Pith. sign in

REVIEW 4 major objections 5 minor 115 references

Addressing the Readout Problem in Quantum Differential Equation Algorithms with Quantum Scientific Machine Learning

T0 review · 4 major / 5 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 Quantum PDE readout can be replaced by trained quantum measurements.

desk verdict A plausible proof of concept for QNN-based readout of PDE solutions, but the readout claim rests on an untested equivalence to real quantum solver outputs. read the letter →

arxiv 2411.14259 v1 pith:BNK42TQS submitted 2024-11-21 quant-ph cond-mat.dis-nnphysics.flu-dyn

Chelsea A. Williams , Stefano Scali , Antonio A. Gentile , Daniel Berger , Oleksandr Kyriienko This is my paper · ORCID
classification quant-phcond-mat.dis-nnphysics.flu-dyn
keywords quantummachinelearningreadoutproblemPDEsolversconvolutionalneuralnetworksBurgersequationNavier-Stokesequationsbasischoicedata
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 differential equation solvers aim to store solutions as $n$-qubit states over $O(2^n)$ grid points, but reading those states out via tomography is exponentially expensive. The paper claims this readout problem can be bypassed when the goal is classification: treat the solver output as quantum data and train a quantum neural network to act as a measurement operator that labels the state directly. It demonstrates the idea on two computational fluid dynamics tasks, classifying Burgers-equation wave solutions as shock or steady and Navier-Stokes cylinder flows as laminar or turbulent. The best Burgers model, a Fourier-basis deep quantum neural network with coarse-graining, reaches 100% accuracy, while the double-QCNN separates laminar from turbulent flow with over 90% accuracy. The paper also shows that the basis in which quantum data is analyzed strongly affects accuracy, making basis choice a central design decision.

What carries the argument

The load-bearing mechanism is a trainable measurement operator built from a quantum neural network. A parameterized circuit $U(\theta)$ acts on the amplitude-encoded PDE solution $\rho$, and an expectation value of a fixed low-dimensional observable produces the label; for the Burgers problem this is the Pauli-$Z$ observable after a depth-4 DQNN, and for Navier-Stokes it is the covariance $\mathrm{cov}(O_1,O_2)$ between measurements on two register halves after a double-QCNN. Two supporting tools do essential work: the quantum Fourier transform changes the analysis basis from real space to momentum space with an $O(n^2)$ circuit, and coarse-graining discards part of the register to concentrate on macroscopic features while reducing the qubit count. The basis choice interacts strongly with the architecture, which is why the same model can vary from 27% to 100% accuracy depending on whether Fourier preprocessing and coarse-graining are applied.

What would settle it

Run the trained Burgers DQNN and Navier-Stokes double-QCNN on states produced by an actual noiseless quantum PDE solver, or by its most faithful simulation, rather than on classically computed amplitude-encoded states; if classification accuracy drops materially below the reported 100% and 92%, the claim that this workflow solves the readout problem for solver output is falsified.

Watch

Extended reading notes

Core claim

The central claim is that the readout bottleneck of quantum PDE algorithms can be addressed by treating solver outputs as quantum data and learning a hypothesis that is itself a measurement operator. Formally, the solver produces states $\rho$ in a Hilbert space, and the learned hypothesis $h(\rho,\theta)=\langle\rho| U^{\dagger}(\theta) M U(\theta)|\rho\rangle$ maps each state to a classification label through a parameterized circuit $U(\theta)$ and a low-dimensional observable $M$. For Burgers shock detection, a deep quantum neural network of depth 4 with Pauli-$Z$ readout achieves 100% accuracy when the data is Fourier-transformed and coarse-grained to half the register; real-space encoding reaches only 27% and Fourier without coarse-graining 45%. For Navier-Stokes flow around a cylinder, a double-QCNN applies separate convolutional networks to the $x$ and $y$ registers and uses the covariance of Pauli-$X$ and Pauli-$Z$ across registers as the hypothesis, reaching 92% average accuracy in real space and 87% in Fourier space with lower variance. The paper reads these results as establishing quantum scientific machine learning as a necessary quantum post-processing step: low-dimensional physical features can be distilled from quantum states without tomographic reconstruction, provided the analysis basis is matched to the problem.

Load-bearing premise

The demonstrations assume that classically computed PDE solutions, amplitude-encoded into quantum states and augmented with white noise, exactly match the states a noiseless quantum differential equation solver would produce; if actual solver states differ in structure, entanglement, or noise, the trained measurement operators may not transfer and the readout advantage would not hold.

Editorial extensions

If this is right

  • Quantum PDE solvers can output physical labels such as shock versus steady or laminar versus turbulent through a small number of measurements instead of full tomography, keeping useful answers accessible as the grid grows.
  • The analysis basis is a decisive hyperparameter: for Burgers solutions, Fourier basis plus coarse-graining is needed for 100% accuracy, while real-space encoding fails on the same task.
  • Coarse-graining can reduce the register size and improve generalization, with the Fourier Navier-Stokes model performing better at small training-sample sizes.
  • Treating solver outputs as quantum data opens a workflow in which quantum machine learning is the readout stage of quantum scientific computing, rather than an optional add-on.

