pith. machine review for the scientific record. sign in

arxiv: 2604.09374 · v2 · submitted 2026-04-10 · 🪐 quant-ph · cs.LG

Recognition: unknown

Variational Quantum Physics-Informed Neural Networks for Hydrological PDE-Constrained Learning with Inherent Uncertainty Quantification

Prasad Nimantha Madusanka Ukwatta Hewage , Midhun Chakkravarthy , Ruvan Kumara Abeysekara
Authors on Pith no claims yet

Pith reviewed 2026-05-10 17:12 UTC · model grok-4.3

classification 🪐 quant-ph cs.LG
keywords hybrid quantum-classical PINNvariational quantum circuitsphysics-informed neural networkshydrological PDEsuncertainty quantificationquantum transfer learningSaint-Venant equationsbarren plateaus
0
0 comments X

The pith

A hybrid quantum-classical PINN constrained by river flow equations learns hydrological predictions with fewer epochs and parameters plus built-in uncertainty from quantum measurements.

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

The paper proposes a Hybrid Quantum-Classical Physics-Informed Neural Network that embeds variational quantum circuits into the PINN framework for learning from multi-source remote sensing data under hydrological constraints. Features are encoded into quantum states via angle encoding, processed by a hardware-efficient variational ansatz with entangling layers, and the outputs are regularized by differentiable loss terms from the Saint-Venant shallow water equations and Manning's flow equation. Quantum measurement stochasticity supplies uncertainty quantification directly, and a transfer learning protocol pre-trains on multi-hazard data before fine-tuning on floods. Numerical tests on Kalu River basin data indicate the hybrid model converges in roughly three times fewer epochs and with about 44 percent fewer parameters than a classical PINN while retaining competitive accuracy, with the physics losses helping to narrow the optimization space.

Core claim

Embedding a hardware-efficient variational quantum ansatz with entangling layers into a physics-informed neural network and constraining outputs with differentiable losses from the Saint-Venant shallow water equations and Manning's flow equation produces a model that converges in approximately three times fewer training epochs, uses about 44 percent fewer trainable parameters, and maintains competitive classification accuracy on multi-modal satellite and meteorological data from the Kalu River basin, while quantum measurement stochasticity provides inherent uncertainty quantification and the physics terms help mitigate barren plateaus.

What carries the argument

Hardware-efficient variational ansatz with entangling layers integrated into the PINN framework, where trainable angle encoding feeds multi-source data into quantum states whose outputs are constrained by differentiable physics losses from the Saint-Venant shallow water equations and Manning's flow equation.

Load-bearing premise

That the hardware-efficient variational ansatz with entangling layers, when combined with the hydrological PDE loss terms, produces the claimed efficiency gains and naturally mitigates barren plateaus in a manner that generalizes beyond the specific Kalu River dataset and simulation setup.

What would settle it

Running the HQC-PINN and an equivalent classical PINN on multi-modal data from a different river basin and checking whether the reported factor-of-three reduction in epochs and 44 percent parameter savings still appear or whether barren plateaus re-emerge without the physics constraints.

Figures

Figures reproduced from arXiv: 2604.09374 by Midhun Chakkravarthy, Prasad Nimantha Madusanka Ukwatta Hewage, Ruvan Kumara Abeysekara.

Figure 1
Figure 1. Figure 1: FIG. 1. Schematic of the HQC-PINN architecture. Classical pre-processing reduces input dimen [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
read the original abstract

We propose a Hybrid Quantum-Classical Physics-Informed Neural Network (HQC-PINN) that integrates parameterized variational quantum circuits into the PINN framework for hydrological PDE-constrained learning. Our architecture encodes multi-source remote sensing features into quantum states via trainable angle encoding, processes them through a hardware-efficient variational ansatz with entangling layers, and constrains the output using the Saint-Venant shallow water equations and Manning's flow equation as differentiable physics loss terms. The inherent stochasticity of quantum measurement provides a natural mechanism for uncertainty quantification without requiring explicit Bayesian inference machinery. We further introduce a quantum transfer learning protocol that pre-trains on multi-hazard disaster data before fine-tuning on flood-specific events. Numerical simulations on multi-modal satellite and meteorological data from the Kalu River basin, Sri Lanka, show that the HQC-PINN achieves convergence in ~3x fewer training epochs and uses ~44% fewer trainable parameters compared to an equivalent classical PINN, while maintaining competitive classification accuracy. Theoretical analysis indicates that hydrological physics constraints narrow the effective optimization landscape, providing a natural mitigation against barren plateaus in variational quantum circuits. This work establishes the first application of quantum-enhanced physics-informed learning to hydrological prediction and demonstrates a viable path toward quantum advantage in environmental science.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 2 minor

