Pith. sign in

REVIEW 3 major objections 3 minor 15 references

Fast simulations of continuous-variable circuits using the coherent state decomposition

T0 review · 3 major / 3 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read lcg_plus merges linear combinations of Gaussians with coherent-state decompositions to simulate continuous-variable circuits fast, and derives analytic gradients for optimizing heralded qunaught-state preparation in lossy Gaussian Boson sam

desk verdict The abstract describes a useful-sounding CV simulator library, but the supplied full text is an unrelated robotics paper, so the technical claims are unverifiable from what's in front of us. read the letter →

arxiv 2508.06175 v1 pith:ODLUDUWG submitted 2025-08-08 quant-ph

Olga Solodovnikova , Ulrik L. Andersen , Jonas S. Neergaard-Nielsen This is my paper · ORCID
classification quant-ph
keywords continuous-variablequantumcomputingcoherentstatedecompositionlinearcombinationofGaussiansWignerfunctionsimulationqunaughtGaussianBosonsamplinggeneraldynedetectioncircuitoptimization
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

The paper presents lcg_plus, an open-source Python library for simulating continuous-variable quantum circuits. Its central idea is to write arbitrary non-Gaussian states as a coherent-state decomposition and combine that with a linear-combination-of-Gaussians representation, so the Wigner function can be tracked through Gaussian operations, loss, and generaldyne or photon-number-resolving measurements. The authors claim this is fast and accurate for multi-mode systems, and they derive analytic gradients of state-quality measures so circuit parameters can be optimized by gradient descent. They demonstrate the method by optimizing heralded qunaught-state preparation in a Gaussian Boson sampling circuit that includes inefficient components. If the claim holds, device-realistic continuous-variable circuit design becomes substantially more tractable without a fixed Fock-basis truncation.

What carries the argument

The load-bearing object is the coherent state decomposition: an arbitrary non-Gaussian state is expanded as a superposition of coherent states, whose Wigner functions are Gaussian. Paired with the linear-combination-of-Gaussians representation, this turns the Wigner function into a weighted sum of Gaussian terms; Gaussian channels and measurements map each term to another Gaussian, so the whole state update is a finite mixture update. This machinery is what carries the speed and accuracy claims and also yields the analytic gradients used for optimization.

What would settle it

Take a two-mode Gaussian Boson sampling circuit with known transmission losses, run lcg_plus with an increasing number of coherent-state terms, and compare the computed qunaught-state fidelity against a numerically exact Fock-truncated simulation; if the fidelity error does not shrink toward zero as the term count grows, the coherent-state decomposition is not converging in the regime the paper targets.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that merging the linear combination of Gaussians methodology with the coherent state decomposition of arbitrary non-Gaussian states creates a practical simulation and optimization tool for continuous-variable circuits. By representing the state through its Wigner function as a mixture of Gaussian components, Gaussian channels, photon loss, and generalized (generaldyne and photon-number-resolving) measurements all act locally on the mixture, so multi-mode evolution can be tracked efficiently. The same representation makes quality measures such as fidelity and purity directly computable and differentiable; the paper derives analytical gradients with resp

Load-bearing premise

The method assumes that a modest number of coherent-state terms can represent the non-Gaussian states appearing in multi-mode lossy circuits well enough for accurate simulation and optimization; the abstract does not supply convergence bounds for this truncation.

Editorial extensions

If this is right

  • Quantum circuit designers can optimize continuous-variable state preparation and measurement parameters directly by gradient descent, including against losses, instead of scanning parameters on a grid.
  • Simulations can incorporate non-Gaussian states and photon-number-resolving detectors without committing to a small Fock-basis truncation, where many exact simulations become expensive.
  • Quality measures such as fidelity and purity are available on the fly from the Wigner mixture, making characterization of prepared states straightforward.
  • Heralded non-Gaussian resources such as qunaught states can be optimized in circuits containing realistic inefficient components, bringing fault-tolerant photonic proposals closer to what experimental hardware actually implements.
  • The open-source implementation makes the method immediately usable and extendable by other researchers.

