Pith. sign in

REVIEW 4 major objections 7 minor 35 references

Realizing Quantum Kernel Models at Scale with Matrix Product State Simulation

T0 review · 4 major / 7 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Using a matrix-product-state simulator, this paper runs quantum kernel classification on 165 features and 6,400 training points and reports that test performance improves with both feature count and data size.

desk verdict Solid engineering and reproducible scaling, but the headline performance claims rest on test-set-selected regularization and need a validation-split redo. read the letter →

arxiv 2411.09336 v1 pith:HMYMPEP7 submitted 2024-11-14 quant-ph

Mekena Metcalf , Pablo Andrés-Martínez , Nathan Fitzpatrick This is my paper · ORCID
classification quant-ph
keywords quantummachinelearningkernelmethodsmatrixproductstatesimulationsupportvectortensornetworksGPUaccelerationEllipticBitcoindatasetbandwidth
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

Quantum kernel methods have been hard to benchmark because simulating quantum feature maps scales exponentially with feature count. The paper builds a Matrix Product State (MPS) simulator that represents the quantum state compactly, with controlled truncation, and exploits the fact that each Gram-matrix entry is an independent inner product that can be spread across processors. Using this framework it trains a quantum-kernel support vector machine on the Elliptic Bitcoin data set with 165 features and 6,400 training points, far beyond prior experiments. The central empirical claim is that the model's test AUC continues to improve as both feature dimension and training data grow, with a 2.44% AUC gain from 100 to 165 features at the largest sample size, and that a tuned quantum kernel outperforms a Gaussian kernel on this data. If true, this is the first evidence that quantum kernel models behave well at a scale meaningful for industry data sets.

What carries the argument

The machinery is a Matrix Product State (MPS) tensor-network representation of the quantum feature state, updated with SVD truncation that keeps discarded squared singular values below $10^{-16}$, so simulation error sits at 64-bit floating-point level. A circuit ansatz encodes each data point by exponentiating an Ising Hamiltonian with single-qubit $Z$ fields and $XX$ couplings along a linear chain of tunable interaction distance $d$ (equations (4) and (5)), and the Gram matrix entry $K_{ij} = |\langle\psi(x_i)|\psi(x_j)\rangle|^2$ is computed by contracting two MPS in $O(m\chi^3)$ time. Because each entry is independent, the workload tiles across parallel processes (no-messaging or round-robin), which is what makes the 6,400-sample kernel feasible.

What would settle it

Run the scale experiment with the SVM regularization coefficient selected on a held-out validation set rather than on the test set; if test AUC then fails to rise with feature count or with sample size across 300, 1500, and 6400 training points, the paper's central scaling claim would be refuted.

Watch

Extended reading notes

Core claim

The paper's central discovery is that a quantum kernel SVM, realized through MPS simulation rather than a quantum computer, can be pushed to 165 qubits (features) and 6,400 training samples while keeping memory per state below 15 KiB for the linear ansatz, and that classification quality rises with scale. The authors report that test AUC at 6,400 samples improves steadily as features increase, that the quantum kernel outperforms the Gaussian kernel at kernel bandwidths $\gamma = 0.5$ and $1.0$, and that increasing circuit depth beyond a few layers degrades performance through kernel concentration. They also identify a runtime crossover between CPU and GPU implementations at interaction distance $d = 10$ and bond dimension $\chi \approx 320$, beyond which the GPU backend is faster for both simulation and inner-product computation.

Load-bearing premise

The reported performance improvements depend on choosing the SVM regularization coefficient that gives the best score on the test data, so those improvements could shrink or vanish if the coefficient were chosen on a separate validation set instead.

Editorial extensions

If this is right

  • Training a 6,400-sample quantum kernel on 165 features completes in about 3 hours on 32 GPUs, and extrapolations put 64,000 samples at roughly 30 hours on 320 GPUs.
  • For the simple linear-chain ansatz ($d=1$), the CPU backend is the cheaper option; the GPU backend only wins once interaction distance reaches $d=10$ and bond dimension reaches roughly 320.
  • Adding features and training samples is predicted to keep improving test AUC for quantum kernels, at least on structured tabular data like Elliptic, provided overfitting is controlled by enough samples.
  • Deeper encoding circuits hurt rather than help: test performance drops as depth grows, consistent with exponential concentration of kernel values.
  • A quantum kernel with bandwidth $\gamma = 0.5$ or $1.0$ can beat a Gaussian kernel on this classification task, while smaller $\gamma$ and larger interaction distances lead to weak encodings or overfitting.