Summary. The manuscript introduces a Hybrid Quantum-Classical Physics-Informed Neural Network (HQC-PINN) that integrates parameterized variational quantum circuits into the PINN framework for hydrological modeling. Multi-source remote sensing features are encoded into quantum states using trainable angle encoding, processed via a hardware-efficient variational ansatz with entangling layers, and constrained by the Saint-Venant shallow water equations and Manning's flow equation as physics loss terms. The approach includes a quantum transfer learning protocol and leverages quantum measurement stochasticity for uncertainty quantification. On data from the Kalu River basin, Sri Lanka, it reports achieving convergence in approximately 3 times fewer training epochs and using 44% fewer trainable parameters than a classical PINN while maintaining competitive accuracy. It further claims that the physics constraints help mitigate barren plateaus in variational quantum circuits.

Significance. Should the efficiency gains and the mechanism for barren plateau mitigation be substantiated, this would constitute a notable contribution as the first application of quantum-enhanced PINNs to hydrological prediction. The inherent UQ without additional Bayesian machinery and the transfer learning protocol are attractive features that could advance quantum machine learning applications in environmental sciences. The work highlights a potential path toward quantum advantage in PDE-constrained learning tasks.

major comments (2)
  1. The assertion in the abstract that 'theoretical analysis indicates that hydrological physics constraints narrow the effective optimization landscape, providing a natural mitigation against barren plateaus in variational quantum circuits' is load-bearing for attributing the reported ~3x convergence improvement and 44% parameter reduction to the quantum component. No derivation, gradient-variance bound, scaling argument, or isolating experiment (such as a plot of Var(∂L/∂θ) versus circuit depth) is provided. Hardware-efficient ansatze are known to exhibit exponentially vanishing gradients; without concrete evidence that the Saint-Venant/Manning residuals alter the Hessian spectrum or keep gradient variance above a threshold, the efficiency claims cannot be confidently linked to the PDE-constrained quantum architecture rather than transfer learning, data specifics, or baseline differences.
  2. The numerical simulation claims of convergence in ~3x fewer training epochs and ~44% fewer trainable parameters (abstract) are presented without details on the equivalent classical PINN baseline architecture (depth, width, activation functions), optimizer settings, data preprocessing for the multi-modal satellite/meteorological inputs, weighting of the PDE residual terms, or error bars from multiple independent runs. This absence undermines verification that the gains are statistically robust and due to the HQC-PINN rather than implementation choices.
minor comments (2)
  1. The acronym HQC-PINN is used in the abstract without an initial definition; ensure it is expanded on first use in the main text.
  2. Clarify in the methods how the quantum circuit output (post-measurement) is interfaced with the classical neural network layers and how the trainable angle encoding scales are optimized jointly with the variational parameters.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and detailed comments, which help clarify the presentation of our results. We respond to each major comment below and indicate the revisions we will make to address the concerns.

read point-by-point responses

  1. Referee: The assertion in the abstract that 'theoretical analysis indicates that hydrological physics constraints narrow the effective optimization landscape, providing a natural mitigation against barren plateaus in variational quantum circuits' is load-bearing for attributing the reported ~3x convergence improvement and 44% parameter reduction to the quantum component. No derivation, gradient-variance bound, scaling argument, or isolating experiment (such as a plot of Var(∂L/∂θ) versus circuit depth) is provided. Hardware-efficient ansatze are known to exhibit exponentially vanishing gradients; without concrete evidence that the Saint-Venant/Manning residuals alter the Hessian spectrum or keep gradient variance above a threshold, the efficiency claims cannot be confidently linked to the PDE-constrained quantum architecture rather than transfer learning, data specifics, or baseline differences.

