Pith. sign in

REVIEW 4 major objections 4 minor 75 references

A quantum circuit whose latent propagator has at most eight trainable parameters matches the accuracy of a 524,288-parameter classical surrogate on three complex flows and stays stable over long rollouts.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 07:30 UTC pith:I3BV5G7Y

load-bearing objection Novel quantum-compressed surrogate with a clean unitarity argument, but the abstract overclaims the parameter count and the stability bound is for the latent operator, not the field-space rollout. the 4 major comments →

arxiv 2607.21688 v1 pith:I3BV5G7Y submitted 2026-07-23 physics.flu-dyn cs.LGquant-ph

Explainable quantum-compressed machine learning for complex fluid flows

Xiao Xue , Maida Wang , Mingyang Gao , Minh Chung , Peter V. Coveney This is my paper
classification physics.flu-dyn cs.LGquant-ph
keywords quantum-compressed machine learningfluid-flow surrogatesKoopman operator learningunitary latent propagatorautoregressive stabilityturbulent channel flowhaemodynamicsparameter-efficient quantum circuits
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

Machine-learning surrogates for complex flows face a trade-off: expressive models are opaque and parameter-heavy, while interpretable reduced-order models lack the nonlinear capacity needed for turbulence and patient-specific haemodynamics. The paper introduces quantum-compressed machine learning (QCML), in which a classical transformer encoder-decoder maps each flow state into a quantum latent space and a structured quantum circuit propagates that state forward in time. The central claim is that the latent propagator can be compressed from 524,288 trainable parameters to as few as eight, because the circuit's exact unitarity pins latent eigenvalues to the unit circle and converts exponential autoregressive error growth into linear accumulation. On turbulent channel flow, stenotic aortic flow, and abdominal aortic aneurysm haemodynamics, QCML matches the classical baseline on pressure spectra, pressure drop, and wall shear stress, and it remains coherent over five Lyapunov times after the classically regularised baseline collapses. A reader should care because the paper offers a concrete route to surrogate models that are stable over long horizons, cheap to run, and transparent enough that each parameter corresponds to a modal frequency or coupling strength.

Core claim

On the paper's own terms, QCML is a hybrid quantum-classical surrogate for nonlinear flows. The encoder-decoder pair (φ, ψ) learns a latent space in which the flow evolution is approximately linear; the latent propagator is implemented not as a dense matrix but as U_q(θ) of Eq. (5), a product of layers of mode-wise phase rotations exp(−iα X_i) and sparse pairwise XY couplings exp(−iβ(X_i X_j + Y_i Y_j)). Because U_q is unitary by construction, every eigenvalue lies exactly on the unit circle for every parameter value, so the latent norm is conserved and the Koopman-Galerkin error bound becomes linear in the number of rollout steps n, ‖K^n g − U_q^n g‖ ≤ n ε_N, rather than exponential. The pa

What carries the argument

The central object is the structured unitary latent propagator U_q(θ) of Eq. (5), generated by layer-wise Hamiltonians H_ℓ = Σ_i α_{ℓ,i} X_i + Σ_{(i,j)∈E} β_{ℓ,(i,j)}(X_i X_j + Y_i Y_j). In the structured variant QCML(S), α and β are shared across all qubits and edges within each layer, leaving only 2L trainable scalars (L=4 gives 8). Unitarity is the load-bearing identity: U_q† U_q = I pins the latent spectrum to the unit circle as an operator identity, making the Euclidean norm a strict Lyapunov function and collapsing the classical exponential rollout bound to the linear Koopman-Galerkin bound ‖K^n g − U_q^n g‖ ≤ n ε_N, where ε_N is the per-step encoder-decoder invariance defect. This mec

Load-bearing premise

The linear-stability result rests on the encoder-decoder keeping a small, roughly constant per-step reconstruction error while the model is fed its own output, and on the eight-parameter circuit being expressive enough to represent the flow's latent dynamics; if either fails, the linear accumulation claim collapses.

What would settle it

Run the same autoregressive window-rollout protocol on a new flow with stronger nonlinearity (e.g., a higher-Reynolds-number channel or a non-periodic patient geometry) and record the per-step encoder-decoder reconstruction defect and the total rollout error as functions of step count n. If the per-step defect grows with n, or if the total error grows faster than linearly as predicted by Eq. (13), the central mechanism is falsified; if the error remains linear with slope equal to the measured defect, the explanation is confirmed.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Long-horizon autoregressive surrogates can drop soft spectral regularisers: the unitary constraint holds exactly for every parameter value, so rollouts over several Lyapunov times avoid the collapse that a soft-penalised classical latent propagator exhibits.
  • The parameter compression to 2L scalars (as few as 8) means the learned dynamical law is readable as modal frequencies and inter-mode couplings, rather than as an opaque 524,288-entry matrix.
  • On stenotic aortic flow and abdominal aortic aneurysm haemodynamics, the structured surrogate preserves pressure-drop and wall-shear-stress statistics, supporting its use as a fast, transparent screening tool in cardiovascular modelling.
  • Latent-propagation cost drops by a factor of roughly 7,300 per cardiac cycle in the emulator, before state-preparation and data-transfer overheads, bringing surrogate inference closer to decision-making timescales.
  • The structured circuit is shallow enough that a 54-qubit processor can run multiple copies in parallel; one-step agreement of 79.69% under device noise suggests the approach is not immediately destroyed by hardware noise, though the cardiovascular cases were evaluated on a noise-aware emulator.