Reading between the lines

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

  • The full text attached to this record is a different manuscript (robot localization with graph neural networks), so the statements above are drawn from the abstract alone; any numerical speed or accuracy figures from the actual library paper need separate verification.
  • Because the method tracks Wigner functions, its performance may degrade for states with strongly negative Wigner regions; testing convergence against exact Fock simulations on such states would clarify the practical domain.
  • The same decomposition could in principle be applied to adaptive or feed-forward continuous-variable circuits, where measurement outcomes change later operations, if the mixture update remains Gaussian; the paper does not claim this, but the mechanism suggests it.
  • A natural benchmark is to compare lcg_plus against a standard Fock-basis simulator on the same lossy Gaussian Boson sampling task; a public error-versus-number-of-coherent-terms curve would make the trade-off concrete.
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

3 major / 3 minor

Summary. The manuscript, arXiv:2508.06175, is ostensibly a quantum-computation paper presenting an open-source Python library called lcg_plus for simulating continuous-variable quantum circuits. The abstract claims that the library merges the linear combination of Gaussians (LCG) methodology with a coherent state decomposition (CSD), tracks the Wigner function, supports generaldyne and photon-number-resolving detection, and provides analytical gradients for parameterized circuit elements. It reports a demonstration optimizing the heralded preparation of a qunaught state in a Gaussian Boson sampling circuit with inefficient components. However, the full text supplied with the manuscript is a different paper entirely: arXiv:2508.06177, 'Graph-based Robot Localization Using a Graph Neural Network with a Floor Camera and a Feature Rich Industrial Floor.' None of the technical content described in the abstract—no derivations, no convergence analysis, no numerical benchmarks, no gradient formulas, no simulation results—appears anywhere in the submitted manuscript.

Significance. If the claims in the abstract are correct, lcg_plus could be a useful contribution to continuous-variable quantum simulation, particularly for non-Gaussian states and detectors with realistic inefficiencies. The combination of LCG and CSD, with analytical gradients, is potentially valuable for variational optimization in photonic quantum computing. However, the significance cannot be assessed from the submitted manuscript because the full text is not the claimed paper. There is no way to verify the convergence properties of the coherent state decomposition, the correctness of the Wigner-function update rules, the validity of the gradient derivation, or the performance of the qunaught-state optimization. The abstract alone provides insufficient evidence for the central claims of speed, accuracy, and generality.

major comments (3)
  1. [Full Text (entire manuscript after abstract)] The full text of the submitted manuscript is arXiv:2508.06177, a robotics paper on graph-based robot localization, and has no connection to continuous-variable quantum circuits, lcg_plus, or the abstract. This is a load-bearing internal inconsistency: the manuscript does not contain any of the claimed technical contributions. The Wigner-function update rules, the coherent state decomposition convergence properties, the analytical gradient formulas, and the numerical results are all absent. The central claim of the paper is therefore entirely unverifiable from the submitted text.
  2. [Abstract, paragraph 1] The abstract states that the framework is 'fast and accurate' and 'convenient,' but gives no quantitative metrics, no comparison to existing simulation methods (e.g., exact Fock-space simulation or phase-space methods), no error bounds, and no complexity scaling. 'Fast and accurate' is an unsupported qualitative claim. A reader cannot evaluate whether the truncation of the coherent state decomposition is controlled or whether the method scales tractably to multi-mode systems, which is the core computational assumption of the approach.
  3. [Abstract, paragraph 2] The claimed demonstration—optimizing heralded qunaught state preparation in a Gaussian Boson sampling circuit with inefficient components—is mentioned only in one sentence. No results, parameters, convergence data, or comparison to baselines are provided. This is not a demonstration that can be assessed; it is an assertion. Since the full text is unrelated, there is no supporting numerical evidence anywhere in the manuscript.
minor comments (3)
  1. [Abstract, paragraph 1] The term 'generaldyne' appears to be a typo; the standard spelling is 'general-dyne' or 'generalized dyne' (often referring to generalized heterodyne measurement). Please verify terminology.
  2. [Abstract, paragraph 2] The phrase 'qunaught state' is unusual. If this is a portmanteau of 'quantum' and 'naught,' it may be field-specific, but it is not defined in the abstract. Consider adding a reference or definition.
  3. [General] The abstract refers to an 'open-source Python library' but provides no repository URL, version number, or installation instructions. If the library is a central contribution, these should be included in the manuscript.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity detectable; the quantum-simulation abstract makes no claim that reduces to its inputs, and the supplied full text is a different paper.