    Authors: We acknowledge that the current manuscript states the mitigating effect of the physics constraints without including a formal derivation, gradient-variance analysis, or isolating experiment. The statement reflects our interpretation of how the Saint-Venant and Manning residuals supply additional gradient information that structures the loss landscape, but we agree this requires explicit support to substantiate the attribution of efficiency gains. In the revised manuscript we will add a dedicated subsection presenting a scaling argument based on the composite loss and an isolating numerical experiment that compares gradient variance with respect to circuit depth in the presence and absence of the PDE terms. revision: partial

  2. Referee: The numerical simulation claims of convergence in ~3x fewer training epochs and ~44% fewer trainable parameters (abstract) are presented without details on the equivalent classical PINN baseline architecture (depth, width, activation functions), optimizer settings, data preprocessing for the multi-modal satellite/meteorological inputs, weighting of the PDE residual terms, or error bars from multiple independent runs. This absence undermines verification that the gains are statistically robust and due to the HQC-PINN rather than implementation choices.

    Authors: We agree that the experimental details provided are insufficient for independent verification. The revised manuscript will expand the methods and experimental sections to specify the classical PINN architecture (depth, width, activation functions), optimizer settings and learning-rate schedule, the exact preprocessing pipeline applied to the multi-modal inputs, the relative weighting chosen for the data and PDE residual terms, and quantitative results accompanied by standard deviations obtained from multiple independent runs. revision: yes

Circularity Check

0 steps flagged

No significant circularity; performance claims rest on numerical simulations and standard components

full rationale

The reported efficiency gains (~3x fewer epochs, 44% fewer parameters) are obtained from direct numerical simulations on Kalu River data and compared against a classical PINN baseline. The physics loss terms are the standard Saint-Venant and Manning residuals already used in classical PINNs; they are not redefined in terms of the quantum outputs. The barren-plateau mitigation is asserted via an unspecified 'theoretical analysis' without any equation that reduces the claim to a self-definition or fitted parameter. No load-bearing self-citation appears; the UQ mechanism simply invokes the known stochasticity of quantum measurement. The derivation chain is therefore self-contained and does not reduce to its own inputs by construction.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 2 invented entities

The central claim rests on the effectiveness of embedding variational quantum circuits into PINNs with hydrological physics constraints; details on how the ansatz is chosen, how parameters are initialized, and the exact form of the loss terms are not specified in the abstract, leaving several standard domain assumptions unexamined.

free parameters (2)
  • Variational quantum circuit parameters
    Trainable angles and weights in the hardware-efficient ansatz that are optimized jointly with the physics loss.
  • Angle encoding scales
    Parameters controlling how remote-sensing features are mapped to quantum rotation angles.
axioms (2)
  • domain assumption The Saint-Venant shallow water equations and Manning's flow equation can be expressed as differentiable loss terms that meaningfully constrain the network output for real hydrological systems.
    Invoked directly in the architecture description as the physics-informed component.
  • domain assumption Stochasticity from quantum measurement supplies calibrated uncertainty estimates without additional Bayesian machinery.
    Stated as an inherent advantage of the quantum component.
invented entities (2)
  • HQC-PINN architecture no independent evidence
    purpose: Hybrid model integrating parameterized variational quantum circuits with PINN for hydrology
    Newly proposed combination not previously applied to this domain.
  • Quantum transfer learning protocol no independent evidence
    purpose: Pre-training on multi-hazard data followed by fine-tuning on flood events
    Introduced as part of the training strategy.

pith-pipeline@v0.9.0 · 5545 in / 1488 out tokens · 66855 ms · 2026-05-10T17:12:29.273352+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

47 extracted references · 8 canonical work pages · 2 internal anchors

  1. [1]

    The classical network handles the heavy lifting of dimensionality reduction from high- dimensional satellite data to the qubit-compatible representation

  2. [2]

    Only the quantum parameters require optimization during fine-tuning, reducing the training cost proportional top= 2n qLrather than the full classical network size

  3. [3]

    Knowledge from the multi-hazard pre-training provides a better initialization land- scape for the quantum optimizer, further mitigating trainability issues. VII. NUMERICAL EXPERIMENTS A. Dataset and Study Area The study area is the Kalu River basin in Ratnapura District, southwestern Sri Lanka (6.68◦N, 80.39 ◦E), encompassing 2,658 km 2 of terrain subject...