Where Pith is reading between the lines

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

  • Editorial inference: The linear stability bound is conditional on the encoder-decoder invariance defect staying small; measuring that defect directly during self-prediction on a higher-Reynolds or more strongly nonlinear flow would show whether the mechanism transfers beyond the three reported cases.
  • Editorial inference: The claimed speedup is algorithmic and excludes state preparation, measurement shots, and classical-quantum data transfer; an end-to-end hardware comparison would be needed to know whether practical wall-clock advantage appears on current devices.
  • Editorial inference: Because the structured ansatz parameterises frequencies and pairwise couplings, the same compression could be tried on other spectral-interpretable surrogates, such as weather or plasma models, where a handful of physically meaningful parameters is more useful than a million-weight black box.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper introduces QCML, a hybrid quantum-classical surrogate for nonlinear fluid flows. A transformer encoder-decoder maps physical fields into a latent space, and a parameterised quantum circuit -- restricted to a structured family of single-qubit rotations and XY couplings -- propagates the latent state. By sharing the phase and coupling parameters across all qubits and edges within each layer, the latent propagator uses only 2L trainable scalars (8 or 16 in the examples). The authors prove that unitarity pins the latent spectrum to the unit circle and yields a linear-in-n Koopman--Galerkin error bound instead of exponential growth. The method is evaluated on turbulent channel flow at Re_tau=180, stenotic aortic flow, and patient-specific abdominal aortic aneurysm haemodynamics, with claims of matching or exceeding a classical machine-learning baseline on spectra, pressure drop, wall shear stress, and long-rollout stability. One benchmark was run on IQM's 54-qubit Emerald processor; the others used a noise-aware emulator.

Significance. If the central claims hold, this is a valuable contribution: a parameter-efficient surrogate with a structural, exact stability constraint (unitarity) that is not merely a soft regularisation target, together with a nontrivial demonstration on three fluid-flow benchmarks of increasing complexity. The unitarity argument in Methods is mathematically clean, and the numerical comparisons are broader than in most quantum-ML fluid papers. The paper is also honest about excluding hardware overheads from its speedup estimate. However, the theoretical bound is not directly validated for the autoregressive loop actually deployed, the headline parameter-count claim is internally inconsistent, and the interpretability/quantum-advantage statements are stronger than the presented evidence. These issues are fixable, and the empirical core is worth preserving.

major comments (4)
  1. [Methods B, Eq. (13); Secs. II-IV] The linear-in-n bound in Eq. (13) is stated for repeated application of U_q to a fixed observable g. The deployed surrogate is autoregressive and re-encodes its own output at every step: z_{t+1} = phi(psi(U_q z_t)). The bound therefore applies only if the encoder-decoder pair is approximately invariant on the latent manifold, with a small, non-accumulating per-step defect epsilon_N. This epsilon_N is never reported, and no test of the invariance assumption on the rollout manifold is provided. The TKE plateau in Fig. 2f and the one-step agreement rates in Table I are aggregate field-space diagnostics and do not test the latent-space mechanism. If epsilon_N grows during autoregression, the linear bound fails and the claimed separation from classical baselines is not explained. Please measure epsilon_N over rollouts, or explicitly restrict Eq. (13) to the idealised no-re-encoding setting an
  2. [Abstract; Sec. III, Fig. 3h] The abstract claims the latent propagator is compressed to 'no more than 8' parameters. Fig. 3h reports 16 trainable parameters for the stenotic-aorta QCML(S) model (L=8, 2L=16). 'No more than 8' is therefore false; the accurate phrasing is 'as few as 8' (with L=4 for the AAA case) or '8-16'. This is a factual error in the headline claim and must be corrected.
  3. [Eq. (5); Sec. I; Sec. VI] Explainability is a central selling point, but the ansatz in Eq. (5) shares one phase alpha_l across all qubits and one coupling beta_l across all edges per layer. Thus a single parameter does not map one-to-one to an identifiable modal frequency or inter-mode interaction; alpha_l is a global common-mode rotation. The statements in Sec. I ('each corresponding to an identifiable modal frequency or inter-mode coupling strength') and Sec. VI ('parameters map one-to-one onto identifiable mode frequencies') overstate what the parameter sharing permits. Please reformulate the interpretability claim precisely, e.g., in terms of layer-wise spectral generators H_l and their collective effect on the latent spectrum.
  4. [Abstract; Sec. V; Fig. 4h] The abstract's phrase 'a concrete contribution towards practical quantum advantage' is not supported by the evidence. Only the turbulent-channel model was run on actual quantum hardware (Emerald), with 79.69% one-step agreement; the cardiovascular benchmarks were evaluated on a noise-aware emulator. The reported inference speedups (Fig. 4h) come from a classical emulator and explicitly exclude state-preparation, measurement-shot, and data-transfer overheads, so they are not a hardware wall-clock advantage. The claims should be softened to 'algorithmic cost reduction in emulation' unless dedicated end-to-end hardware benchmark data are provided.
minor comments (4)
  1. [Table I] The agreement metric averages losses within each batch and then over test-loader batches. If batch sizes vary, this is not the same as a mean over all test samples; please state the aggregation precisely.
  2. [Fig. 2f, Sec. II] The caption calls the plotted quantity 'cumulative turbulent-kinetic-energy error relative to the reference (relative L2)', while the text says a value of 1 corresponds to a vanishing predicted TKE. Clarify the normalisation (e.g., ||TKE_pred - TKE_ref||_2 / ||TKE_ref||_2) so the interpretation is unambiguous.
  3. [Sec. II] The Lyapunov time T_lambda used for the nondimensional time t* is not defined in the main text. Please give its value or estimation method in the main text or refer to the specific supplementary equation number.
  4. [Sec. V / Methods C] Please report the total number of hardware executions and the number of measurement shots per expectation value in the main text or Table I caption, since the 79.69% figure is a single one-step number and reproducibility would benefit from this context.