Reading between the lines

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

  • If the improvement with feature count and data size reproduces under an independent validation-set rule for choosing the SVM regularization coefficient, the result would strengthen the case that quantum kernels can be useful classical surrogates at industrial scale.
  • The same MPS framework could be pointed at other high-dimensional tabular benchmarks (fraud, credit scoring, drug discovery) to test whether the Elliptic result is data-specific; a natural next experiment is to fix depth and bandwidth while sweeping dataset size.
  • The crossover at $\chi \approx 320$ gives a simple rule of thumb for practitioners: monitor the MPS bond dimension, and only move to GPUs when it approaches a few hundred.
  • Because the simulator is essentially noiseless, comparing these kernels against hardware noise models could reveal how much of the observed advantage survives on real devices, where exponential concentration may set in earlier.
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

4 major / 7 minor

Summary. The paper proposes a Matrix Product State (MPS) simulation framework for constructing quantum kernels, with an ITensors-based CPU backend and a new pytket-cutensornet GPU backend, together with two parallelization strategies for Gram-matrix construction. The authors benchmark runtime and memory scaling as a function of interaction distance, qubit count, and dataset size, and report a CPU/GPU crossover near interaction distance d = 10 with bond dimension about 320. They then apply the simulator to the Elliptic Bitcoin dataset with up to 165 features and 6400 training points, reporting that quantum kernel AUC improves with feature count and training set size and that a quantum circuit ansatz with suitable hyperparameters outperforms a Gaussian kernel on test metrics. The manuscript includes a detailed artifact appendix linking each figure and table to reproducible scripts and archived data.

Significance. If the performance claims survive a proper evaluation protocol, the paper would make a meaningful contribution: it demonstrates a practical MPS-based route to quantum kernel experiments at a scale (165 qubits/features, 6400 training points) well beyond what state-vector simulation permits, and it ships unusually complete computational artifacts, including machine-readable raw outputs and scripts for every figure and table. The runtime scaling study, the crossover analysis, and the parallelization strategy are valuable and appear well supported by the supplied artifacts. The classification-performance claims, however, are currently the weakest link: they rest on test-set-based selection of the SVM regularization coefficient, and the reported metric differences are presented without uncertainty estimates, so the headline "performance improves with scale" and the quantum-versus-Gaussian comparison are not yet established at the standard required for a claim of "first evidence of quantum model performance at scale."