    1987

  4. [4]

    The circuit depth is shallow (L= 3, total depth∼15 for 8 qubits), well within the coherence times of current superconducting processors [30]

  5. [5]

    Comparison with prior quantum approaches to flood/environmental prediction

    The parameter-shift rule is robust to readout errors (they manifest as a multiplica- tive damping factor on the gradient, which can be mitigated by zero-noise extrapola- 20 TABLE VI. Comparison with prior quantum approaches to flood/environmental prediction. Our work is the first to combine quantum circuits with hydrological PDE constraints and multi-moda...

  6. [6]

    Hardware validation on IBM Quantum processors is left to future work

    The physics constraints act as a regularizer that may partially compensate for noise- induced degradation by penalizing unphysical outputs. Hardware validation on IBM Quantum processors is left to future work. C. Comparison with Prior Work Table VI summarizes the landscape. Our work is differentiated by the combination of (i) domain-specific hydrological ...

  7. [7]

    Satellite imaging reveals increased proportion of population exposed to floods,

    B. Tellman, J. A. Sullivan, C. Kuhn, A. J. Kettner, C. S. Doyle, G. R. Brakenridge, T. A. Er- ickson, and D. A. Slayback, “Satellite imaging reveals increased proportion of population exposed to floods,”Nature596, 80–86 (2021)

    2021

  8. [8]

    E. F. Toro,Shock-Capturing Methods for Free-Surface Shallow Flows(John Wiley & Sons, Chichester, 2001)

    2001

  9. [9]

    Flood prediction using machine learning models: Literature review,

    A. Mosavi, P. Ozturk, and K. W. Chau, “Flood prediction using machine learning models: Literature review,”Water10, 1536 (2018)

    2018

  10. [10]

    Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial dif- ferential equations,

    M. Raissi, P. Perdikaris, and G. E. Karniadakis, “Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial dif- ferential equations,”J. Comput. Phys.378, 686–707 (2019)

    2019

  11. [11]

    Characterizing possible failure modes in physics-informed neural networks,

    A. S. Krishnapriyan, A. Gholami, S. Zhe, R. M. Kirby, and M. W. Mahoney, “Characterizing possible failure modes in physics-informed neural networks,” inAdvances in Neural Informa- tion Processing Systems (NeurIPS)(2021), Vol. 34

    2021

  12. [12]

    Variational quantum algorithms,

    M. Cerezo, A. Arrasmith, R. Babbush, S. C. Benjamin, S. Endo, K. Fujii, J. R. McClean, K. Mitarai, X. Yuan, L. Cincio, and P. J. Coles, “Variational quantum algorithms,”Nat. Rev. Phys.3, 625–644 (2021)

    2021

  13. [13]

    Parameterized quantum circuits as machine learning models,

    M. Benedetti, E. Lloyd, S. Sack, and M. Fiorentini, “Parameterized quantum circuits as machine learning models,”Quantum Sci. Technol.4, 043001 (2019)

    2019

  14. [14]

    Supervised learning with quantum-enhanced feature spaces,

    V. Havl´ ıˇ cek, A. D. C´ orcoles, K. Temme, A. W. Harrow, A. Kandala, J. M. Chow, and J. M. Gambetta, “Supervised learning with quantum-enhanced feature spaces,”Nature567, 209–212 (2019)

    2019

  15. [15]

    Schuld and F

    M. Schuld and F. Petruccione,Machine Learning with Quantum Computers, 2nd ed. (Springer, 23 Cham, 2021)

    2021

  16. [16]

    Quantum-Enhanced Convergence of Physics-Informed Neural Networks

    N. Klement, V. Eyring, and M. Schwabe, “Explaining the advantage of quantum-enhanced physics-informed neural networks,” arXiv:2601.15046 (2026)

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  17. [17]

    AQ-PINNs: Attention-enhanced quantum physics-informed neural networks for carbon-efficient climate modeling,