Circularity Check

0 steps flagged

No significant circularity: the stability bound is a structural consequence of unitarity, benchmarks are held-out, and the parameter compression is by construction.

full rationale

The paper's derivation chain is not circular. The central stability claim, Eq. (13), follows from the constructed unitarity of U_q(θ) in Eq. (5), the Lyapunov identity in Eq. (12), and a Koopman–Galerkin telescoping bound. This is a mathematical consequence of the model class, not a quantity fitted to the benchmark outputs. The 524,288-to-8 parameter compression is by construction of the shared-phase/shared-coupling ansatz, and the empirical results (turbulent channel stability, pressure spectra, pressure drop, WSS) are evaluated on held-out trajectories under autoregressive rollout. The only potentially load-bearing self-citation is ref. 42, which defines the classical soft-unitarity baseline; however, the paper states that this baseline 'was trained on the same split for comparison' (Sec. II), so it is an independently executed empirical baseline rather than an unverified self-citation that the central claim reduces to. The reviewer's concern that the per-step encoder–decoder defect ε_N is not measured is an assumption/verification gap, not circularity: Eq. (13) is conditional on a small per-step defect, and the paper does not claim to have measured it or to have fitted it from the reported metrics. No equation is equivalent to its input by construction, and no prediction reduces to a fitted parameter renamed as a result.

Axiom & Free-Parameter Ledger

4 free parameters · 6 axioms · 0 invented entities

The theoretical stability argument adds no ad-hoc fitted constants; the fitted content is the model's trained angles and hyperparameters. The main unproved premises are the Koopman-invariance of the learned latent subspace, the constancy of ε_N, and the sufficiency of the highly shared ansatz. No new physical entities (particles, forces, dimensions) are introduced.

free parameters (4)
  • Structured ansatz angles α_l, β_l (2L scalars) = L=4 for AAA (8 params), L=8 for stenosis (16 params); numeric values not reported
    Trained on CFD/LBM data; the headline 8-parameter compression refers to these fitted scalars.
  • Transformer encoder–decoder parameters = not reported
    Not included in the latent-propagator count but essential to the surrogate; the compression claim explicitly excludes them.
  • Training hyperparameters (Nq, L, optimizer, loss weights, shot count) = not fully reported; supplement S3.4/S3.5 referenced
    Needed to reproduce the figures; currently only partially described in the main text.
  • Encoder–decoder invariance defect ε_N = not measured
    This is the slope of the linear stability bound Eq. (13); the paper does not report its value, so the bound cannot be evaluated quantitatively.
axioms (6)
  • domain assumption Koopman operator framework: nonlinear dynamics linearize on an invariant subspace of observables, and the encoder–decoder approximates this subspace.
    Eq. (3) and Methods A; the entire latent-propagation formulation depends on this approximation.
  • domain assumption The resolved flows are measure-preserving/statistically stationary over the rollout window, so ρ(K)=1 is the correct latent target.
    Methods B, Eq. (10); needed for the claim that exact unitarity is desirable rather than spuriously non-dissipative.
  • domain assumption The per-step encoder–decoder defect ε_N remains constant and does not itself accumulate over the rollout.
    Eqs. (7) and (13); no measured value or proof of non-accumulation is supplied.
  • standard math Gelfand's spectral radius formula and Lyapunov's stability criterion apply as used.
    Methods B, refs. 48 and 72; standard results, correctly invoked.
  • ad hoc to paper The structured ansatz with globally shared α,β per layer retains enough expressivity for turbulent and cardiovascular latent dynamics.
    Eq. (5); only empirically validated on three benchmarks, with no expressivity bound or comparison to exactly-unitary classical models.
  • domain assumption Amplitude encoding and expectation-value reconstruction faithfully represent the latent state.
    Eq. (4) and Supplementary S4; the latent-to-quantum embedding is assumed to be lossless enough for the accuracy claims.

pith-pipeline@v1.3.0-alltime-deepseek · 18207 in / 17233 out tokens · 156925 ms · 2026-08-01T07:30:36.848404+00:00 · methodology

0 comments
read the original abstract

Machine-learning surrogates of physical systems face a paradox: explainable models facing the challenge of expressivity to capture complex nonlinear flows, whereas expressive deep surrogates match high-fidelity simulations only through massive parameterisations that turn the learned dynamics into a black box. Here, we introduce quantum-compressed machine learning (QCML), which resolves this tension by compressing the latent propagator of a flow surrogate from $524{,}288$ trainable parameters to no more than $8$. This parameter reduction brings the learned dynamical law to the parameter scale of a physical constitutive relation rather than a black-box neural network, making the surrogate directly interpretable and controllable without sacrificing expressivity. The compression is realised by a structured quantum circuit whose unitary propagator constrains the latent spectrum to the unit circle exactly and by construction, replacing exponential error growth with linear accumulation over autoregressive rollouts. Classical regularisation only approximates this constraint: even a quantum-inspired classical baseline penalised towards unitarity collapses within one Lyapunov time on turbulent channel flow, whereas QCML remains stable over the full rollout. Shared phase and coupling angles parameterising the circuit correspond directly to modal frequencies and inter-mode interactions, giving the learned dynamics a physical interpretation in spectral space. On two patient-specific cardiovascular benchmarks, the structured QCML propagator matches the predictive accuracy of its classical counterpart on surface pressure spectra, pressure drop, and wall shear stress. These results establish QCML as a working component of scientific machine learning and a concrete contribution towards practical quantum advantage in real-world prediction.

Figures

Figures reproduced from arXiv: 2607.21688 by Maida Wang, Mingyang Gao, Minh Chung, Peter V. Coveney, Xiao Xue.