Reading between the lines

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

  • If actual noise-bearing quantum PDE solver outputs differ from the amplitude-encoded classical solutions used here in entanglement structure, amplitude distribution, or device noise, the trained measurement operators may need to be retrained on solver-native states; direct transfer is not guaranteed by the paper.
  • The basis-dependence result suggests a design principle beyond the two examples: choose an analysis basis matched to the symmetries of the target feature, such as momentum space for translationally invariant structures, which could reduce sample complexity in other PDE classification tasks.
  • The same concept-learning readout could apply to other low-dimensional diagnostics from PDE states, including instability detection, phase-boundary identification, or vortex detection, whenever the desired output is a label rather than the full field.
  • Combining this readout with shadow-tomography or classical-simulation bounds could reveal when quantum classification is genuinely needed rather than reproducible by classical post-processing of classically computed solutions.
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 / 5 minor

Summary. The manuscript proposes a quantum scientific machine learning (QuaSciML) pipeline to address the readout problem for quantum differential equation solvers. It treats solver outputs as quantum data and trains parametrized quantum circuits—a deep QNN for Burgers shock detection and a double-QCNN for Navier-Stokes laminar versus turbulent classification—to act as problem-specific measurement operators that output binary labels directly from quantum states. Training is performed on amplitude-encoded classical simulations of Burgers and Navier-Stokes solutions, with variants using real-space and Fourier bases and with or without coarse-graining. The paper reports 100% accuracy for shock detection with the Fourier plus coarse-grained model and about 92% and 87% accuracy for turbulence classification in real and Fourier bases, respectively, and concludes that learned measurement operators can bypass tomography-based readout.

Significance. If the central equivalence assumption held, the paper would be a useful proof-of-concept that quantum machine learning can serve as a readout layer for quantum PDE solvers, reducing high-dimensional quantum states to low-dimensional classification labels without full tomography. The framing of quantum data as a resource, the emphasis on basis choice and coarse-graining, and the physics-informed covariance hypothesis are valuable ideas. The paper also draws on a relevant body of literature on QCNNs and quantum data learning. However, the demonstration is a classical simulation study with an asserted rather than verified connection to actual quantum solver outputs, and the statistical support for the headline results is limited. The strengths are conceptual and methodological; the experimental evidence is not yet sufficient for the central claim as stated.

major comments (4)
  1. [Results, paragraph beginning 'For simplicity...'] The central claim rests on the assertion, 'The results from the analysis presented here are the same as if the quantum data inputs were measured directly from a noiseless quantum PDE solver without any classical processing.' This equivalence is not demonstrated. For the nonlinear PDEs considered, standard quantum algorithms (e.g., Carleman embedding, Schrodingerisation) prepare states in enlarged Hilbert spaces with truncation errors and normalization effects, which need not coincide with the simple amplitude-encoded grid functions used here. If the actual solver states differ in structure or entanglement, the trained measurement operators may not transfer, and the claimed solution to the readout problem would fail in the intended application. Please either run a concrete quantum PDE solver on small instances and test the trained models on its output states, or explicitly restrict the claim to amplitude-encoded quantum states and justify why this restricted setting is representative.
  2. [Fig. 2(c,d) and surrounding text] The 100% shock-detection accuracy is obtained from only 22 solutions (11 per class), and the optimal model is selected among the three variants (real, Fourier, Fourier plus coarse-graining) after examining the full-data accuracy curves. The accuracy plotted in Fig. 2(c) is evaluated on the entire data set after training on a subset, which includes training points, so it is not a held-out generalization measure; the confusion matrix in Fig. 2(d) uses an 11-sample test set. With this sample size, 100% accuracy has very wide confidence intervals, and post-hoc model selection without a separate validation set inflates the reported performance. Please provide confidence intervals, a nested cross-validation or a pre-specified model selection rule, and a larger number of independent solutions.
  3. [Results (Navier-Stokes) and Fig. 4] No classical baseline is provided for either benchmark. A classical machine learning classifier operating on the same truncated and subsampled features, or on the amplitude vectors, is needed to determine whether the quantum measurement layer is actually necessary or whether the tasks become trivially separable after the coarse-graining and late-time truncation. Without such a baseline, the claim that QuaSciML 'addresses the readout problem' is not established beyond showing that a particular quantum circuit can fit this dataset.
  4. [Hypothesis definitions and Eq. (4)] Training and evaluation use exact statevector expectation values, so the measurement overhead required to estimate h(ρ,θ) and cov(O1,O2) on actual hardware is not analyzed. Since the readout problem is specifically about the cost of extracting information from quantum states, an analysis of the number of shots needed to estimate these expectation values to the accuracy required for the reported classification results is essential. Without it, the proposal may simply shift the tomographic cost into repeated measurements of a covariance hypothesis.
minor comments (5)
  1. [Fig. 2 caption and Burgers results text] The text states that Burgers solutions are encoded into 12-qubit states, while the Fig. 2 caption says 6-qubit quantum states; please reconcile these values.
  2. [General notation] The terms 'double-QNN' and 'double-QCNN' are used interchangeably; please choose one consistent name for the two-register architecture.
  3. [Fig. 4(b-e)] Fig. 4(b,d) report results over four random seeds, while Fig. 4(c,e) report averages over ten seeds; please make the number of seeds consistent or explain the discrepancy.
  4. [Reference [96]] Reference [96] appears to lack author information and a full citation; please complete it.
  5. [Fig. 2(a-b)] The color legend for steady waves versus shock waves is not explicitly defined in the caption; please add a clear legend.