    S. Dutta, N. Innan, S. Ben Yahia, and M. Shafique, “AQ-PINNs: Attention-enhanced quantum physics-informed neural networks for carbon-efficient climate modeling,” arXiv:2409.01626 (2024)

    work page arXiv 2024

  18. [18]

    Flood prediction using classical and quantum machine learning models,

    M. Grzesiak and P. Thakkar, “Flood prediction using classical and quantum machine learning models,” arXiv:2407.01001 (2024)

    work page arXiv 2024

  19. [19]

    Pattnaik et al

    A. Farea, S. Khan, and M. S. Celebi, “QCPINN: Quantum-classical physics-informed neural networks for solving PDEs,” arXiv:2503.16678 (2025)

    work page arXiv 2025

  20. [20]

    Solving nonlinear differential equations with differentiable quantum circuits,

    O. Kyriienko, A. E. Paine, and V. E. Elfving, “Solving nonlinear differential equations with differentiable quantum circuits,”Phys. Rev. A103, 052416 (2021)

    2021

  21. [21]

    Protocols for solving nonlinear ODEs without grid evalua- tion,

    O. Kyriienko and V. E. Elfving, “Protocols for solving nonlinear ODEs without grid evalua- tion,” arXiv:2308.01827 (2023)

    work page arXiv 2023

  22. [22]

    Probabilistic quantum physics-informed neural networks,

    O. Kyriienkoet al., “Probabilistic quantum physics-informed neural networks,”Phys. Rev. Research6, 033291 (2024)

    2024

  23. [23]

    Hybrid quantum physics-informed neural network: Towards efficient learn- ing of high-speed flows,

    S. A. Steinet al., “Hybrid quantum physics-informed neural network: Towards efficient learn- ing of high-speed flows,” arXiv:2503.02202 (2025)

    work page arXiv 2025

  24. [24]

    Transfer learning in hybrid classical-quantum neural networks,

    A. Mari, T. R. Bromley, J. Izaac, M. Schuld, and N. Killoran, “Transfer learning in hybrid classical-quantum neural networks,”Quantum4, 340 (2020)

    2020

  25. [25]

    Satellite image classification with neural quantum kernels,

    P. Rodriguez-Grasa, R. Farzan-Rodriguez, G. Novelli, Y. Ban, and M. Sanz, “Satellite image classification with neural quantum kernels,”Mach. Learn.: Sci. Technol.6, 015043 (2025)

    2025

  26. [26]

    Quantum circuit learning,

    K. Mitarai, M. Negoro, M. Kitagawa, and K. Fujii, “Quantum circuit learning,”Phys. Rev. A98, 032309 (2018)

    2018

  27. [27]

    Evaluating analytic gradients on quantum hardware,

    M. Schuld, V. Bergholm, C. Gogolin, J. Izaac, and N. Killoran, “Evaluating analytic gradients on quantum hardware,”Phys. Rev. A99, 032331 (2019)

    2019

  28. [28]

    PennyLane: Automatic differentiation of hybrid quantum-classical computations

    V. Bergholmet al., “PennyLane: Automatic differentiation of hybrid quantum-classical com- putations,” arXiv:1811.04968 (2018)

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  29. [29]

    Focal loss for dense object detection,

    T. Y. Lin, P. Goyal, R. Girshick, K. He, and P. Doll´ ar, “Focal loss for dense object detection,” inProceedings of the IEEE International Conference on Computer Vision (ICCV)(2017), pp. 2980–2988. 24

    2017

  30. [30]

    V. T. Chow,Open-Channel Hydraulics(McGraw-Hill, New York, 1959)

    1959

  31. [31]

    Dropout as a Bayesian approximation: Representing model uncertainty in deep learning,

    Y. Gal and Z. Ghahramani, “Dropout as a Bayesian approximation: Representing model uncertainty in deep learning,” inProceedings of the 33rd International Conference on Machine Learning (ICML)(2016), pp. 1050–1059

    2016

  32. [32]

    Barren plateaus in quantum neural network training landscapes,

    J. R. McClean, S. Boixo, V. N. Smelyanskiy, R. Babbush, and H. Neven, “Barren plateaus in quantum neural network training landscapes,”Nat. Commun.9, 4812 (2018)

    2018

  33. [33]

    Cost function dependent barren plateaus in shallow parametrized quantum circuits,