Figure 1
Figure 1. Figure 1: FIG. 1 [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4 [PITH_FULL_IMAGE:figures/full_fig_p013_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5 [PITH_FULL_IMAGE:figures/full_fig_p022_5.png] view at source ↗

discussion (0)

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

Reference graph

Works this paper leans on

75 extracted references · 1 canonical work pages

  1. [1]

    Duraisamy , author G

    author author K. Duraisamy , author G. Iaccarino ,\ and\ author H. Xiao ,\ title title Turbulence modeling in the age of data , \ @noop journal journal Annual review of fluid mechanics \ volume 51 ,\ pages 357--377 ( year 2019 ) NoStop

    2019

  2. [2]

    Bi , author L

    author author K. Bi , author L. Xie , author H. Zhang , author X. Chen , author X. Gu ,\ and\ author Q. Tian ,\ title title Accurate medium-range global weather forecasting with 3d neural networks , \ @noop journal journal Nature \ volume 619 ,\ pages 533--538 ( year 2023 ) NoStop

    2023

  3. [3]

    Lam , author A

    author author R. Lam , author A. Sanchez-Gonzalez , author M. Willson , author P. Wirnsberger , author M. Fortunato , author F. Alet , author S. Ravuri , author T. Ewalds , author Z. Eaton-Rosen , author W. Hu , et al. ,\ title title Learning skillful medium-range global weather forecasting , \ @noop journal journal Science \ volume 382 ,\ pages 1416--142...

    2023

  4. [4]

    Cho , author A

    author author J. Cho , author A. Lazarian ,\ and\ author E. T. \ Vishniac ,\ title title Simulations of magnetohydrodynamic turbulence in a strongly magnetized medium , \ @noop journal journal The Astrophysical Journal \ volume 564 ,\ pages 291 ( year 2002 ) NoStop

    2002

  5. [5]

    Gopakumar , author S

    author author V. Gopakumar , author S. Pamela , author L. Zanisi , author Z. Li , author A. Gray , author D. Brennand , author N. Bhatia , author G. Stathopoulos , author M. Kusner , author M. P. \ Deisenroth , et al. ,\ title title Plasma surrogate modelling using fourier neural operators , \ @noop journal journal Nuclear Fusion \ volume 64 ,\ pages 0560...

    2024

  6. [6]

    Feiger , author J

    author author B. Feiger , author J. Gounley , author D. Adler , author J. A. \ Leopold , author E. W. \ Draeger , author R. Chaudhury , author J. Ryan , author G. Pathangey , author K. Winarta , author D. Frakes , et al. ,\ title title Accelerating massively parallel hemodynamic models of coarctation of the aorta using neural networks , \ @noop journal jo...

    2020

  7. [7]

    Xue , author T

    author author X. Xue , author T. M. \ Athawale , author J. W. \ McCullough , author S. C. \ Lo , author I. Zacharoudiou , author B. Joo , author A. Georgiadou ,\ and\ author P. V. \ Coveney ,\ title title An uncertainty visualization framework for large-scale cardiovascular flow simulations: A case study on aortic stenosis , \ @noop journal journal arXiv ...

    Pith/arXiv arXiv 2025

  8. [8]

    Tanade , author S

    author author C. Tanade , author S. J. \ Chen , author J. A. \ Leopold ,\ and\ author A. Randles ,\ title title Analysis identifying minimal governing parameters for clinically accurate in silico fractional flow reserve , \ @noop journal journal Frontiers in Medical Technology \ volume 4 ,\ pages 1034801 ( year 2022 ) NoStop

    2022

  9. [9]

    Leuprecht , author S

    author author A. Leuprecht , author S. Kozerke , author P. Boesiger ,\ and\ author K. Perktold ,\ title title Blood flow in the human ascending aorta: a combined MRI and CFD study , \ @noop journal journal Journal of engineering mathematics \ volume 47 ,\ pages 387--404 ( year 2003 ) NoStop

    2003

  10. [10]

    Groen , author J

    author author D. Groen , author J. Hetherington , author H. B. \ Carver , author R. W. \ Nash , author M. O. \ Bernabeu ,\ and\ author P. V. \ Coveney ,\ title title Analysing and modelling the performance of the HemeLB lattice- B oltzmann simulation environment , \ @noop journal journal Journal of Computational Science \ volume 4 ,\ pages 412--422 ( year...

    2013

  11. [11]

    author author M. D. \ Mazzeo \ and\ author P. V. \ Coveney ,\ title title HemeLB: A high performance parallel lattice- B oltzmann code for large scale fluid flow in complex geometries , \ @noop journal journal Computer Physics Communications \ volume 178 ,\ pages 894--914 ( year 2008 ) NoStop

    2008

  12. [12]

    author author S. L. \ Brunton , author B. R. \ Noack ,\ and\ author P. Koumoutsakos ,\ title title Machine learning for fluid mechanics , \ @noop journal journal Annual review of fluid mechanics \ volume 52 ,\ pages 477--508 ( year 2020 ) NoStop

    2020

  13. [13]

    Xue , author T

    author author X. Xue , author T. Yang , author M. Gao , author L. Pan , author M. Wang , author K. Zhu , author S. Wang , author J. Li , author M. F. \ ten Eikelder ,\ and\ author P. V. \ Coveney ,\ title title Uni-flow: a unified autoregressive-diffusion model for complex multiscale flows , \ @noop journal journal arXiv preprint arXiv:2602.15592 \ ( year...

    arXiv 2026

  14. [14]

    Price , author A

    author author I. Price , author A. Sanchez-Gonzalez , author F. Alet , author T. R. \ Andersson , author A. El-Kadi , author D. Masters , author T. Ewalds , author J. Stott , author S. Mohamed , author P. Battaglia , et al. ,\ title title Probabilistic weather forecasting with machine learning , \ @noop journal journal Nature \ volume 637 ,\ pages 84--90 ...