Circularity Check

1 steps flagged · score 4.0 of 10

Central 'quantum data' bridge is assumed by definition: classical amplitude-encoded PDE solutions are declared identical to noiseless quantum solver outputs, making the readout claim depend on the premise.

  1. self definitional [Results, opening paragraph before the Burgers equation subsection]
    "For simplicity in exhibiting how the quantum readout classification is performed, the solutions used for training and testing are obtained from classical simulations. These solutions are converted into quantum states and used as inputs to the QuaSciML models. The results from the analysis presented here are the same as if the quantum data inputs were measured directly from a noiseless quantum PDE solver without any classical processing. Therefore, the quantum data ρ represent the vectors encoding the quantum state PDE solutions in a given basis."

    The paper's headline claim is that QuaSciML can address the readout problem 'in scenarios where data samples come directly from quantum differential equation solvers.' The demonstration never runs a quantum PDE solver: the inputs are classical Burgers and Navier-Stokes solutions converted into amplitude-encoded states. The quoted sentence asserts, without evidence, that these states are identical to noiseless quantum solver outputs, and then uses 'Therefore' to redefine the classical states as 'the quantum state PDE solutions.' The applicability of the trained measurement operators to actual quantum solver outputs is therefore not a derived result; it is assumed by this definitional identification.

full rationale

The paper is otherwise not circular. Training follows standard supervised learning: labels are used to fit QNN parameters, and the models are evaluated on held-out splits, so no fitted parameter is renamed as a prediction and no test-set quantity reduces to a training quantity. The self-citations (e.g., Refs. [44], [69], [73], [92], [95], [103]) are contextual and methodological; none is used to forbid alternatives or to import a uniqueness theorem that forces the chosen model. The QCNN and DQNN architectures are cited from the external literature, not from an unverified self-citation chain. The one genuinely circular element is the definitional identification of amplitude-encoded classical simulation data with quantum PDE solver outputs. Because the intended application is readout of quantum solvers, and the demonstration substitutes classical data while asserting exact equivalence, the central transfer claim reduces to that assertion. Still, the learning results themselves are independent and non-trivial, so the circularity is partial rather than total.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The paper's load-bearing assumptions are: (1) classical amplitude-encoded solutions equal noiseless quantum solver outputs, (2) the classification labels are learnable, and (3) ideal quantum primitives. These are stated but not verified. No code or data is shipped.

free parameters (2)
  • Variational parameters theta (DQNN and double-QCNN)
    Optimized by gradient descent on the training loss; values not reported in the paper.
  • Activation scaling parameters a and b
    Used in tanh[a(|f|-b)] to shift and scale input data for the double-QCNN; unclear whether trained or hand-chosen, no values given.
assumptions (3)
  • domain assumption Amplitude-encoded classical PDE solutions faithfully represent outputs of a noiseless quantum PDE solver
    Explicitly stated in Results; the entire readout-problem claim depends on this equivalence.
  • domain assumption The labels (shock vs steady, laminar vs turbulent) are well-defined and learnable from the quantum state features
    The paper assumes the chosen classification tasks are meaningful and that the QNN can represent the decision boundary.
  • standard math Standard quantum computing primitives (amplitude encoding, QFT, Pauli measurements) behave ideally
    Assumed throughout; no noise model is considered.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Addressing the Readout Problem in Quantum Differential Equation Algorithms with Quantum Scientific Machine Learning." pith.science (2026). https://pith.science/paper/BNK42TQS

@misc{pith2026241114259,
  author       = {Pith},
  title        = {Pith review of: Addressing the Readout Problem in Quantum Differential Equation Algorithms with Quantum Scientific Machine Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BNK42TQS}},
  note         = {Machine review of arXiv:2411.14259}
}
abstract

Quantum differential equation solvers aim to prepare solutions as $n$-qubit quantum states over a fine grid of $O(2^n)$ points, surpassing the linear scaling of classical solvers. However, unlike classically stored vectors of solutions, the readout of exact quantum states poses a bottleneck due to the complexity of tomography. Here, we show that the readout problem can be addressed with quantum learning tools where we focus on distilling the relevant features. Treating outputs of quantum differential equation solvers as quantum data, we demonstrate that low-dimensional output can be extracted using a measurement operator adapted to detect relevant features. We apply this quantum scientific machine learning approach to classify solutions for shock wave detection and turbulence modeling in scenarios where data samples come directly from quantum differential equation solvers. We show that the basis chosen for performing analysis greatly impacts classification accuracy. Our work opens up the area of research where quantum machine learning for quantum datasets is inherently required.

Figures

Figures reproduced from arXiv: 2411.14259 by the authors.

