SQUIDs created by focused helium-ion-beam irradiation of YBa2Cu3O7,

    Z. Lu, P.-X. Shen, and D.-L. Deng, “Markovian Quan- tum Neuroevolution for Machine Learning,”Physical Review Applied, vol. 16, no. 4, p. 044 039, Oct. 21, 2021,ISSN: 2331-7019.DOI: 10.1103/PhysRevApplied. 16.044039

    work page doi:10.1103/physrevapplied 2021

  30. [30]

    Curriculum re- inforcement learning for quantum architecture search under hardware errors

    Y . J. Patel, A. Kundu, M. Ostaszewski, X. Bonet- Monroig, V . Dunjko, and O. Danaci. “Curriculum re- inforcement learning for quantum architecture search under hardware errors. ”[Online]. Available: http://arxiv. org/abs/2402.03500 pre-published

    work page arXiv

  31. [31]

    Quantum Neural Architecture Search with Quantum Circuits Metric and Bayesian Optimization

    T. Duong, S. T. Truong, M. Tam, B. Bach, J.-Y . Ryu, and J.-K. K. Rhee. “Quantum Neural Architecture Search with Quantum Circuits Metric and Bayesian Optimization. ”[Online]. Available: http://arxiv.org/abs/ 2206.14115 pre-published

    work page arXiv

  32. [32]

    Quantum Circuit Design Search,

    M. Pirhooshyaran and T. Terlaky. “Quantum Circuit Design Search,” Accessed: Oct. 23, 2025. [Online]. Available: http : / / arxiv . org / abs / 2012 . 04046 pre- published

    work page 2025

  33. [33]

    Quantum Archi- tecture Search with Unsupervised Representation Learn- ing

    Y . Sun, Z. Wu, Y . Ma, and V . Tresp. “Quantum Archi- tecture Search with Unsupervised Representation Learn- ing.” version 1, Accessed: Nov. 15, 2025. [Online]. Available: http : / / arxiv . org / abs / 2401 . 11576 pre- published

    work page 2025

  34. [34]

    Automated Quantum Circuit Design With Nested Monte Carlo Tree Search,

    P. Wang, M. Usman, U. Parampalli, L. C. L. Hollenberg, and C. R. Myers, “Automated Quantum Circuit Design With Nested Monte Carlo Tree Search,”IEEE Transac- tions on Quantum Engineering, vol. 4, pp. 1–20, 2023, ISSN: 2689-1808.DOI: 10.1109/TQE.2023.3265709

    work page doi:10.1109/tqe.2023.3265709 2023

  35. [35]

    Quantum Circuit Design using a Progres- sive Widening Enhanced Monte Carlo Tree Search,

    V . Lipardi, D. Dibenedetto, G. Stamoulis, and M. H. Winands, “Quantum Circuit Design using a Progres- sive Widening Enhanced Monte Carlo Tree Search,” Advanced Quantum Technologies, vol. 8, no. 10, e2500093, 2025,ISSN: 2511-9044.DOI: 10.1002/qute. 202500093

    work page doi:10.1002/qute 2025

  36. [36]

    Medmnist classification decathlon: A lightweight automl benchmark for medical image analysis,

    J. Yang, R. Shi, and B. Ni, “Medmnist classification decathlon: A lightweight automl benchmark for medical image analysis,” inIEEE 18th International Symposium on Biomedical Imaging (ISBI), 2021, pp. 191–195

    work page 2021

  37. [37]

    Medmnist v2-a large-scale lightweight benchmark for 2d and 3d biomedical image classifica- tion,

    J. Yang et al., “Medmnist v2-a large-scale lightweight benchmark for 2d and 3d biomedical image classifica- tion,”Scientific Data, vol. 10, no. 1, p. 41, 2023

    work page 2023

  38. [38]

    Paszke et al.,Pytorch: An imperative style, high- performance deep learning library, 2019

    A. Paszke et al.,Pytorch: An imperative style, high- performance deep learning library, 2019

    work page 2019

  39. [39]

    Expressibility and Entangling Capability of Parameterized Quantum Circuits for Hybrid Quantum-Classical Algorithms,

    S. Sim, P. D. Johnson, and A. Aspuru-Guzik, “Express- ibility and entangling capability of parameterized quan- tum circuits for hybrid quantum-classical algorithms,” Advanced Quantum Technologies, vol. 2, no. 12, Oct. 2019,ISSN: 2511-9044.DOI: 10.1002/qute.201900070

    work page doi:10.1002/qute.201900070 2019

  40. [40]

    Reducing Overfitting in Deep Networks by Decorrelating Representations

    M. Cogswell, F. Ahmed, R. Girshick, L. Zitnick, and D. Batra. “Reducing Overfitting in Deep Networks by Decorrelating Representations,” Accessed: Apr. 3, 2026. [Online]. Available: http://arxiv.org/abs/1511.06068 pre-published

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  41. [41]

    Pruning Filters for Efficient ConvNets

    H. Li, A. Kadav, I. Durdanovic, H. Samet, and H. P. Graf. “Pruning Filters for Efficient ConvNets,” Ac- cessed: Apr. 3, 2026. [Online]. Available: http://arxiv. org/abs/1608.08710 pre-published

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  42. [42]

    Hrank: Filter pruning using high-rank feature map,

    M. Lin et al., “Hrank: Filter pruning using high-rank feature map,” inProceedings of the IEEE/CVF confer- ence on computer vision and pattern recognition, 2020, pp. 1529–1538

    work page 2020

  43. [43]

    Bergholm et al.,Pennylane: Automatic differentiation of hybrid quantum-classical computations, 2022

    V . Bergholm et al.,Pennylane: Automatic differentiation of hybrid quantum-classical computations, 2022

    work page 2022

  44. [44]

    Bandit Based Monte- Carlo Planning,

    L. Kocsis and C. Szepesvári, “Bandit Based Monte- Carlo Planning,” inMachine Learning: ECML 2006, J. Fürnkranz, T. Scheffer, and M. Spiliopoulou, Eds., Berlin, Heidelberg: Springer, 2006, pp. 282–293,ISBN: 978-3-540-46056-5.DOI: 10.1007/11871842_29

    work page doi:10.1007/11871842_29 2006

  45. [45]

    Finite-time analysis of the multiarmed bandit problem

    P. Auer, N. Cesa-Bianchi, and P. Fischer, “Finite-time Analysis of the Multiarmed Bandit Problem,”Machine Learning, vol. 47, no. 2, pp. 235–256, May 1, 2002, ISSN: 1573-0565.DOI: 10.1023/A:1013689704352

    work page doi:10.1023/a:1013689704352 2002

  46. [46]

    En- coding patterns for quantum algorithms,

    M. Weigold, J. Barzen, F. Leymann, and M. Salm, “En- coding patterns for quantum algorithms,”IET Quantum Communication, vol. 2, no. 4, pp. 141–152, 2021. 12

    work page 2021