    2025

  15. [15]

    Pathak , author Y

    author author J. Pathak , author Y. Cohen , author P. Garg , author P. Harrington , author N. Brenowitz , author D. Durran , author M. Mardani , author A. Vahdat , author S. Xu , author K. Kashinath , et al. ,\ title title Kilometer-scale convection-allowing model emulation using generative diffusion modeling , \ @noop journal journal Science Advances \ v...

    2026

  16. [16]

    Passaro , author G

    author author S. Passaro , author G. Corso , author J. Wohlwend , author M. Reveiz , author S. Thaler , author V. R. \ Somnath , author N. Getz , author T. Portnoi , author J. Roy , author H. Stark , et al. ,\ title title Boltz-2 : Towards accurate and efficient binding affinity prediction , \ @noop journal journal BioRxiv \ ( year 2025 ) NoStop

    2025

  17. [17]

    author author A. P. \ Bhati , author S. Wan , author H. H. \ Loeffler , author M. Kl \"a hn , author M. Bieniek , author K. Liu , author S. Ahmad , author L. Halabelian , author J. Dong , author X. Zhang , et al. ,\ title title An integrated workflow comprising AI , physics and experiment: Discovery of nanomolar-potent inhibitors , \ @noop \ ( year 2026 ) NoStop

    2026

  18. [18]

    Wan , author X

    author author S. Wan , author X. Zhang , author X. Xue ,\ and\ author P. V. \ Coveney ,\ title title On the reliability of AI methods in drug discovery: Evaluation of Boltz-2 for structure and binding affinity prediction , \ @noop journal journal arXiv preprint arXiv:2603.05532 \ ( year 2026 ) NoStop

    arXiv 2026

  19. [19]

    author author G. E. \ Karniadakis , author I. G. \ Kevrekidis , author L. Lu , author P. Perdikaris , author S. Wang ,\ and\ author L. Yang ,\ title title Physics-informed machine learning , \ @noop journal journal Nature Reviews Physics \ volume 3 ,\ pages 422--440 ( year 2021 ) NoStop

    2021

  20. [20]

    Li , author N

    author author Z. Li , author N. Kovachki , author K. Azizzadenesheli , author B. Liu , author K. Bhattacharya , author A. Stuart ,\ and\ author A. Anandkumar ,\ title title Fourier neural operator for parametric partial differential equations , \ @noop journal journal arXiv preprint arXiv:2010.08895 \ ( year 2020 ) NoStop

    Pith/arXiv arXiv 2010

  21. [21]

    Oommen , author A

    author author V. Oommen , author A. Bora , author Z. Zhang ,\ and\ author G. E. \ Karniadakis ,\ title title Integrating neural operators with diffusion models improves spectral representation in turbulence modelling , \ https://doi.org/10.1098/rspa.2024.0819 journal journal Proceedings of the Royal Society A: Mathematical, Physical and Engineering Scienc...

    arXiv 2024

  22. [22]

    Raissi , author P

    author author M. Raissi , author P. Perdikaris ,\ and\ author G. E. \ Karniadakis ,\ title title Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations , \ @noop journal journal Journal of Computational physics \ volume 378 ,\ pages 686--707 ( year 2019 ) NoStop

    2019

  23. [23]

    Preskill ,\ title title Quantum computing in the NISQ era and beyond , \ @noop journal journal Quantum \ volume 2 ,\ pages 79 ( year 2018 ) NoStop

    author author J. Preskill ,\ title title Quantum computing in the NISQ era and beyond , \ @noop journal journal Quantum \ volume 2 ,\ pages 79 ( year 2018 ) NoStop

    2018

  24. [24]

    \ Huang , author S

    author author H.-Y. \ Huang , author S. Choi , author J. R. \ McClean ,\ and\ author J. Preskill ,\ title title The vast world of quantum advantage , \ @noop journal journal arXiv preprint arXiv:2508.05720 \ ( year 2025 ) NoStop

    Pith/arXiv arXiv 2025

  25. [25]

    Lubasch , author J

    author author M. Lubasch , author J. Joo , author P. Moinier , author M. Kiffner ,\ and\ author D. Jaksch ,\ title title Variational quantum algorithms for nonlinear problems , \ @noop journal journal Physical Review A \ volume 101 ,\ pages 010301 ( year 2020 ) NoStop

    2020

  26. [26]

    Sanavio \ and\ author S

    author author C. Sanavio \ and\ author S. Succi ,\ title title Lattice B oltzmann-- C arleman quantum algorithm and circuit for fluid flows at moderate R eynolds number , \ @noop journal journal AVS Quantum Science \ volume 6 ( year 2024 ) NoStop

    2024

  27. [27]

    Sanavio , author E

    author author C. Sanavio , author E. Mauri ,\ and\ author S. Succi ,\ title title Explicit quantum circuit for simulating the advection-diffusion-reaction dynamics , \ @noop journal journal IEEE Transactions on Quantum Engineering \ ( year 2025 ) NoStop

    2025

  28. [28]

    Kocherla , author Z

    author author S. Kocherla , author Z. Song , author F. E. \ Chrit , author B. Gard , author E. F. \ Dumitrescu , author A. Alexeev ,\ and\ author S. H. \ Bryngelson ,\ title title Fully quantum algorithm for mesoscale fluid simulations with application to partial differential equations , \ @noop journal journal AVS Quantum Science \ volume 6 ( year 2024 ) NoStop

    2024

  29. [29]

    Wawrzyniak , author J

    author author D. Wawrzyniak , author J. Winter , author S. Schmidt , author T. Indinger , author U. Schramm , author C. Jan en ,\ and\ author N. A. \ Adams ,\ title title Unitary quantum algorithm for the lattice- B oltzmann method , \ @noop journal journal arXiv preprint arXiv:2405.13391 \ ( year 2024 ) NoStop