Figure 1
Figure 1. FIG. 1. A quantum CFD pipeline that uses a combination of physics-based modeling (stress, pressure, etc.) and ML to predict the behavior [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. The classification of wave solutions with a DQNN. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. The encoding of quantum data into a double-QCNN cir [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: FIG. 4. The classification of flow solutions with a double-QCNN. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

115 extracted references · 47 canonical work pages

  1. [1]

    Montanaro, npj Quantum Information 2, 15023 (2016)

    A. Montanaro, npj Quantum Information 2, 15023 (2016)

    2016

  2. [2]

    A. M. Dalzell, S. McArdle, M. Berta, P. Bienias, C.-F. Chen, A. Gily´en, C. T. Hann, M. J. Kastoryano, E. T. Khabiboulline, A. Kubica, G. Salton, S. Wang, and F. G. S. L. Brand˜ao, Quan- tum algorithms: A survey of applications and end-to-end com- plexities (2023), arXiv:2310.03011 [quant-ph]

    arXiv 2023

  3. [3]

    Babbush, N

    R. Babbush, N. Wiebe, J. McClean, J. McClain, H. Neven, and G. K.-L. Chan, Phys. Rev. X 8, 011044 (2018)

    2018

  4. [4]

    K. J. Satzinger, Y .-J. Liu, A. Smith, C. Knapp, M. Newman, C. Jones, Z. Chen, C. Quintana, X. Mi, A. Dunsworth, C. Gidney, I. Aleiner, F. Arute, K. Arya, J. Atalaya, et al. , Science 374, 1237 (2021), https://www.science.org/doi/pdf/10.1126/science.abi8378

    doi:10.1126/science.abi8378 2021

  5. [5]

    Clinton, T

    L. Clinton, T. Cubitt, B. Flynn, F. M. Gambetta, J. Klassen, A. Montanaro, S. Piddock, R. A. Santos, and E. Sheridan, Na- ture Communications 15, 211 (2024)

    2024

  6. [6]

    Alexeev et al., Future Generation Computer Systems 160, 666 (2024)

    Y . Alexeev et al., Future Generation Computer Systems 160, 666 (2024)

    2024

  7. [7]

    Di Meglio, K

    A. Di Meglio, K. Jansen, I. Tavernelli, C. Alexandrou, S. Arunachalam, C. W. Bauer, K. Borras, S. Carrazza, A. Crippa, V . Croft, R. de Putter, A. Delgado, V . Dunjko, D. J. Egger, E. Fern ´andez-Combarro, et al. , PRX Quantum 5, 037001 (2024)

    2024

  8. [8]

    J. J. Mart ´ınez de Lejarza, L. Cieri, M. Grossi, S. Vallecorsa, and G. Rodrigo, Phys. Rev. D 110, 074031 (2024)

    2024

Show all 115 references
  1. [9]

    McArdle, S

    S. McArdle, S. Endo, A. Aspuru-Guzik, S. C. Benjamin, and X. Yuan, Rev. Mod. Phys.92, 015003 (2020)

    2020

  2. [10]

    Y . Cao, J. Romero, J. P. Olson, M. Degroote, P. D. Johnson, M. Kieferov´a, I. D. Kivlichan, T. Menke, B. Peropadre, N. P. Sawaya, et al., Chemical reviews 119, 10856 (2019)

    2019

  3. [11]

    J. Lee, D. W. Berry, C. Gidney, W. J. Huggins, J. R. McClean, N. Wiebe, and R. Babbush, PRX Quantum 2, 030305 (2021)

    2021

  4. [12]

    Robledo-Moreno, M

    J. Robledo-Moreno, M. Motta, H. Haas, A. Javadi-Abhari, P. Jurcevic, W. Kirby, S. Martiel, K. Sharma, S. Sharma, T. Shirakawa, I. Sitdikov, R.-Y . Sun, K. J. Sung, M. Takita, M. C. Tran, et al. , Chemistry beyond exact solutions on a quantum-centric supercomputer (2024), arXiv...

    2024 arXiv

  5. [13]

    Santagati, A

    R. Santagati, A. Aspuru-Guzik, R. Babbush, M. Degroote, L. Gonz´alez, E. Kyoseva, N. Moll, M. Oppel, R. M. Parrish, N. C. Rubin, M. Streif, C. S. Tautermann, H. Weiss, N. Wiebe, and C. Utschig-Utschig, Nature Physics 20, 549 (2024)

    2024

  6. [14]

    Y . Zhou, J. Chen, J. Cheng, G. Karemore, M. Zitnik, F. T. Chong, J. Liu, T. Fu, and Z. Liang, Quantum-machine- assisted drug discovery: Survey and perspective (2024), arXiv:2408.13479 [quant-ph]

    2024

  7. [15]

    Herman, C

    D. Herman, C. Googin, X. Liu, Y . Sun, A. Galda, I. Safro, M. Pistoia, and Y . Alexeev, Nature Reviews Physics 5, 450 (2023)

    2023

  8. [16]

    Mironowicz, A

    P. Mironowicz, A. S. H., A. Mandarino, A. E. Yilmaz, and T. Ankenbrand, Applications of quantum machine learning for quantitative finance (2024), arXiv:2405.10119 [quant-ph]

    2024 arXiv

  9. [17]

    have also attracted attention. While in each case there are hard instances to calculate, the largest workload of high- performance computing comes from solving differential equa- tions (computational fluid dynamics and alike), reaching ex- ascale [18–20]. This makes scientific...