major comments (4)
  1. [Appendix G (Artifact A5) and Appendix H (Artifact A6)] The central performance claims in Figures 9-10 and Table II are undermined by test-set-based selection of the SVM regularization coefficient C. Artifact A5 states that plot.py "picks the best AUC score among the different possible choices of the regularization coefficient," and Artifact A6 states that to_csv.py "picks the regularization coefficient with highest AUC" after averaging metrics over six samples. Because the same test data used to report performance is also used to choose C, the reported AUCs are optimistically biased, and the monotone improvement seen in Figure 10 (e.g., the 2.44% gain from 100 to 165 features at 6400 samples) could be inflated by selecting the best of many C values per experiment. The same issue affects the quantum-versus-Gaussian comparison in Table II, since the Gaussian kernel's regularization coefficient is selected in the same way. I request a reanalysis using a held-out validation set for C selection, or nested cross-validation, with the selected C then applied to the reported test set. Without this, the claims C2.1 and C2.2 are not supported.
  2. [Section III-B and Table II] No uncertainty quantification is provided for the averaged metrics. The text says "Each run consists of 6 data samples and the metrics are averaged over the 6 runs," but Table II reports single numbers without standard deviations or confidence intervals. The reported differences are small: for example, the Gaussian AUC is 0.892 and the best quantum AUC is 0.904, and Figure 10 reports a 2.44% improvement. Without error bars or repeated data-split estimates, it is impossible to assess whether these gaps are meaningful or within sampling noise. I ask the authors to report means and standard deviations across multiple random splits and, ideally, a paired comparison between quantum and Gaussian kernels on the same splits.
  3. [Abstract and Conclusion; Appendix A (C2.1)] The claim "first evidence of quantum model performance at scale" overreaches the experimental design. The experiments show that one specific MPS-simulated quantum kernel combined with an SVM improves on one financial dataset under hyperparameters selected on the test set. Even after the validation-set reanalysis requested above, this would be evidence about this quantum kernel implementation on this dataset, not a general statement about "quantum model performance." The manuscript should temper the abstract and conclusion accordingly, for instance by saying "these results provide initial evidence that this MPS-simulated quantum kernel can scale to 165 features and 6400 points" and by acknowledging that the comparison is to a single Gaussian kernel baseline with a fixed bandwidth rule.
  4. [Section II-A] The feature map itself is not prespecified: the text states that "the Pauli operators have been chosen empirically based on model performance." This is another, earlier level of model selection on the same data, and it compounds the test-set C-selection problem. For the performance claims to be clean, the authors should either specify the full model-selection protocol (including how the Pauli operators and feature map hyperparameters were chosen) or present the results as an exploratory study that requires confirmation on held-out data.
minor comments (7)
  1. [Title page] The affiliation line contains a typo: "V entures" should read "Ventures."
  2. [Section I] "state of art" should be "state of the art" (or "state-of-the-art").
  3. [Section III-A] "embarrasingly parallel" should be "embarrassingly parallel."
  4. [Figure 10 caption] The caption reads "sample sizes. on the Elliptic Bitcoin data set" with an extra period; it should read "sample sizes on the Elliptic Bitcoin data set."
  5. [Section III-B] The phrase "down selected and seeded to a specified dimension" would be clearer as "down-selected and seeded to a specified size," and the hyphen should be used consistently.
  6. [Table I] The table caption says "Average of the largest bond dimension" but the columns are labeled χ; please clarify in the caption or table that the reported values are the average over data points of the maximum virtual bond dimension after simulation.
  7. [Appendix G/H] The artifact descriptions are very thorough, which is commendable; however, the descriptions of the regularization selection should be highlighted in the main text, because they determine the interpretation of the headline performance claims.

Circularity Check

1 steps flagged · score 6.0 of 10

Test-set selection of the SVM regularization coefficient makes the headline performance claims (improvement with features/data; quantum beats Gaussian) constructed from the test set rather than predicted.

  1. fitted input called prediction [Appendix G (Artifact A5, Figs. 9-10); Appendix H (Artifact A6, Table II); Section III-B]
    "The script plot.py gathers the AUC metrics and, for each experiment, it picks the best AUC score among the different possible choices of the regularization coefficient; ... it picks the regularization coefficient with highest AUC and reports the average score of each metric for said regularization coefficient, which are displayed in Table II."

    The reported test AUCs are not the predictive performance of a model with a pre-specified regularization parameter; by the artifact's own description, the scripts select the regularization coefficient C that maximizes the test AUC for each experiment and then report that maximum as the model's score. The test labels are therefore used twice: once to choose C, and once to compute the reported metric. Consequently the claimed improvement with features and data size (C2.1) and the quantum-over-Gaussian result (C2.2) are, by construction, comparisons of test-set-fitted maxima rather than out-of-sample predictions.

full rationale

The paper's core engineering contributions (C1.1-C1.3) are self-contained empirical scaling studies: the MPS memory/runtime behavior is measured with public artifacts A1-A4 and does not rest on any fitted parameter or self-citation. No uniqueness theorem or ansatz is imported from the authors' prior work; citations to ITensors, cuTensorNet, and the QML literature are external and not load-bearing in a circular way. The circularity is confined to the model-performance claims. Appendices G and H state explicitly that the regularization coefficient is selected by taking the highest test AUC, and the figures and table then report those selected values. This matches the 'fitted input called prediction' pattern: the reported test-set AUC is, by construction, the maximum of the AUCs over the C grid. Thus C2.1 and C2.2 are partially constructed from the test set rather than independently predicted. The comparison with the Gaussian kernel uses the same selection workflow, so the bias is symmetric, but that does not rescue the absolute scaling claim in Figure 10, which lacks error bars and validation-split separation. The score is 6 rather than higher because the runtime/memory framework contributions and the raw Gram-matrix computations remain genuine measurements, and the selection rule is disclosed in the artifacts rather than hidden.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The central claims rest on standard MPS mathematics, the empirical ansatz choice, and the data preparation procedure. The main ad hoc elements are the truncation threshold and the test-set-based model selection. No new physical entities are introduced.