    Pith/arXiv arXiv 2024

  30. [30]

    Mezi \'c ,\ title title Koopman operator, geometry, and learning of dynamical systems , \ @noop journal journal Not

    author author I. Mezi \'c ,\ title title Koopman operator, geometry, and learning of dynamical systems , \ @noop journal journal Not. Am. Math. Soc \ volume 68 ,\ pages 1087--1105 ( year 2021 ) NoStop

    2021

  31. [31]

    author author S. L. \ Brunton , author M. Budi s i\' c , author E. Kaiser ,\ and\ author J. N. \ Kutz ,\ title title Modern K oopman theory for dynamical systems , \ https://doi.org/10.1137/21M1401243 journal journal SIAM Review \ volume 64 ,\ pages 229--340 ( year 2022 ) NoStop

    doi:10.1137/21m1401243 2022

  32. [32]

    author author M. O. \ Williams , author I. G. \ Kevrekidis ,\ and\ author C. W. \ Rowley ,\ title title A data-driven approximation of the K oopman operator: Extended dynamic mode decomposition , \ @noop journal journal Journal of Nonlinear Science \ volume 25 ,\ pages 1307--1346 ( year 2015 ) NoStop

    2015

  33. [33]

    Korda \ and\ author I

    author author M. Korda \ and\ author I. Mezi \'c ,\ title title On convergence of extended dynamic mode decomposition to the K oopman operator , \ @noop journal journal Journal of Nonlinear Science \ volume 28 ,\ pages 687--710 ( year 2018 ) NoStop

    2018

  34. [34]

    Lu , author P

    author author L. Lu , author P. Jin , author G. Pang , author Z. Zhang ,\ and\ author G. E. \ Karniadakis ,\ title title Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators , \ @noop journal journal Nature machine intelligence \ volume 3 ,\ pages 218--229 ( year 2021 ) NoStop

    2021

  35. [35]

    Kovachki , author Z

    author author N. Kovachki , author Z. Li , author B. Liu , author K. Azizzadenesheli , author K. Bhattacharya , author A. Stuart ,\ and\ author A. Anandkumar ,\ title title Neural operator: Learning maps between function spaces with applications to PDE s , \ @noop journal journal Journal of Machine Learning Research \ volume 24 ,\ pages 1--97 ( year 2023 ) NoStop

    2023

  36. [36]

    Li , author M

    author author Z. Li , author M. Liu-Schiaffini , author N. Kovachki , author B. Liu , author K. Azizzadenesheli , author K. Bhattacharya , author A. Stuart ,\ and\ author A. Anandkumar ,\ title title Learning dissipative dynamics in chaotic systems , \ @noop journal journal arXiv preprint arXiv:2106.06898 \ ( year 2021 ) NoStop

    Pith/arXiv arXiv 2021

  37. [37]

    McCabe , author P

    author author M. McCabe , author P. Harrington , author S. Subramanian ,\ and\ author J. Brown ,\ title title Towards stability of autoregressive neural operators , \ @noop journal journal arXiv preprint arXiv:2306.10619 \ ( year 2023 ) NoStop

    Pith/arXiv arXiv 2023

  38. [38]

    Dunjko , author J

    author author V. Dunjko , author J. M. \ Taylor ,\ and\ author H. J. \ Briegel ,\ title title Quantum-enhanced machine learning , \ @noop journal journal Physical review letters \ volume 117 ,\ pages 130501 ( year 2016 ) NoStop

    2016

  39. [39]

    Ghazi Vakili , author C

    author author M. Ghazi Vakili , author C. Gorgulla , author J. Snider , author A. Nigam , author D. Bezrukov , author D. Varoli , author A. Aliper , author D. Polykovsky , author K. M. \ Padmanabha Das , author H. Cox Iii , et al. ,\ title title Quantum-computing-enhanced algorithm unveils potential KRAS inhibitors , \ @noop journal journal Nature biotech...

    1954

  40. [40]

    author author J. R. \ McClean , author S. Boixo , author V. N. \ Smelyanskiy , author R. Babbush ,\ and\ author H. Neven ,\ title title Barren plateaus in quantum neural network training landscapes , \ https://doi.org/10.1038/s41467-018-07090-4 journal journal Nature Communications \ volume 9 ,\ pages 4812 ( year 2018 ) NoStop

    doi:10.1038/s41467-018-07090-4 2018

  41. [41]

    Cerezo , author G

    author author M. Cerezo , author G. Verdon , author H.-Y. \ Huang , author L. Cincio ,\ and\ author P. J. \ Coles ,\ title title Challenges and opportunities in quantum machine learning , \ @noop journal journal Nature Computational Science \ volume 2 ,\ pages 567--576 ( year 2022 ) NoStop

    2022

  42. [42]

    Wang , author X

    author author M. Wang , author X. Xue , author M. Gao ,\ and\ author P. V. \ Coveney ,\ title title Quantum-informed machine learning for predicting spatiotemporal chaos with practical quantum advantage , \ @noop journal journal Science Advances \ volume 12 ,\ pages eaec5049 ( year 2026 ) NoStop

    2026

  43. [43]

    author author S. W. \ Zamir , author A. Arora , author S. Khan , author M. Hayat , author F. S. \ Khan ,\ and\ author M.-H. \ Yang ,\ title title Restormer: Efficient transformer for high-resolution image restoration , \ in\ @noop booktitle Proceedings of the IEEE/CVF conference on computer vision and pattern recognition \ ( year 2022 )\ pp.\ pages 5728--...