    2024 arXiv

  10. [18]

    Abbas, A

    A. Abbas, A. Ambainis, B. Augustino, A. B ¨artschi, H. Buhrman, C. Coffrin, G. Cortiana, V . Dunjko, D. J. Egger, B. G. Elmegreen, N. Franco, F. Fratini, B. Fuller, J. Gacon, C. Gonciulea, et al., Nature Reviews Physics 10.1038/s42254- 024-00770-9 (2024)

    2024 doi

  11. [19]

    Heldens, P

    S. Heldens, P. Hijma, B. V . Werkhoven, J. Maassen, A. S. Z. Belloum, and R. V . Van Nieuwpoort, ACM Comput. Surv.53, 10.1145/3372390 (2020)

    2020 doi

  12. [20]

    Mullowney, R

    P. Mullowney, R. Li, S. Thomas, S. Ananthan, A. Sharma, J. S. Rood, A. B. Williams, and M. A. Sprague, in Proceedings of the International Conference for High Performance Comput- ing, Networking, Storage and Analysis , SC ’21 (Association for Computing Machinery, New York, NY ...

    2021

  13. [21]

    C. Coti, Y . Pfau-Kempf, M. Battarbee, U. Ganse, S. Shende, K. Huck, J. Rodriquez, L. Kotipalo, J. Faj, J. J. Williams, I. Peng, A. D. Malony, S. Markidis, and M. Palmroth, in 2024 IEEE International Parallel and Distributed Processing Sym- posium Workshops (IPDPSW) (2024) pp....

    2024

  14. [22]

    M ¨oller, Quantum scientific computing, in Impact of Sci- entific Computing on Science and Society , edited by P

    M. M ¨oller, Quantum scientific computing, in Impact of Sci- entific Computing on Science and Society , edited by P. Neit- taanm¨aki and M.-L. Rantalainen (Springer International Pub- lishing, Cham, 2023) pp. 357–371

    2023

  15. [23]

    A. W. Harrow, A. Hassidim, and S. Lloyd, Physical review letters 103, 150502 (2009)

    2009

  16. [24]

    Martin, R

    A. Martin, R. Ibarrondo, and M. Sanz, Phys. Rev. Appl. 19, 064056 (2023)

    2023

  17. [25]

    D. W. Berry, Journal of Physics A: Mathematical and Theoret- ical 47, 105301 (2014)

    2014

  18. [26]

    Montanaro and S

    A. Montanaro and S. Pallister, Phys. Rev. A 93, 032324 (2016)

    2016

  19. [27]

    P. C. S. Costa, S. Jordan, and A. Ostrander, Phys. Rev. A 99, 012323 (2019)

    2019

  20. [28]

    Scherer, B

    A. Scherer, B. Valiron, S.-C. Mau, S. Alexander, E. van den Berg, and T. E. Chapuran, Quantum Information Processing 16, 60 (2017). 7

    2017

  21. [29]

    Krovi, Quantum 7, 913 (2023)

    H. Krovi, Quantum 7, 913 (2023)

    2023

  22. [30]

    Daribayev, A

    B. Daribayev, A. Mukhanbet, and T. Imankulov, Applied Sci- ences 13, 10.3390/app132011491 (2023)

    2023 doi

  23. [31]

    A. M. Childs, R. Kothari, and R. D. Somma, SIAM Journal on Computing 46, 1920 (2017)

    2017

  24. [32]

    A. M. Childs and N. Wiebe, Quantum Info. Comput. 12, 901–924 (2012)

    2012

  25. [33]

    Lin and Y

    L. Lin and Y . Tong, Quantum4, 361 (2020)

    2020

  26. [34]

    J. M. Martyn, Z. M. Rossi, A. K. Tan, and I. L. Chuang, PRX Quantum 2, 040203 (2021)

    2021

  27. [35]

    Subas ¸ı, R

    Y . Subas ¸ı, R. D. Somma, and D. Orsucci, Physical Review Letters 122, 060504 (2019)

    2019

  28. [36]

    P. C. S. Costa, P. Schleich, M. E. S. Morales, and D. W. Berry, Further improving quantum algorithms for nonlinear differen- tial equations via higher-order methods and rescaling (2023), arXiv:2312.09518 [quant-ph]

    2023 arXiv

  29. [37]

    Jennings, M

    D. Jennings, M. Lostaglio, S. Pallister, A. T. Sornborger, and Y . Subas ¸ı, Efficient quantum linear solver algorithm with de- tailed running costs (2023), arXiv:2305.11352 [quant-ph]

    2023

  30. [38]

    Succi, F

    S. Succi, F. Fillion-Gourdeau, and S. Palpacelli, EPJ Quantum Technology 2, 12 (2015)

    2015

  31. [39]

    Steijl, Applied Sciences 13 (2023)

    R. Steijl, Applied Sciences 13 (2023)

    2023

  32. [40]

    M. A. Schalkers and M. M ¨oller, E fficient and fail- safe collisionless quantum boltzmann method (2022), arXiv:2211.14269 [quant-ph]

    2022 arXiv

  33. [41]

    M. A. Schalkers and M. M ¨oller, On the importance of data encoding in quantum boltzmann methods (2023), arXiv:2302.05305 [quant-ph]

    2023 arXiv

  34. [42]

    Succi, W

    S. Succi, W. Itani, K. Sreenivasan, and R. Steijl, Europhysics Letters 144, 10001 (2023)

    2023

  35. [43]

    Succi, W

    S. Succi, W. Itani, C. Sanavio, K. R. Sreenivasan, and R. Steijl, Computers & Fluids 270, 106148 (2024)

    2024

  36. [44]