full rationale

The supplied full text is arXiv:2508.06177, a robot-localization paper, not the quantum-simulation manuscript arXiv:2508.06175. Consequently, the only in-scope content from the quantum paper is its abstract. That abstract presents lcg_plus as a library that merges linear combination of Gaussians with coherent state decomposition, tracks the Wigner function, derives analytical gradients, and demonstrates optimization of qunaught-state preparation. No derivation chain is visible in the abstract, so there is no equation I can quote to show that any output is defined in terms of an input, that a fitted parameter is renamed as a prediction, or that a load-bearing premise rests solely on a self-citation. The claimed coherent-state decomposition of arbitrary non-Gaussian states is a representational assumption, but it is not asserted to be derived from the target conclusion; the optimization example is an application of the method, not a benchmark defined by the method's own outputs. The mismatch between the claimed paper and the supplied full text is a serious verifiability failure, and the absence of convergence analysis, gradient derivations, and numerical details is a correctness/evidence gap, but per the hard rules circularity may only be found when a specific reduction is exhibited. None can be exhibited here. Score 0 is therefore the honest, proportionate finding.

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

No free parameters or invented entities are identified from the abstract. The main axiomatic assumptions are standard for continuous-variable quantum simulations: the validity of the Wigner function representation and the convergence of the coherent state decomposition. The abstract gives no evidence to verify these assumptions.

assumptions (2)
  • domain assumption The Wigner function is a valid and sufficient representation for the simulated continuous-variable states.
    The abstract states the framework tracks the Wigner function; this assumes that the Wigner function is non-negative or that the simulation can handle negative regions appropriately for all states considered.
  • domain assumption Coherent state decomposition can represent arbitrary non-Gaussian states with a tractable number of terms.
    The abstract claims the method merges LCG with coherent state decomposition; this requires that the decomposition converges to accurate results for multi-mode systems, which is not demonstrated in the abstract.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Fast simulations of continuous-variable circuits using the coherent state decomposition." pith.science (2026). https://pith.science/paper/ODLUDUWG

@misc{pith2026250806175,
  author       = {Pith},
  title        = {Pith review of: Fast simulations of continuous-variable circuits using the coherent state decomposition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ODLUDUWG}},
  note         = {Machine review of arXiv:2508.06175}
}
read the original abstract

We present \texttt{lcg\_plus}, an open-source Python library for the simulation of continuous-variable quantum circuits with both generaldyne and photon-number-resolving detector capabilities. Our framework merges the linear combination of Gaussians methodology with the coherent state decomposition of arbitrary non-Gaussian states, forming a bridge between the Gaussian and Fock basis representations. By tracking the Wigner function, we can simulate the action of Gaussian channels and measurements on multi-mode systems in a fast and accurate numerical framework. The calculation of the quality measures of quantum states is convenient in this formalism, and we derive expressions for the analytical gradients of these measures with respect to parameterized circuit elements. We demonstrate the utility of this methodology by optimizing the heralded preparation of a qunaught state, a crucial component for building a fault-tolerant photonic quantum computer, with a Gaussian Boson sampling circuit containing inefficient components.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

15 extracted references · 11 canonical work pages

  1. [1]

    In: Proceedings of the twelfth ACM international conference on web search and data mining

    Bai, Y., Ding, H., Bian, S., Chen, T., Sun, Y., Wang, W.: Simgnn: A neural network approach to fast graph similarity computation. In: Proceedings of the twelfth ACM international conference on web search and data mining. pp. 384–392 (2019)

    work page 2019

  2. [2]

