pith. sign in

arxiv: 2301.03995 · v2 · submitted 2023-01-10 · ✦ hep-lat

Efficiently unquenching QCD+QED at O(α)

Tim Harris , Vera G\"ulpers , Antonin Portelli , James Richings This is my paper

Pith reviewed 2026-05-24 09:37 UTC · model grok-4.3

classification ✦ hep-lat
keywords lattice QCDQCD+QEDisospin breakingdisconnected diagramsstochastic estimatorsvariance reductionelectromagnetic correctionsdomain wall fermions
0
0 comments X

The pith

Variance analysis of stochastic estimators for quark propagator traces reduces computational cost for sea-quark electromagnetic effects in lattice QCD+QED.

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

The paper presents a strategy for including the electromagnetic interactions of sea quarks when simulating QCD+QED at leading order in the fine-structure constant. Sea-quark charges generate quark-line disconnected diagrams that enter isospin-breaking corrections and are expensive to evaluate precisely with stochastic methods. The authors show that the variance of the relevant traces depends on flavour combinations and space-time decompositions, and that certain choices yield appreciably lower variance. They demonstrate the effect with numerical measurements on an existing Nf=2+1 domain-wall ensemble. If the variance reduction carries over to the full set of diagrams, the approach makes unquenched electromagnetic corrections more practical without changing the underlying theory.

Core claim

An analysis of the variance of stochastic estimators for the relevant traces of quark propagators improves the situation for certain flavour combinations and space-time decompositions when computing isospin breaking corrections at leading order in the electromagnetic coupling.

What carries the argument

Variance analysis of stochastic estimators for traces of quark propagators, used to select lower-variance flavour combinations and space-time decompositions for disconnected diagrams.

If this is right

  • Certain flavour combinations produce lower variance in the disconnected contributions.
  • Specific space-time decompositions can be chosen to further suppress the variance.
  • The variance reduction is measurable on existing Nf=2+1 domain-wall ensembles.
  • The same estimator choices apply directly to any hadronic quantity whose leading electromagnetic correction involves sea-quark charges.

Where Pith is reading between the lines

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

  • The same variance-ranking procedure could be repeated on other lattice actions or volumes to identify action-specific optimal decompositions.
  • If the low-variance choices remain stable across quark masses, the method may extend to calculations that require multiple sea-quark mass points.
  • Combining the variance reduction with existing noise-reduction techniques such as dilution or hierarchical probing might compound the efficiency gain.

Load-bearing premise

Lower variance identified for selected flavour combinations and decompositions will translate into a net gain in precision or efficiency for the complete collection of disconnected diagrams without introducing bias or prohibitive extra cost.

What would settle it

A head-to-head comparison, at fixed computational budget, of the final statistical error on a representative isospin-breaking observable when the full set of diagrams is evaluated with the high-variance versus low-variance estimator choices.

Figures

Figures reproduced from arXiv: 2301.03995 by Antonin Portelli, James Richings, Tim Harris, Vera G\"ulpers.

Figure 1
Figure 1. Figure 1: Wick contractions which appear at leading order in the expansion of a hadronic observable 𝑂 in the electromagnetic coupling. Each closed fermion line has contributions from all of the quark flavours u, d,s, . . . with the appropriate charge factors. conversely the iso-spin breaking corrections themselves). Otherwise the predictions of QCD+QED are unambiguously defined, up to its intrinsic accuracy, by fixi… view at source ↗
Figure 2
Figure 2. Figure 2: Left: Comparison of the variance versus the number of sources for the 𝑊1 quark-line disconnected diagram, using a single flavour (red squares), the standard estimator for u, d,s flavours (blue circles) and the split-even estimator (green triangles). The dashed line shows 1/𝑁 2 s scaling. In this figure, the (local) currents are not renormalized and the charge factors are not included. standard estimators f… view at source ↗
Figure 3
Figure 3. Figure 3: Left: the variance for the stochastic estimator (red squares) and point source estimator (blue circles) for the minimum number of inversions required, for the contribution with fixed separation between the currents |𝑟|. The green triangle indicates the gauge variance for the point 𝑟 = 0. Right: the variance for the short-distance (red squares) and long-distance (blue circles) for the choice 𝑅/𝑎 = 4, versus… view at source ↗
read the original abstract

We outline a strategy to efficiently include the electromagnetic interactions of the sea quarks in QCD+QED. When computing iso-spin breaking corrections to hadronic quantities at leading order in the electromagnetic coupling, the sea-quark charges result in quark-line disconnected diagrams which are challenging to compute precisely. An analysis of the variance of stochastic estimators for the relevant traces of quark propagators helps us to improve the situation for certain flavour combinations and space-time decompositions. We present preliminary numerical results for the variances of the corresponding contributions using an ensemble of $N_\mathrm{f}=2+1$ domain-wall fermions generated by the RBC/UKQCD collaboration.