    Shao and H

    C. Shao and H. Xiang, Physical Review A 101 (2020)

    2020

  37. [45]

    C. A. Williams, A. A. Gentile, V . E. Elfving, D. Berger, and O. Kyriienko, Quantum iterative methods for solving differen- tial equations with application to computational fluid dynam- ics (2024), arXiv:2404.08605 [quant-ph]

    2024

  38. [46]

    O. M. Raisuddin and S. De, Quantum relaxation method for linear systems in finite element problems (2023), arXiv:2308.01377 [quant-ph]

    2023 arXiv

  39. [47]

    O. M. Raisuddin and S. De, Quantum multigrid algorithm for finite element problems (2024), arXiv:2404.07466 [quant-ph]

    2024 arXiv

  40. [48]

    Jin and N

    S. Jin and N. Liu, Quantum simulation of discrete linear dy- namical systems and simple iterative methods in linear algebra via schrodingerisation (2023), arXiv:2304.02865 [quant-ph]

    2023 arXiv

  41. [49]

    M. A. Nielsen and I. L. Chuang, Quantum computation and quantum information (Cambridge university press, 2010)

    2010

  42. [50]

    Huang, R

    H.-Y . Huang, R. Kueng, and J. Preskill, Nature Physics 16, 1050 (2020)

    2020

  43. [51]

    Jerbi, C

    S. Jerbi, C. Gyurik, S. C. Marshall, R. Molteni, and V . Dunjko, Nature Communications 15, 5676 (2024)

    2024

  44. [52]

    Zhou and Z

    Y . Zhou and Z. Liu, A hybrid framework for estimating non- linear functions of quantum states (2023), arXiv:2208.08416 [quant-ph]

    2023 arXiv

  45. [53]

    Biamonte, P

    J. Biamonte, P. Wittek, N. Pancotti, P. Rebentrost, N. Wiebe, and S. Lloyd, Nature 549, 195 (2017)

    2017

  46. [54]

    Linden, A

    N. Linden, A. Montanaro, and C. Shao, Communications in Mathematical Physics 395, 601 (2022)

    2022

  47. [55]

    Lloyd, M

    S. Lloyd, M. Mohseni, and P. Rebentrost, Nature Physics 10, 631 (2014)

    2014

  48. [56]

    Benedetti, E

    M. Benedetti, E. Lloyd, S. Sack, and M. Fiorentini, Quantum Science and Technology 4, 043001 (2019)

    2019

  49. [57]

    Wiebe, D

    N. Wiebe, D. Braun, and S. Lloyd, Phys. Rev. Lett. 109, 050505 (2012)

    2012

  50. [58]

    Rebentrost, M

    P. Rebentrost, M. Mohseni, and S. Lloyd, Phys. Rev. Lett.113, 130503 (2014)

    2014

  51. [59]

    Schuld, I

    M. Schuld, I. Sinayskiy, and F. Petruccione, Phys. Rev. A 94, 022342 (2016)

    2016

  52. [60]

    Schuld and N

    M. Schuld and N. Killoran, PRX Quantum 3, 030101 (2022)

    2022

  53. [62]

    Schuld and N

    M. Schuld and N. Killoran, Phys. Rev. Lett. 122, 040504 (2019)

    2019

  54. [63]

    T. Goto, Q. H. Tran, and K. Nakajima, Phys. Rev. Lett. 127, 090506 (2021)

    2021

  55. [64]

    Schuld, Supervised quantum machine learning models are kernel methods (2021), arXiv:2101.11020 [quant-ph]

    M. Schuld, Supervised quantum machine learning models are kernel methods (2021), arXiv:2101.11020 [quant-ph]

    2021 arXiv

  56. [65]

    Schuld, R

    M. Schuld, R. Sweke, and J. J. Meyer, Phys. Rev. A 103, 032430 (2021)

    2021

  57. [66]

    Abbas, D

    A. Abbas, D. Sutter, C. Zoufal, A. Lucchi, A. Figalli, and S. Woerner, Nature Computational Science 1, 403 (2021)

    2021

  58. [67]

    Liu and L

    J.-G. Liu and L. Wang, Phys. Rev. A 98, 062324 (2018)

    2018

  59. [68]

    Zoufal, A

    C. Zoufal, A. Lucchi, and S. Woerner, npj Quantum Informa- tion 5, 103 (2019)

    2019

  60. [69]

    Coyle, D

    B. Coyle, D. Mills, V . Danos, and E. Kashefi, npj Quantum Information 6, 60 (2020)

    2020

  61. [70]

    Kyriienko, A

    O. Kyriienko, A. E. Paine, and V . E. Elfving, Phys. Rev. Res. 6, 033291 (2024)

    2024

  62. [71]

    Kasture, O

    S. Kasture, O. Kyriienko, and V . E. Elfving, Phys. Rev. A108, 042406 (2023)

    2023

  63. [72]

    M. S. Rudolph, S. Lerch, S. Thanasilp, O. Kiss, O. Shaya, S. Vallecorsa, M. Grossi, and Z. Holmes, npj Quantum Infor- mation 10, 116 (2024)

    2024

  64. [73]