    Brömmel, P., Brämer, D., Urbann, O., Kleingarn, D.: Robot localization using a learned keypoint detector and descriptor with a floor camera and a feature rich industrial floor (2025),https://arxiv.org/abs/2504.03249

    work page Pith review arXiv 2025

  3. [3]

    Artificial neural networks pp

    Chicco, D.: Siamese neural networks: An overview. Artificial neural networks pp. 73–94 (2021)

    work page 2021

  4. [4]

    In: 2005 IEEE computer society conference on computer vision and pattern recognition (CVPR’05)

    Chopra, S., Hadsell, R., LeCun, Y.: Learning a similarity metric discriminatively, with application to face verification. In: 2005 IEEE computer society conference on computer vision and pattern recognition (CVPR’05). vol. 1, pp. 539–546. IEEE (2005)

    2005

  5. [5]

    Diplomová práce

    Dubrofsky, E.: Homography estimation. Diplomová práce. Vancouver: Univerzita Britské Kolumbie 5 (2009)

    work page 2009

  6. [6]

    Advances in neural information processing systems33, 18661–18673 (2020)

    Khosla, P., Teterwak, P., Wang, C., Sarna, A., Tian, Y., Isola, P., Maschinot, A., Liu, C., Krishnan, D.: Supervised contrastive learning. Advances in neural information processing systems33, 18661–18673 (2020)

    2020

  7. [7]

    arXiv preprint arXiv:1609.02907 (2016)

    Kipf, T.N., Welling, M.: Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907 (2016)

    arXiv 2016

  8. [8]

    In: 2016 IEEE/ION Position, Location and Navigation Sym- posium (PLANS)

    Kozak, K., Alban, M.: Ranger: A ground-facing camera-based localization system for ground vehicles. In: 2016 IEEE/ION Position, Location and Navigation Sym- posium (PLANS). pp. 170–178. IEEE (2016)

    work page 2016

Show all 15 references
  1. [9]

    arXiv preprint arXiv:2002.11948 (2020)

    Schmid, J.F., Simon, S.F., Mester, R.: Features for ground texture based localization–a survey. arXiv preprint arXiv:2002.11948 (2020)

    2002 arXiv

  2. [10]

    ACM Transactions on Database Systems (TODS) 42(3), 1–21 (2017)

    Schubert, E., Sander, J., Ester, M., Kriegel, H.P., Xu, X.: Dbscan revisited, revis- ited: why and how you should (still) use dbscan. ACM Transactions on Database Systems (TODS) 42(3), 1–21 (2017)

    2017

  3. [11]

    International Journal of Computer Vision130(6), 1526–1565 (2022)

    Soviany, P., Ionescu, R.T., Rota, P., Sebe, N.: Curriculum learning: A survey. International Journal of Computer Vision130(6), 1526–1565 (2022)

    2022

  4. [12]

    Expert Syst

    Wang, Q., Pan, Z., Hou, J., Yu, L.: High-precision offline mapping and localiza- tion system based on ground texture with binary descriptors. Expert Syst. Appl. 240, 122650 (2024). https://doi.org/10.1016/J.ESWA.2023.122650, https:// doi.org/10.1016/j.eswa.2023.122650

    2024

  5. [13]

    In: 2015 IEEE Interna- tional Conference on Robotics and Biomimetics (ROBIO)

    Zhang, H., Zhang, C., Yang, W., Chen, C.Y.: Localization and navigation us- ing qr code for mobile robot in indoor environment. In: 2015 IEEE Interna- tional Conference on Robotics and Biomimetics (ROBIO). pp. 2501–2506 (2015). https://doi.org/10.1109/ROBIO.2015.7419715

    2015

  6. [14]

    In: 2019 International Conference on Robotics and Automation (ICRA)

    Zhang, L., Finkelstein, A., Rusinkiewicz, S.: High-precision localization using ground texture. In: 2019 International Conference on Robotics and Automation (ICRA). pp. 6381–6387. IEEE (2019)

    2019

  7. [15]

    In: Proceedings of the IEEE conference on computer vision and pattern recognition

    Zhang, L., Rusinkiewicz, S.: Learning to detect features in texture images. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 6325–6333 (2018)

    2018

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.