pith. machine review for the scientific record. sign in

arxiv: 2605.12915 · v1 · submitted 2026-05-13 · 🧮 math.NA · cs.NA· physics.flu-dyn

Recognition: 1 theorem link

· Lean Theorem

Fully Discrete Active Flux Method based on Transported Acoustic Increments for the Compressible Euler Equations

Karthik Duraisamy

Authors on Pith no claims yet

Pith reviewed 2026-05-14 19:03 UTC · model grok-4.3

classification 🧮 math.NA cs.NAphysics.flu-dyn
keywords Active Flux methodCompressible Euler equationsTransported acoustic incrementsUnsplit evolutionFinite-volume methodsHyperbolic conservation lawsHigh-order numerical methods
0
0 comments X

The pith

Reconstructing acoustic increments as cellwise quadratic fields and evaluating them at convective feet yields the exact unsplit frozen evolution for the compressible Euler equations.

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

The method updates cell averages through unsplit flux quadrature while evolving point values by transporting a reconstructed acoustic increment to the convective foot instead of adding separate acoustic and advective steps. For constant coefficients this transport produces the exact composition of the two operators whenever they commute, removing the split defect that otherwise appears in additive formulations. The same reconstruction preserves the compact stencil, the conservative one-stage average update, and the exact locally linearized acoustic operator. Tests on isentropic vortex convection recover third-order accuracy with smaller error constants and a wider stable CFL range, while low-Mach and Kelvin-Helmholtz runs show coherent structures and low dissipation without limiters.

Core claim

The paper shows that the acoustic increment, reconstructed as a cellwise Q2 field and evaluated at the convective foot of each target point, produces a point update that reduces exactly to the transported composition of acoustic and advective generators for frozen coefficients. When those generators commute, the update therefore coincides with the true unsplit frozen evolution operator, eliminating the additive-split error that remains in earlier Active Flux formulations.

What carries the argument

The transported acoustic increment: a cellwise quadratic (Q2) reconstruction of the acoustic update that is evaluated at the foot of the characteristic trajectory and then composed with the advective operator.

If this is right

  • Third-order point accuracy is recovered for the transported update, while the additive version remains only second-order.
  • Isentropic vortex convection tests exhibit third-order convergence, reduced error constants, and an enlarged empirical CFL range.
  • Nonlinear Gaussian acoustic pulses preserve radial symmetry with near-third-order decay of symmetry error.
  • Low-Mach shear layers produce coherent vorticity and ultra-low entropy dissipation without secondary vortices on coarse grids.
  • Under-resolved compressible Kelvin-Helmholtz evolution remains stable without limiters and shows consistent entropy dissipation to late times.

Where Pith is reading between the lines

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

  • The same transport construction could be applied to other operator-split hyperbolic schemes to remove splitting errors without enlarging the stencil.
  • Because the acoustic operator is kept exact at the linear level, the method may allow larger time steps in low-Mach regimes before acoustic stiffness appears.
  • Extending the Q2 reconstruction to higher even degrees would test whether the observed third-order point accuracy can be raised systematically while retaining the compact stencil.

Load-bearing premise

Reconstructing the acoustic increment as a cellwise Q2 field and evaluating it at the convective foot accurately captures the unsplit evolution operator even for the full nonlinear Euler equations.

What would settle it

A direct numerical check on a linear frozen system with non-commuting acoustic and advective operators would show whether the transported point update remains exact; if the error fails to drop to machine zero, the exact-unsplit claim is false.

Figures

Figures reproduced from arXiv: 2605.12915 by Karthik Duraisamy.