    H.-Y . Wu, V . E. Elfving, and O. Kyriienko, Ad- vanced Quantum Technologies , 2400337 (2024), https://onlinelibrary.wiley.com/doi/pdf/10.1002/qute.202400337

    2024 doi

  65. [74]

    Kyriienko, A

    O. Kyriienko, A. E. Paine, and V . E. Elfving, Physical Review A 103, 052416 (2021)

    2021

  66. [75]

    N. Heim, A. Ghosh, O. Kyriienko, and V . E. Elfving, Quantum model-discovery (2021), arXiv:2111.06376 [quant-ph]

    2021 arXiv

  67. [76]

    Markidis, Frontiers in Applied Mathematics and Statistics 8 (2022)

    S. Markidis, Frontiers in Applied Mathematics and Statistics 8 (2022)

    2022

  68. [77]

    A. E. Paine, V . E. Elfving, and O. Kyriienko, Advanced Quan- tum Technologies 6, 2300065 (2023)

    2023

  69. [78]

    A. E. Paine, V . E. Elfving, and O. Kyriienko, Physical Review A 107, 032428 (2023)

    2023

  70. [79]

    Jaderberg, A

    B. Jaderberg, A. A. Gentile, Y . A. Berrada, E. Shishenina, and V . E. Elfving, Phys. Rev. A109, 042421 (2024)

    2024

  71. [80]

    Jaderberg, A

    B. Jaderberg, A. A. Gentile, A. Ghosh, V . E. Elfving, C. Jones, D. V odola, J. Manobianco, and H. Weiss, Phys. Rev. A 110, 052423 (2024)

    2024

  72. [81]

    Holmes, K

    Z. Holmes, K. Sharma, M. Cerezo, and P. J. Coles, PRX Quan- tum 3, 010313 (2022)

    2022

  73. [82]

    J. R. McClean, S. Boixo, V . N. Smelyanskiy, R. Babbush, and H. Neven, Nature communications 9, 4812 (2018)

    2018

  74. [83]

    Fontana, D

    E. Fontana, D. Herman, S. Chakrabarti, N. Kumar, R. Yalovet- zky, J. Heredge, S. H. Sureshbabu, and M. Pistoia, The adjoint is all you need: Characterizing barren plateaus in quantum ans¨atze (2023), arXiv:2309.07902 [quant-ph]

    2023 arXiv

  75. [84]

    S. H. Sack, R. A. Medina, A. A. Michailidis, R. Kueng, and M. Serbyn, PRX Quantum 3, 020365 (2022)

    2022

  76. [85]

    Pesah, M

    A. Pesah, M. Cerezo, S. Wang, T. V olkoff, A. T. Sornborger, and P. J. Coles, Physical Review X11, 041011 (2021)

    2021

  77. [86]

    Larocca, N

    M. Larocca, N. Ju, D. Garc ´ıa-Mart´ın, P. J. Coles, and M. Cerezo, Nature Computational Science 3, 542 (2023)

    2023

  78. [87]

    R. Puig, M. Drudis, S. Thanasilp, and Z. Holmes, Variational quantum simulation: a case study for understanding warm starts (2024), arXiv:2404.10044 [quant-ph]. 8

    2024 arXiv

  79. [88]

    Cerezo, M

    M. Cerezo, M. Larocca, D. Garc ´ıa-Mart´ın, N. L. Diaz, P. Braccia, E. Fontana, M. S. Rudolph, P. Bermejo, A. Ijaz, S. Thanasilp, E. R. Anschuetz, and Z. Holmes, Does prov- able absence of barren plateaus imply classical simulability? or, why we need to rethink variational qua...

    2023 arXiv

  80. [89]

    Huang, M

    H.-Y . Huang, M. Broughton, J. Cotler, S. Chen, J. Li, M. Mohseni, H. Neven, R. Babbush, R. Kueng, J. Preskill, and J. R. McClean, Science 376, 1182 (2022)

    2022

  81. [90]

    Cerezo, G

    M. Cerezo, G. Verdon, H.-Y . Huang, L. Cincio, and P. J. Coles, Nature Computational Science 2, 567 (2022)

    2022

  82. [91]

    M. C. Caro, H.-Y . Huang, M. Cerezo, K. Sharma, A. Sorn- borger, L. Cincio, and P. J. Coles, Nature Communications13, 4919 (2022)

    2022

  83. [92]

    M. C. Caro, H.-Y . Huang, N. Ezzell, J. Gibbs, A. T. Sorn- borger, L. Cincio, P. J. Coles, and Z. Holmes, Nature Commu- nications 14, 3751 (2023)

    2023

  84. [93]

    Umeano, A

    C. Umeano, A. E. Paine, V . E. Elfving, and O. Kyriienko, What can we learn from quantum convolutional neural net- works? (2023), arXiv:2308.16664 [quant-ph]

    2023 arXiv

  85. [94]

    Gil-Fuster, J

    E. Gil-Fuster, J. Eisert, and C. Bravo-Prieto, Understanding quantum machine learning also requires rethinking general- ization (2023), arXiv:2306.13461 [quant-ph]

    2023 arXiv

  86. [95]

    Umeano, V

    C. Umeano, V . E. Elfving, and O. Kyriienko, Geometric quan- tum machine learning of bqpa protocols and latent graph clas- sifiers (2024), arXiv:2402.03871 [quant-ph]