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 outlines a strategy for efficiently including sea-quark electromagnetic interactions in QCD+QED at O(α) when computing iso-spin breaking corrections. It analyzes the variance of stochastic estimators for traces of quark propagators to identify lower-variance flavour combinations and space-time decompositions for the resulting disconnected diagrams, and presents preliminary numerical results on variances using a single Nf=2+1 domain-wall fermion ensemble from RBC/UKQCD.

Significance. If the variance reductions translate into net gains in precision or cost for the full set of diagrams, the work would be useful for lattice calculations of electromagnetic corrections, where disconnected contributions are a known bottleneck. The approach builds on existing stochastic methods without introducing new parameters.

major comments (2)
  1. [Numerical results] Numerical results section: variances are reported only for individual flavour combinations and space-time decompositions on a single ensemble; the manuscript does not show the total variance or effective error of the linear combination of all required disconnected diagrams that enters a physical iso-spin-breaking observable (e.g., the full set needed for a meson mass splitting).
  2. [Strategy outline] Strategy outline: while lower variance is demonstrated for selected subsets, there is no explicit accounting for the overhead of the decomposition or the contribution of any remaining high-variance terms to the summed estimator, so the claim of practical efficiency improvement for the complete calculation is not yet demonstrated.
minor comments (2)
  1. [Abstract] Abstract and introduction: the phrase 'helps us to improve the situation' is vague; replace with a quantitative statement of what is improved (e.g., variance reduction factor for specific terms).
  2. [Figures/Tables] Figure captions and table labels should explicitly state the ensemble parameters (volume, spacing, quark masses) and the number of stochastic sources used for each variance estimate.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their detailed review and constructive comments. Our manuscript presents a variance analysis of stochastic estimators for sea-quark electromagnetic contributions at O(α) together with preliminary numerical results on individual flavour and space-time decompositions. We address the two major comments below.

read point-by-point responses

  1. Referee: Numerical results section: variances are reported only for individual flavour combinations and space-time decompositions on a single ensemble; the manuscript does not show the total variance or effective error of the linear combination of all required disconnected diagrams that enters a physical iso-spin-breaking observable (e.g., the full set needed for a meson mass splitting).

    Authors: We agree that the total variance of the full linear combination entering a physical observable is not computed or shown. The scope of this work is to analyse the variances of the individual contributions in order to identify lower-variance flavour and space-time combinations; constructing the complete estimator for a specific observable (including all cross-correlations) lies outside the present preliminary study. revision: no

  2. Referee: Strategy outline: while lower variance is demonstrated for selected subsets, there is no explicit accounting for the overhead of the decomposition or the contribution of any remaining high-variance terms to the summed estimator, so the claim of practical efficiency improvement for the complete calculation is not yet demonstrated.

    Authors: The manuscript does not claim a net efficiency gain for a complete physical calculation. It demonstrates that certain flavour and space-time decompositions yield substantially lower variance than the naive estimator. The proposed decomposition re-uses the same stochastic sources and therefore incurs no additional overhead; any remaining high-variance terms can be treated with the original estimator or with increased statistics as needed. A full cost-benefit analysis for a specific observable would require further dedicated work. revision: no

Circularity Check

0 steps flagged

No circularity; variance analysis is an independent numerical study of existing estimators

full rationale

The paper outlines a computational strategy for handling disconnected diagrams in QCD+QED at O(α) by analyzing variances of stochastic estimators for quark propagator traces. No equations, derivations, or predictions are presented that reduce to self-definitions, fitted inputs renamed as predictions, or load-bearing self-citations. Preliminary results are reported directly from an Nf=2+1 ensemble without any chain that equates outputs to inputs by construction. The central claim rests on external numerical evidence rather than internal redefinition.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No free parameters, axioms, or invented entities are identifiable from the abstract alone.

pith-pipeline@v0.9.0 · 5636 in / 1154 out tokens · 24106 ms · 2026-05-24T09:37:10.927710+00:00 · methodology

discussion (0)

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

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Variance reduction strategies for lattice QCD

    hep-lat 2026-05 unverdicted novelty 2.0

    Variance reduction schemes based on decompositions of quark propagators have proven useful for precision lattice QCD observables and may help reduce the computational cost of reaching large volumes.