Figure 1
Figure 1. Figure 1: Schematic of the transported-increment. The dashed red circle is the original acoustic disk centered at the grid node 𝑃 . That acoustic evolution is naturally associated with the material label that started at 𝑃 , which is located at 𝑃 +𝒖0 𝜏 after time 𝜏. To obtain the Eulerian value at the fixed node 𝑃 at time 𝜏, one instead needs acoustic information associated with the convective foot 𝑃𝑓 , represented h… view at source ↗
Figure 2
Figure 2. Figure 2: Mixed packet, 562 cells: primitive method-minus-exact error contours, scaled by the packet amplitude 𝜖 = 10−6 . K. Duraisamy: Preprint submitted to Elsevier Page 14 of 31 [PITH_FULL_IMAGE:figures/full_fig_p014_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Mixed packet, 562 cells: centerline primitive error profiles. Purple: RB, Green: RB-TAI [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Isentropic vortex convection: Average density error (L1) after 1 period: 322 , 642 ,1282 grids. K. Duraisamy: Preprint submitted to Elsevier Page 15 of 31 [PITH_FULL_IMAGE:figures/full_fig_p015_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Isentropic vortex convection: Average density error (L1) after 1 period with transported method: 162 , 322 , 642 grids. K. Duraisamy: Preprint submitted to Elsevier Page 16 of 31 [PITH_FULL_IMAGE:figures/full_fig_p016_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Evolution of Gaussian Pulse at t=2.5 at Δ𝑥 = 0.1, 0.05, 0.025, 0.0125 with RB-TAI. Cell average values shown. K. Duraisamy: Preprint submitted to Elsevier Page 17 of 31 [PITH_FULL_IMAGE:figures/full_fig_p017_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Evolution of Gaussian Pulse at t=2.5. Black: Δ𝑥 = 0.1; Blue: Δ𝑥 = 0.05; Red: Δ𝑥 = 0.025; Magenta: Δ𝑥 = 0.0125. Nodal values shown. K. Duraisamy: Preprint submitted to Elsevier Page 18 of 31 [PITH_FULL_IMAGE:figures/full_fig_p018_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Maximum error and asymmetry in pressure for Gaussian Pulse at t=2.5 [PITH_FULL_IMAGE:figures/full_fig_p019_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Zoomed in view of Gaussian pulse evolution at t=2.5, Δ𝑥 = 0.1. Blue: Semi Discrete; Black: Discrete (RB-TAI); Red: Discrete (RB); Magenta: Reference. K. Duraisamy: Preprint submitted to Elsevier Page 19 of 31 [PITH_FULL_IMAGE:figures/full_fig_p019_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Density contours for shear problem on 64 × 32 and 256 × 128 grids using Discrete (RB-TAI) [PITH_FULL_IMAGE:figures/full_fig_p020_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Shear Problem on 64 × 32 grids. Top: Discrete (RB-TAI), Bottom: DG (P1) K. Duraisamy: Preprint submitted to Elsevier Page 20 of 31 [PITH_FULL_IMAGE:figures/full_fig_p020_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Normalized integral quantities. Red: 64 × 32; Blue: 128 × 64; Black: 256 × 128. Solid lines: Discrete (RB-TAI); Dashed lines: Semi Discrete [PITH_FULL_IMAGE:figures/full_fig_p022_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Under-resolved KH problem on 2562 mesh, Discrete RB-TAI 𝜌(𝑥, 𝑦, 0) = 1 2 + 3 4 𝐵(𝑦), 𝑢(𝑥, 𝑦, 0) = 1 2 ( 𝐵(𝑦) − 1) , 𝑣(𝑥, 𝑦, 0) = 1 10 sin(2𝜋𝑥), 𝑝(𝑥, 𝑦, 0) = 1. The goal is to evolve the solution to 𝑡 = 15. Most of the numerical methods in the above publications (Glaubitz et al., 2025; Ranocha et al., 2025; Bercik et al., 2026) fail by 𝑡 < 5 units. All three active flux variants spanning under-resolved mes… view at source ↗
Figure 14