free parameters (5)
  • kernel bandwidth gamma = 0.1, 0.5, 1.0 in Table II
    Controls the angle of RZ and RXX gates through the Hamiltonian coefficients. The paper states that gamma is needed to scale to larger quantum feature spaces and chooses values empirically.
  • SVM regularization C = selected from [0.01, 4.0] to maximize test AUC
    Chosen per experiment by picking the value with the highest AUC on the test set (Appendix G and H). This is a test-set-based selection and inflates reported performance.
  • interaction distance d = 1, 2, 4, 6 in Table II; up to 12 in Figure 5
    Controls the number of RXX gates and thus the expressivity and bond dimension. Chosen as a hyperparameter to study the effect of circuit complexity.
  • number of layers r = 2, 4, 8, 12, 16, 20 in Table III
    Determines circuit depth. Chosen to study the effect of increasing expressivity on model performance.
  • Gaussian kernel bandwidth alpha = 1/[m * var(X)]
    For the classical baseline, the bandwidth is set by a standard heuristic rather than tuned, which may disadvantage the Gaussian kernel in the comparison.
assumptions (6)
  • standard math SVD truncation error formula (Eq. 8) gives valid bounds for canonical MPS approximations.
    Used to justify truncating singular values with sum of squared values below 1e-16. Based on standard tensor network theory (Refs [29,30]).
  • standard math The inner product of two MPS states computed via the contraction order in Figure 2 is exact up to floating point precision.
    The contraction algorithm is standard for MPS inner products with complexity O(m*chi^3), cited to Ref [31].
  • domain assumption The quantum kernel K_ij = |<psi(x_i), psi(x_j)>|^2 is a valid kernel for an SVM classifier.
    This follows from the quantum feature map construction of Havlicek et al. (Ref [22]); the paper adopts this as a standard approach.
  • domain assumption The empirical choice of Z and XX Pauli operators in the feature map produces a feature space that is expressive enough for the Elliptic dataset.
    The paper states the Pauli operators are 'chosen empirically based on model performance' (Section II-A), so the choice is not derived from first principles.
  • ad hoc to paper Truncation error below 1e-16 is negligible and does not affect the reported classification metrics.
    The paper argues that the truncation error is at the scale of 64-bit floating point precision, but no direct comparison to exact statevector results is provided for the quantum kernel values.
  • domain assumption The downsampled balanced data splits are representative of the Elliptic dataset and do not introduce bias.
    The paper describes data preparation and seeding but does not report sensitivity to different seeds or preprocessing choices.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Realizing Quantum Kernel Models at Scale with Matrix Product State Simulation." pith.science (2026). https://pith.science/paper/HMYMPEP7

@misc{pith2026241109336,
  author       = {Pith},
  title        = {Pith review of: Realizing Quantum Kernel Models at Scale with Matrix Product State Simulation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HMYMPEP7}},
  note         = {Machine review of arXiv:2411.09336}
}
read the original abstract

Data representation in quantum state space offers an alternative function space for machine learning tasks. However, benchmarking these algorithms at a practical scale has been limited by ineffective simulation methods. We develop a quantum kernel framework using a Matrix Product State (MPS) simulator and employ it to perform a classification task with 165 features and 6400 training data points, well beyond the scale of any prior work. We make use of a circuit ansatz on a linear chain of qubits with increasing interaction distance between qubits. We assess the MPS simulator performance on CPUs and GPUs and, by systematically increasing the qubit interaction distance, we identify a crossover point beyond which the GPU implementation runs faster. We show that quantum kernel model performance improves as the feature dimension and training data increases, which is the first evidence of quantum model performance at scale.

Figures

Figures reproduced from arXiv: 2411.09336 by the authors.