    2022

  44. [44]

    Cerezo , author A

    author author M. Cerezo , author A. Arrasmith , author R. Babbush , author S. C. \ Benjamin , author S. Endo , author K. Fujii , author J. R. \ McClean , author K. Mitarai , author X. Yuan , author L. Cincio ,\ and\ author P. J. \ Coles ,\ title title Variational quantum algorithms , \ https://doi.org/10.1038/s42254-021-00348-9 journal journal Nature Revi...

    doi:10.1038/s42254-021-00348-9 2021

  45. [45]

    Mitarai , author M

    author author K. Mitarai , author M. Negoro , author M. Kitagawa ,\ and\ author K. Fujii ,\ title title Quantum circuit learning , \ @noop journal journal Physical Review A \ volume 98 ,\ pages 032309 ( year 2018 ) NoStop

    2018

  46. [46]

    Goto , author Q

    author author T. Goto , author Q. H. \ Tran ,\ and\ author K. Nakajima ,\ title title Universal approximation property of quantum machine learning models in quantum-enhanced feature spaces , \ @noop journal journal Physical Review Letters \ volume 127 ,\ pages 090506 ( year 2021 ) NoStop

    2021

  47. [47]

    Wang , author J

    author author M. Wang , author J. Jiang ,\ and\ author P. V. \ Coveney ,\ title title Parameter-efficient quantum anomaly detection method on a superconducting quantum processor , \ @noop journal journal Physical Review Research \ volume 7 ,\ pages 043094 ( year 2025 ) NoStop

    2025

  48. [48]

    author author A. M. \ Lyapunov ,\ title title The general problem of the stability of motion , \ @noop journal journal International Journal of Control \ volume 55 ,\ pages 531--773 ( year 1992 ) ,\ note english translation by A. T. Fuller of the 1892 Russian doctoral dissertation, Kharkov Mathematical Society NoStop

    1992

  49. [49]

    author author H. K. \ Khalil ,\ @noop title Nonlinear Systems ,\ edition 3rd \ ed.\ ( publisher Prentice Hall ,\ address Upper Saddle River, NJ ,\ year 2002 ) NoStop

    2002

  50. [50]

    author author B. O. \ Koopman ,\ title title Hamiltonian systems and transformation in H ilbert space , \ @noop journal journal Proceedings of the National Academy of Sciences \ volume 17 ,\ pages 315--318 ( year 1931 ) NoStop

    1931

  51. [51]

    Bengio , author P

    author author Y. Bengio , author P. Simard ,\ and\ author P. Frasconi ,\ title title Learning long-term dependencies with gradient descent is difficult , \ @noop journal journal IEEE transactions on neural networks \ volume 5 ,\ pages 157--166 ( year 1994 ) NoStop

    1994

  52. [52]

    Cerezo , author A

    author author M. Cerezo , author A. Sone , author T. Volkoff , author L. Cincio ,\ and\ author P. J. \ Coles ,\ title title Cost function dependent barren plateaus in shallow parametrized quantum circuits , \ https://doi.org/10.1038/s41467-021-21728-w journal journal Nature Communications \ volume 12 ,\ pages 1791 ( year 2021 b ) NoStop

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

  53. [53]

    Larocca , author S

    author author M. Larocca , author S. Thanasilp , author S. Wang , author K. Sharma , author J. Biamonte , author P. J. \ Coles , author L. Cincio , author J. R. \ McClean , author Z. Holmes ,\ and\ author M. Cerezo ,\ title title A review of barren plateaus in variational quantum computing , \ https://doi.org/10.1038/s42254-025-00813-9 journal journal Nat...

    doi:10.1038/s42254-025-00813-9 2025

  54. [54]

    author author M. C. \ Caro , author H.-Y. \ Huang , author M. Cerezo , author K. Sharma , author A. Sornborger , author L. Cincio ,\ and\ author P. J. \ Coles ,\ title title Generalization in quantum machine learning from few training data , \ @noop journal journal Nature communications \ volume 13 ,\ pages 4919 ( year 2022 ) NoStop

    2022

  55. [55]

    Fukami , author Y

    author author K. Fukami , author Y. Nabae , author K. Kawai ,\ and\ author K. Fukagata ,\ title title Synthetic turbulent inflow generator using machine learning , \ @noop journal journal Physical Review Fluids \ volume 4 ,\ pages 064603 ( year 2019 ) NoStop

    2019

  56. [56]

    Fukami , author K

    author author K. Fukami , author K. Fukagata ,\ and\ author K. Taira ,\ title title Super-resolution reconstruction of turbulent flows with machine learning , \ @noop journal journal Journal of Fluid Mechanics \ volume 870 ,\ pages 106--120 ( year 2019 ) NoStop

    2019

  57. [57]

    Xue , author H.-D

    author author X. Xue , author H.-D. \ Yao ,\ and\ author L. Davidson ,\ title title Synthetic turbulence generator for lattice B oltzmann method at the interface between RANS and LES , \ @noop journal journal Physics of Fluids \ volume 34 ,\ pages 055118 ( year 2022 ) NoStop

    2022

  58. [58]

    Latt , author B

    author author J. Latt , author B. Chopard , author O. Malaspinas , author M. Deville ,\ and\ author A. Michler ,\ title title Straight velocity boundaries in the lattice B oltzmann method , \ @noop journal journal Physical Review E \ volume 77 ,\ pages 056703 ( year 2008 ) NoStop

    2008

  59. [59]

    Smagorinsky ,\ title title General circulation experiments with the primitive equations: I

    author author J. Smagorinsky ,\ title title General circulation experiments with the primitive equations: I. the basic experiment , \ @noop journal journal Monthly weather review \ volume 91 ,\ pages 99--164 ( year 1963 ) NoStop

    1963

  60. [60]