    2024 arXiv

  87. [96]

    Umeano, S

    C. Umeano, S. Scali, and O. Kyriienko, Can geometric quan- tum machine learning lead to advantage in barcode classifica- tion? (2024), arXiv:2409.01496 [quant-ph]

    2024 arXiv

  88. [97]

    V olume 2: Multiphase Flow (MFTC); Computational Fluid Dynamics (CFDTC); Micro and Nano Fluid Dynamics (MNFDTC) (2022)

    Is it Turbulent or Laminar? Convolutional Neural Network Predictions, Fluids Engineering Division Summer Meeting, V ol. V olume 2: Multiphase Flow (MFTC); Computational Fluid Dynamics (CFDTC); Micro and Nano Fluid Dynamics (MNFDTC) (2022)

    2022

  89. [98]

    L. Deng, Y . Wang, Y . Liu, F. Wang, S. Li, and J. Liu, J. Vis. 22, 65–78 (2019)

    2019

  90. [99]

    Chen and R

    F. Chen and R. Samtaney, Physics of Fluids 34, 10.1063/5.0105098 (2022)

    2022 doi

  91. [100]

    Sciml: Open source software for scientific machine learning, https://sciml.ai/ (2023), accessed: 30-07-2023

    2023

  92. [101]

    Rackauckas, M

    C. Rackauckas, M. Innes, Y . Ma, J. Bettencourt, L. White, and V . Dixit, arXiv preprint arXiv:1902.02376 (2019)

    2019 arXiv

  93. [102]

    Rackauckas, Y

    C. Rackauckas, Y . Ma, J. Martensen, C. Warner, K. Zubov, R. Supekar, D. Skinner, and A. Ramadhan, arXiv preprint arXiv:2001.04385 (2020)

    2020 arXiv

  94. [103]

    Lubasch, J

    M. Lubasch, J. Joo, P. Moinier, M. Ki ffner, and D. Jaksch, Physical Review A 101, 010301 (2020)

    2020

  95. [104]

    A. E. Paine, V . E. Elfving, and O. Kyriienko, Physics-informed quantum machine learning: Solving nonlinear di fferential equations in latent spaces without costly grid evaluations (2023), arXiv:2308.01827 [quant-ph]

    2023 arXiv

  96. [105]

    An, J.-P

    D. An, J.-P. Liu, D. Wang, and Q. Zhao, A theory of quantum differential equation solvers: limitations and fast-forwarding (2023), arXiv:2211.05246 [quant-ph]

    2023 arXiv

  97. [106]

    B. Y . Gan, P.-W. Huang, E. Gil-Fuster, and P. Rebentrost, Con- cept learning of parameterized quantum models from limited measurements (2024), arXiv:2408.05116 [quant-ph]

    2024 arXiv

  98. [107]

    Gyurik and V

    C. Gyurik and V . Dunjko, Exponential separations between classical and quantum learners (2023), arXiv:2306.16028 [quant-ph]

    2023 arXiv

  99. [108]

    I. Cong, S. Choi, and M. D. Lukin, Nature Physics 15, 1273–1278 (2019)

    2019

  100. [109]

    Herrmann, S

    J. Herrmann, S. M. Llima, A. Remm, P. Zapletal, N. A. McMa- hon, C. Scarato, F. Swiadek, C. K. Andersen, C. Hellings, S. Krinner, N. Lacroix, S. Lazar, M. Kerschbaum, D. C. Zanuz, G. J. Norris, et al., Nature Communications 13, 4144 (2022)

    2022

  101. [110]

    Zapletal, N

    P. Zapletal, N. A. McMahon, and M. J. Hartmann, Error-tolerant quantum convolutional neural networks for symmetry-protected topological phases (2023), arXiv:2307.03711 [quant-ph]

    2023 arXiv

  102. [111]

    Shen and J

    F. Shen and J. Liu, Qfcnn: Quantum fourier convolutional neu- ral network (2021), arXiv:2106.10421 [quant-ph]

    2021 arXiv

  103. [112]

    Chinzei, Q

    K. Chinzei, Q. H. Tran, K. Maruyama, H. Oshima, and S. Sato, Phys. Rev. Res. 6, 023042 (2024)

    2024

  104. [113]

    L. C. Sander, N. A. McMahon, P. Zapletal, and M. J. Hartmann, Quantum convolutional neural network for phase recognition in two dimensions (2024), arXiv:2407.04114 [quant-ph]

    2024

  105. [114]

    Bermejo, P

    P. Bermejo, P. Braccia, M. S. Rudolph, Z. Holmes, L. Cincio, and M. Cerezo, Quantum convolutional neural networks are (effectively) classically simulable (2024), arXiv:2408.12739 [quant-ph]

    2024 arXiv

  106. [115]

    Huang, M

    H.-Y . Huang, M. Broughton, J. Cotler, S. Chen, J. Li, M. Mohseni, H. Neven, R. Babbush, R. Kueng, J. Preskill, and J. R. McClean, Science 376, 1182–1186 (2022)

    2022

  107. [116]

    K. Beer, D. Bondarenko, T. Farrelly, T. J. Osborne, R. Salz- mann, D. Scheiermann, and R. Wolf, Nature Communications 11 (2019)

    2019

Pith tools

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