    M. Cerezo, A. Sone, T. Volkoff, L. Cincio, and P. J. Coles, “Cost function dependent barren plateaus in shallow parametrized quantum circuits,”Nat. Commun.12, 1791 (2021)

    2021

  34. [34]

    A hybrid AI frame- work for multi-modal flood prediction: Integrating Bayesian neural networks, physics-informed constraints, and multi-task learning,

    P. N. M. Ukwatta Hewage, M. Chakkravarthy, and R. K. Abeysekara, “A hybrid AI frame- work for multi-modal flood prediction: Integrating Bayesian neural networks, physics-informed constraints, and multi-task learning,”Nat. Hazards(2026, submitted)

    2026

  35. [35]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” inProceedings of the 3rd International Conference on Learning Representations (ICLR)(2015)

    2015

  36. [36]

    Evidence for the utility of quantum computing before fault tolerance,

    Y. Kimet al., “Evidence for the utility of quantum computing before fault tolerance,”Nature 618, 500–505 (2023)

    2023

  37. [37]

    Error mitigation for short-depth quantum cir- cuits,

    K. Temme, S. Bravyi, and J. M. Gambetta, “Error mitigation for short-depth quantum cir- cuits,”Phys. Rev. Lett.119, 180509 (2017)

    2017

  38. [38]

    Classification with Quantum Neural Networks on Near Term Processors

    E. Farhi and H. Neven, “Classification with quantum neural networks on near term proces- sors,” arXiv:1802.06002 (2018)

    work page Pith review arXiv 2018

  39. [39]

    A variational eigenvalue solver on a photonic quantum processor,

    A. Peruzzo, J. McClean, P. Shadbolt, M.-H. Yung, X.-Q. Zhou, P. J. Love, A. Aspuru-Guzik, and J. L. O’Brien, “A variational eigenvalue solver on a photonic quantum processor,”Nat. Commun.5, 4213 (2014)

    2014

  40. [40]

    Physics-informed graph neural networks for flood forecasting,

    Y. Liuet al., “Physics-informed graph neural networks for flood forecasting,”J. Hydrol.620, 129376 (2023)

    2023

  41. [41]

    A new application of deep neural network (LSTM) and RUSLE for soil erosion prediction,

    S. Senanayake, B. Pradhan, A. Alamri, and H. J. Park, “A new application of deep neural network (LSTM) and RUSLE for soil erosion prediction,”Geocarto Int.37, 1–23 (2022)

    2022

  42. [42]

    Sentinel-1-based flood mapping: A fully automated processing chain,

    A. Twele, W. Cao, S. Plank, and S. Martinis, “Sentinel-1-based flood mapping: A fully automated processing chain,”Int. J. Remote Sens.37, 2990–3004 (2016)

    2016

  43. [43]

    Classification of remote sensing images with parameterized quantum gates,

    S. Otgonbaatar and M. Datcu, “Classification of remote sensing images with parameterized quantum gates,”IEEE Geosci. Remote Sens. Lett.19, 1–5 (2022)

    2022

  44. [44]

    Quantum 25 support vector machine algorithms for remote sensing data classification,

    A. Delilbasic, G. Cavallaro, M. Willsch, F. Melgani, M. Riedel, and K. Michielsen, “Quantum 25 support vector machine algorithms for remote sensing data classification,” in2021 IEEE International Geoscience and Remote Sensing Symposium (IGARSS)(2021), pp. 2608–2611

    2021

  45. [45]

    The power of quantum neural networks,

    A. Abbas, D. Sutter, C. Zoufal, A. Lucchi, A. Figalli, and S. Woerner, “The power of quantum neural networks,”Nat. Comput. Sci.1, 403–409 (2021)

    2021

  46. [46]

    Quantum physics-informed neural net- works,

    I. Garc´ ıa-Barrenechea, S. Borr` as, and J. Latorre, “Quantum physics-informed neural net- works,”Entropy26, 649 (2024)

    2024

  47. [47]

    Trainable embedding quantum physics informed neural networks,

    H. Tezukaet al., “Trainable embedding quantum physics informed neural networks,”Sci. Rep. 15, 3894 (2025). 26

    2025