    Hou , author J

    author author S. Hou , author J. Sterling , author S. Chen ,\ and\ author G. Doolen ,\ title title A lattice B oltzmann subgrid model for high R eynolds number flows , \ @noop journal journal arXiv preprint comp-gas/9401004 \ ( year 1994 ) NoStop

    Pith/arXiv arXiv 1994

  61. [61]

    Koda \ and\ author F.-S

    author author Y. Koda \ and\ author F.-S. \ Lien ,\ title title The lattice B oltzmann method implemented on the GPU to simulate the turbulent flow over a square cylinder confined in a channel , \ @noop journal journal Flow, Turbulence and Combustion \ volume 94 ,\ pages 495--512 ( year 2015 ) NoStop

    2015

  62. [62]

    Xue , author J

    author author X. Xue , author J. W. \ McCullough , author S. C. \ Lo , author I. Zacharoudiou , author B. Jo \'o ,\ and\ author P. V. \ Coveney ,\ title title The lattice B oltzmann based large eddy simulations for the stenosis of the aorta , \ in\ @noop booktitle International Conference on Computational Science \ ( organization Springer Nature Switzerla...

    2024

  63. [63]

    author author N. M. \ Wilson , author A. K. \ Ortiz ,\ and\ author A. B. \ Johnson ,\ title title The Vascular Model Repository: A Public Resource of Medical Imaging Data and Blood Flow Simulation Results , \ https://doi.org/10.1115/1.4025983 journal journal Journal of Medical Devices \ volume 7 ( year 2013 ),\ 10.1115/1.4025983 NoStop

    work page doi:10.1115/1.4025983 2013

  64. [64]

    Zacharoudiou , author J

    author author I. Zacharoudiou , author J. McCullough ,\ and\ author P. Coveney ,\ title title Development and performance of a HemeLB GPU code for human-scale blood flow simulation , \ @noop journal journal Computer Physics Communications \ volume 282 ,\ pages 108548 ( year 2023 ) NoStop

    2023

  65. [65]

    Succi ,\ @noop title The lattice B oltzmann equation: for fluid dynamics and beyond \ ( publisher Oxford University Press ,\ year 2001 ) NoStop

    author author S. Succi ,\ @noop title The lattice B oltzmann equation: for fluid dynamics and beyond \ ( publisher Oxford University Press ,\ year 2001 ) NoStop

    2001

  66. [66]

    author author S. C. \ Lo , author J. W. \ McCullough , author X. Xue ,\ and\ author P. V. \ Coveney ,\ title title Uncertainty quantification of the impact of peripheral arterial disease on abdominal aortic aneurysms in blood flow simulations , \ @noop journal journal Journal of the Royal Society Interface \ volume 21 ,\ pages 20230656 ( year 2024 ) NoStop

    2024

  67. [67]

    Guo , author L

    author author Y. Guo , author L. Kleiser ,\ and\ author N. Adams ,\ title title A comparison study of an improved temporal DNS and spatial DNS of compressible boundary layer transition , \ in\ @noop booktitle Fluid Dynamics Conference \ ( year 1994 )\ p.\ pages 2371 NoStop

    1994

  68. [68]

    author author N. A. \ Adams ,\ title title Direct simulation of the turbulent boundary layer along a compression ramp at m= 3 and re = 1685 , \ @noop journal journal Journal of Fluid Mechanics \ volume 420 ,\ pages 47--83 ( year 2000 ) NoStop

    2000

  69. [69]

    author author S. C. \ Lo , author A. Zingaro , author J. W. \ McCullough , author X. Xue , author P. Gonzalez-Martin , author B. Joo , author M. V \'a zquez ,\ and\ author P. V. \ Coveney ,\ title title A multi-component, multi-physics computational model for solving coupled cardiac electromechanics and vascular haemodynamics , \ @noop journal journal Com...

    2025

  70. [70]

    Gonzalez-Conde , author T

    author author J. Gonzalez-Conde , author T. W. \ Watts , author P. Rodriguez-Grasa ,\ and\ author M. Sanz ,\ title title Efficient quantum amplitude encoding of polynomial functions , \ @noop journal journal Quantum \ volume 8 ,\ pages 1297 ( year 2024 ) NoStop

    2024

  71. [71]

    Pascanu , author T

    author author R. Pascanu , author T. Mikolov ,\ and\ author Y. Bengio ,\ title title On the difficulty of training recurrent neural networks , \ in\ @noop booktitle International Conference on Machine Learning (ICML) \ ( year 2013 )\ pp.\ pages 1310--1318 NoStop

    2013

  72. [72]

    author author I. M. \ Gelfand ,\ title title Normierte ringe , \ @noop journal journal Recueil Math\'ematique [Matematicheski Sbornik] N.S. \ volume 9(51) ,\ pages 3--24 ( year 1941 ) NoStop

    1941

  73. [73]

    author author P. D. \ Nation , author H. Kang , author N. Sundaresan ,\ and\ author J. M. \ Gambetta ,\ title title Scalable mitigation of measurement errors on quantum computers , \ @noop journal journal PRX Quantum \ volume 2 ,\ pages 040326 ( year 2021 ) NoStop

    2021

  74. [74]

    He , author B

    author author A. He , author B. Nachman , author W. A. \ de Jong ,\ and\ author C. W. \ Bauer ,\ title title Zero-noise extrapolation for quantum-gate error mitigation with identity insertions , \ @noop journal journal Physical Review A \ volume 102 ,\ pages 012426 ( year 2020 ) NoStop

    2020

  75. [75]

    author author J. J. \ Wallman \ and\ author J. Emerson ,\ title title Noise tailoring for scalable quantum computation via randomized compiling , \ @noop journal journal Physical Review A \ volume 94 ,\ pages 052325 ( year 2016 ) NoStop

    2016