Figure 1
Figure 1. Gate application on a state represented as an MPS (turquoise) [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Calculation of an inner product between states [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. The circuit ansatz preparing the state U(x)|+⟩m for each data point x of m features. Circuits have m qubits and a tunable number of layers (r) and interaction distance (d). (a) Overall circuit structure; the parameter r indicates how many copies of e−iHXX(x) · e−iHZ (x) are applied; the angles of the RZ gates correspond to the coefficients in equation (4). (b) The structure of the circuit for e−iHXX(x) on m = 5 qubi… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Distributed computation of the Gram matrix. Two strategies are presented: (a) no-messaging and (b) round-robin. The figure displays the sequence [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Runtime scaling as the qubit interaction distance is increased. CPU corresponds to ITensors, GPU corresponds to our library. Each circuit comprises [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Memory required to store the MPS throughout the simulation of a [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 8
Figure 8. Figure 8: Breakdown of wall-clock runtime to compute the Gram matrix for [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: Training set Area Under the Curve prediction score with 6400 (blue), [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 24 canonical work pages

  1. [20]

    cutn-qsvm: cutensornet-accelerated quantum support vector machine with cuquantum sdk,

    K.-C. Chen, T.-Y . Li, Y .-Y . Wang, S. See, C.-C. Wang, R. Wille, N.-Y . Chen, A.-C. Yang, and C.-Y . Lin, “cutn-qsvm: cutensornet-accelerated quantum support vector machine with cuquantum sdk,” ArXiv, p. 2405.02630, 2024

    arXiv 2024

  2. [1]

    An introduction to quantum machine learning,

    I. S. Maria Schuld and F. Petruccione, “An introduction to quantum machine learning,” Contemporary Physics , vol. 56, no. 2, pp. 172– 185, 2015. [Online]. Available: https://doi.org/10.1080/00107514.2014. 964942

    arXiv 2015

  3. [2]

    Unsupervised quantum machine learning for fraud detection,

    O. Kyriienko and E. B. Magnusson, “Unsupervised quantum machine learning for fraud detection,” 2022

    work page 2022

  4. [3]

    Quantum multiple kernel learning in financial classification tasks,

    S. Miyabe, B. Quanz, N. Shimada, A. Mitra, T. Yamamoto, V . Ras- tunkov, D. Alevras, M. Metcalf, D. J. King, M. Mamouei, M. D. Jackson, M. Brown, P. Intallura, and J.-E. Park, “Quantum multiple kernel learning in financial classification tasks,” 2023

    work page 2023

  5. [4]

    Photovoltaic power forecasting using quantum machine learning,

    A. Sagingalieva, S. Komornyik, A. Senokosov, A. Joshi, A. Sedykh, C. Mansell, O. Tsurkan, K. Pinto, M. Pflitsch, and A. Melnikov, “Photovoltaic power forecasting using quantum machine learning,” 2023

    work page 2023

  6. [5]

    Quantum machine learning for credit scoring,

    N. Schetakis, D. Aghamalyan, M. Boguslavsky, A. Rees, M. Raktoma- lala, and P. Griffin, “Quantum machine learning for credit scoring,” 2023

    work page 2023

  7. [6]

    Quantum machine learning algorithms for drug discovery applications,

    K. Batra, K. M. Zorn, D. H. Foil, E. Minerali, V . O. Gawriljuk, T. R. Lane, and S. Ekins, “Quantum machine learning algorithms for drug discovery applications,” Journal of Chemical Information and Modeling, vol. 61, no. 6, pp. 2641–2647, 2021, pMID: 34032436. [Online]. Available: https://doi.org/10.1021/acs.jcim.1c00166

    doi:10.1021/acs.jcim.1c00166 2021

  8. [7]

    Better than classical? the subtle art of benchmarking quantum machine learning models,

    J. Bowles, S. Ahmed, and M. Schuld, “Better than classical? the subtle art of benchmarking quantum machine learning models,” 2024

    work page 2024

Show all 35 references
  1. [8]

    Numerical evidence against advantage with quantum fidelity kernels on classical data,

    L. Slattery, R. Shaydulin, S. Chakrabarti, M. Pistoia, S. Khairy, and S. M. Wild, “Numerical evidence against advantage with quantum fidelity kernels on classical data,” Phys. Rev. A, vol. 107, p. 062417, Jun