Reference graph

Works this paper leans on

24 extracted references · 24 canonical work pages · cited by 1 Pith paper

  1. [1]

    Aoyama et al

    T. Aoyama et al. In:Phys. Rept.887 (2020), pp. 1–166. arXiv:2/zero.alt3/zero.alt36./zero.alt34822 [hep-ph]

    work page 2020

  2. [2]

    Aoki et al

    Y. Aoki et al. In:Eur. Phys. J. C82.10 (2022), p. 869. arXiv:2111./zero.alt39849 [hep-lat]

    work page 2022

  3. [3]

    Aoki et al

    S. Aoki et al. In:Phys. Rev. D86 (2012), p. 034507. arXiv:12/zero.alt35.2961 [hep-lat]

    work page 2012

  4. [4]

    Ishikawa et al

    T. Ishikawa et al. In:Phys. Rev. Lett.109 (7 Aug. 2012), p. 072002

    work page 2012

  5. [5]

    Horsley et al

    R. Horsley et al. In:J. Phys. G46 (2019), p. 115004. arXiv:19/zero.alt34./zero.alt323/zero.alt34 [hep-lat]

    work page 2019

  6. [6]

    Bushnaq et al

    L. Bushnaq et al. In: (Sept. 2022). arXiv:22/zero.alt39.13183 [hep-lat]

    work page 2022

  7. [7]

    Borsanyi et al

    S. Borsanyi et al. In:Nature593.7857 (2021), pp. 51–55. arXiv:2/zero.alt3/zero.alt32.12347 [hep-lat]

    work page 2021

  8. [8]

    G. M. de Divitiis et al. In:JHEP04 (2012), p. 124. arXiv:111/zero.alt3.6294 [hep-lat]

    work page 2012

  9. [9]

    G. M. de Divitiis et al. In: Phys. Rev. D87.11 (2013), p. 114505. arXiv: 13/zero.alt33 . 4896 [hep-lat]

    work page 2013

  10. [10]

    Giusti et al

    L. Giusti et al. In:Eur. Phys. J. C79.7 (2019), p. 586. arXiv:19/zero.alt33.1/zero.alt3447 [hep-lat]

    work page 2019

  11. [11]

    Capitani et al

    S. Capitani et al. In:Nucl. Phys. B593 (2001), pp. 183–228. arXiv:hep-lat//zero.alt3/zero.alt3/zero.alt37/zero.alt3/zero.alt34

    work page 2001

  12. [12]

    Blum et al

    T. Blum et al. In:Phys. Rev. D93.7 (2016), p. 074505. arXiv:1411.7/zero.alt317 [hep-lat]

    work page 2016

  13. [13]

    San José et al

    T. San José et al. In:JHEP 08 (2022), p. 220. arXiv:22/zero.alt33./zero.alt38676 [hep-lat]

    work page 2022

  14. [14]

    Chao et al

    E.-H. Chao et al. In:Eur. Phys. J. C81.7 (2021), p. 651. arXiv:21/zero.alt34./zero.alt32632 [hep-lat]

    work page 2021

  15. [15]

    Salg et al

    M. Salg et al. In:39th International Symposium on Lattice Field Theory. Nov. 2022. arXiv: 2211.17/zero.alt349 [hep-lat]

    work page 2022

  16. [16]

    Boucaud et al

    P. Boucaud et al. In:Comput. Phys. Commun.179 (2008), pp. 695–715. arXiv:/zero.alt38/zero.alt33./zero.alt3224 [hep-lat]

    work page 2008

  17. [17]

    Allton et al

    C. Allton et al. In:Phys. Rev. D78 (2008), p. 114509. arXiv:/zero.alt38/zero.alt34./zero.alt3473 [hep-lat]

    work page 2008

  18. [18]

    120(2008),pp.413–441.arXiv: /zero.alt38/zero.alt34.2/zero.alt344 [hep-ph]

    M.HayakawaandS.Uno.In: Prog.Theor.Phys. 120(2008),pp.413–441.arXiv: /zero.alt38/zero.alt34.2/zero.alt344 [hep-ph]

    work page 2008

  19. [19]

    G.M.deDivitiisetal.In: Phys.Lett.B 382(1996),pp.393–397.arXiv: hep-lat/96/zero.alt33/zero.alt32/zero.alt3

    work page 1996

  20. [20]

    Li et al

    Y. Li et al. In:Phys. Rev. D103.1 (2021), p. 014514. arXiv:2/zero.alt3/zero.alt39./zero.alt31/zero.alt329 [hep-lat]

    work page 2021

  21. [21]

    124.13(2020),p.132002.arXiv: 1911./zero.alt38123 [hep-lat]

    T.Blumetal.In: Phys.Rev.Lett. 124.13(2020),p.132002.arXiv: 1911./zero.alt38123 [hep-lat]

    work page 2020

  22. [22]

    Boyle et al

    P. Boyle et al. In: (Nov. 2022). arXiv:2211.12865 [hep-lat]

    work page arXiv 2022

  23. [23]

    Boyle et al

    P. Boyle et al. In:39th International Symposium on Lattice Field Theory. Dec. 2022. arXiv: 2212./zero.alt347/zero.alt39 [hep-lat]

    work page 2022

  24. [24]

    P. A. Boyle et al. In:PoSLATTICE2015 (2016), p. 023. 9

    work page 2016