  2. [9]

    A hyperparameter study for quantum kernel methods,

    S. Egginger, A. Sakhnenko, and J. M. Lorenz, “A hyperparameter study for quantum kernel methods,” 2023

    2023

  3. [10]

    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,” Phys. Rev. A , vol. 103, p. 032430, Mar 2021. [Online]. Available: https://link.aps.org/doi/10.1103/PhysRevA.103.032430

    2021 doi

  4. [11]

    Generalization despite overfitting in quantum machine learning models,

    E. Peters and M. Schuld, “Generalization despite overfitting in quantum machine learning models,” Quantum, vol. 7, p. 1210, Dec. 2023. [Online]. Available: https://doi.org/10.22331/q-2023-12-20-1210

    2023 doi

  5. [12]

    Power of data in quantum machine learning,

    H.-Y . Huang, M. Broughton, M. Mohseni, R. Babbush, S. Boixo, H. Neven, and J. R. McClean, “Power of data in quantum machine learning,” Nat Commun , vol. 12, p. 2631, 2021. [Online]. Available: https://doi.org/10.1038/s41467-021-22539-9

    2021 doi

  6. [13]

    Understanding quantum machine learning also requires rethinking generalization,

    E. Gil-Fuster, J. Eisert, and C. Bravo-Prieto, “Understanding quantum machine learning also requires rethinking generalization,” Nat Commun, vol. 15, p. 2277, 2024

    2024

  7. [14]

    Noise-induced barren plateaus in variational quantum algorithms,

    S. Wang, E. Fontana, M. Cerezo, K. Sharma, A. Sone, L. Cincio, and P. J. Coles, “Noise-induced barren plateaus in variational quantum algorithms,” Nat Commun , vol. 12, p. 6961, 2021

    2021

  8. [15]

    Exponential concentration and untrainability in quantum kernel methods,

    S. Thanasilp, S. Wang, M. Cerezo, and Z. Holmes, “Exponential concentration and untrainability in quantum kernel methods,” Nature Communications, vol. 15, no. 5200, 2024

    2024

  9. [16]

    Hyper-optimized tensor network contraction,

    J. Gray and S. Kourtis, “Hyper-optimized tensor network contraction,” Quantum, vol. 5, p. 410, Mar. 2021. [Online]. Available: https: //doi.org/10.22331/q-2021-03-15-410

    2021 doi

  10. [17]

    Efficient tensor network simulation of ibm’s eagle kicked ising experiment,

    J. Tindall, M. Fishman, E. M. Stoudenmire, and D. Sels, “Efficient tensor network simulation of ibm’s eagle kicked ising experiment,” PRX Quantum , vol. 5, p. 010308, Jan 2024. [Online]. Available: https://link.aps.org/doi/10.1103/PRXQuantum.5.010308

    2024 doi

  11. [18]

    Fast and converged classical simulations of evidence for the utility of quantum computing before fault tolerance,

    T. Begu ˇsi´c, J. Gray, and G. K.-L. Chan, “Fast and converged classical simulations of evidence for the utility of quantum computing before fault tolerance,” Science Advances , vol. 10, no. 3, p. eadk4321,

  12. [19]

    Density-matrix renormalization group algorithm for simulating quantum circuits with a finite fidelity,

    T. Ayral, T. Louvet, Y . Zhou, C. Lambert, E. M. Stoudenmire, and X. Waintal, “Density-matrix renormalization group algorithm for simulating quantum circuits with a finite fidelity,” PRX Quantum , vol. 4, p. 020304, Apr 2023. [Online]. Available: https://link.aps.org/ doi/10.1...

    2023 doi

  13. [21]

    Hyperoptimized approximate contraction of tensor networks with arbitrary geometry,

    J. Gray and G. K.-L. Chan, “Hyperoptimized approximate contraction of tensor networks with arbitrary geometry,” Phys. Rev. X , vol. 14, p. 011009, Jan 2024. [Online]. Available: https://link.aps.org/doi/10.1103/ PhysRevX.14.011009

    2024

  14. [22]

    Supervised learning with quantum- enhanced feature spaces,

    V . Havl ´ıˇcek, A. D. C `orcoles, K. Temme, A. K. Aram W. Harrow, J. M. Chow, and J. M. Gambetta, “Supervised learning with quantum- enhanced feature spaces,” Nature, vol. 567, pp. 209–212, 2019

    2019

  15. [23]

    Public repository of pytket-cutensornet

    “Public repository of pytket-cutensornet.” [Online]. Available: https: //github.com/CQCL/pytket-cutensornet

  16. [24]

    The ITensor Software Library for Tensor Network Calculations,

    M. Fishman, S. R. White, and E. M. Stoudenmire, “The ITensor Software Library for Tensor Network Calculations,” SciPost Phys. Codebases, p. 4, 2022. [Online]. Available: https://scipost.org/10.21468/ SciPostPhysCodeb.4

    2022

  17. [25]

    cuquantum sdk: A high-performance library for accelerating quantum science,

    H. Bayraktar, A. Charara, D. Clark, S. Cohen, T. Costa, Y .-L. L. Fang, Y . Gao, J. Guan, J. Gunnels, A. Haidar, A. Hehn, M. Hohnerbach, M. Jones, T. Lubowe, D. Lyakh, S. Morino, P. Springer, S. Stanwyck, I. Terentyev, S. Varadhan, J. Wong, and T. Yamaguchi, “cuquantum sdk: A ...

    2023

  18. [26]

    Importance of kernel bandwidth in quantum machine learning,

    R. Shaydulin and S. M. Wild, “Importance of kernel bandwidth in quantum machine learning,” Phys. Rev. A , vol. 106, no. 042407, 2022

    2022

  19. [27]

    Bandwidth enables generalization in quantum kernel models,

    A. Canatar, C. P. Evan Peters, S. M. Wild, and R. Shaydulin, “Bandwidth enables generalization in quantum kernel models,” 2022

    2022

  20. [28]

    Large-scale simulation of shor’s quantum factoring algorithm,

    D. Willsch, M. Willsch, F. Jin, H. De Raedt, and K. Michielsen, “Large-scale simulation of shor’s quantum factoring algorithm,” Mathematics, vol. 11, no. 19, 2023. [Online]. Available: https: //www.mdpi.com/2227-7390/11/19/4222

    2023

  21. [29]

    What limits the simulation of quantum computers?

    Y . Zhou, E. M. Stoudenmire, and X. Waintal, “What limits the simulation of quantum computers?” Phys. Rev. X , vol. 10, p. 041038, Nov 2020. [Online]. Available: https://link.aps.org/doi/10.1103/ PhysRevX.10.041038

    2020

  22. [30]

    Time-evolution methods for matrix-product states,

    S. Paeckel, T. K ¨ohler, A. Swoboda, S. R. Manmana, U. Schollw ¨ock, and C. Hubig, “Time-evolution methods for matrix-product states,” Annals of Physics , vol. 411, p. 167998, 2019. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0003491619302532

    2019

  23. [31]

    A practical introduction to tensor networks: Matrix product states and projected entangled pair states,

    R. Or ´us, “A practical introduction to tensor networks: Matrix product states and projected entangled pair states,” Annals of Physics, vol. 349, pp. 117–158, 2014. [Online]. Available: https: //www.sciencedirect.com/science/article/pii/S0003491614001596

    2014

  24. [32]

    Elliptic dataset from kaggle

    “Elliptic dataset from kaggle.” [Online]. Available: https://www.kaggle. com/datasets/ellipticco/elliptic-data-set

  25. [33]

    Receiver operating characteristic (roc) curves,

    T. Gneiting and P. V ogels, “Receiver operating characteristic (roc) curves,” 2018. APPENDIX ARTIFACT DESCRIPION A. Paper’s Main Contributions The contributions of our paper can be classified in two groups. The first group of contributions ( C1) refers to the development and b...

    2018

  26. [2023]

    Available: https://link.aps.org/doi/10.1103/PhysRevA

    [Online]. Available: https://link.aps.org/doi/10.1103/PhysRevA. 107.062417

    doi

  27. [2024]

    Available: https://www.science.org/doi/abs/10.1126/ sciadv.adk4321

    [Online]. Available: https://www.science.org/doi/abs/10.1126/ sciadv.adk4321

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.