Figure 14. Figure 14: Normalized Entropy. Red: 642 ; Blue: 1282 ; Black: 2562 . Solid lines: Discrete (RB-TAI); Dashed lines: Semi Discrete [PITH_FULL_IMAGE:figures/full_fig_p023_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Density contours for under-resolved KH problem on 642 mesh. Left: Discrete RB, Right: Discrete RB-TAI It is possible to write the pressure update in the following form (Appendix B). 𝑝 𝑛+1 ac (𝑃 ) = 𝑃𝑅 + 𝑃𝐿. Set 𝜉 ≜ 𝑘𝑥ℎ, 𝜂 ≜ 𝑘𝑦ℎ, 𝜈 ≜ 𝑐0Δ𝑡 ℎ , 𝑍0 ≜ 𝜌0 𝑐0 . The scalar pressure diagnostic is obtained by substituting a single linearized acoustic Fourier branch. Let 𝜅 ≜ 𝐾ℎ = √ 𝜉 2 + 𝜂 2, ̂𝑘𝑥 ≜ 𝜉 𝜅 , ̂𝑘𝑦 ≜ 𝜂 𝜅 ,… view at source ↗
Figure 16
Figure 16. Figure 16: Complex-plane locus of the intended-family vertical-edge symbol 𝐺𝑗𝑗 as the wave angle varies. Dashed curves are the exact diagonal entries; solid curves are the Discrete (RB) vertical-edge point symbol [PITH_FULL_IMAGE:figures/full_fig_p032_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Zoomed complex-plane comparison of the exact convected acoustic pressure gain for a pure 𝑠 = +1 acoustic input for 𝜈 = 0.25. K. Duraisamy: Preprint submitted to Elsevier Page 32 of 31 [PITH_FULL_IMAGE:figures/full_fig_p032_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: Polar plot of the scalar pressure-gain error as the wave-vector angle is swept at fixed 𝐾ℎfor a pure 𝑠 = 1 acoustic input for 𝜈 = 0.25. K. Duraisamy: Preprint submitted to Elsevier Page 33 of 31 [PITH_FULL_IMAGE:figures/full_fig_p033_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Polar plot of the scalar pressure-gain amplification for a pure 𝑠 = 1 acoustic input for 𝜈 = 0.5. K. Duraisamy: Preprint submitted to Elsevier Page 34 of 31 [PITH_FULL_IMAGE:figures/full_fig_p034_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: Shear Problem on 64 × 32 and 256 × 128 mesh, Discrete (RB-TAI) [PITH_FULL_IMAGE:figures/full_fig_p035_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: Density and Vorticity contours in shear problem with P1 Discontinuous Galerkin on 256 × 128 grid. K. Duraisamy: Preprint submitted to Elsevier Page 35 of 31 [PITH_FULL_IMAGE:figures/full_fig_p035_21.png] view at source ↗
Figure 22
Figure 22. Figure 22: Vorticity contours in shear problem with 5th order Continuous Galerkin: 32 × 16; 64 × 32; 128 × 64 grids. K. Duraisamy: Preprint submitted to Elsevier Page 36 of 31 [PITH_FULL_IMAGE:figures/full_fig_p036_22.png] view at source ↗
Figure 23
Figure 23. Figure 23: Vorticity contours for under-resolved KH problem. Rows: 642 , 1282 , 2562 mesh. Columns: Discrete RB, Discrete RB-TAI K. Duraisamy: Preprint submitted to Elsevier Page 37 of 31 [PITH_FULL_IMAGE:figures/full_fig_p037_23.png] view at source ↗
read the original abstract

A fully discrete Active Flux method is proposed for the 2D compressible Euler equations. The method builds on the evolution-operator formulation proposed by Roe in which conservative cell averages are updated by unsplit flux quadrature while primitive point values are evolved by acoustic and advective subsolvers. The proposed method reconstructs the acoustic increment as a cellwise Q2 field and evaluates this field at the convective foot of the target point. For constant frozen coefficients, the resulting point update reduces to the transported composition, eliminating the additive split defect and yielding the exact unsplit frozen evolution when the acoustic and advective generators commute. The resulting method preserves the exact locally linearized acoustic evolution operator of Barsukow (2025), the compact stencil, and the conservative one-stage average update. Numerical experiments probe several facets of the numerical method. A mixed Fourier wave packet isolates the split error and shows third-order point accuracy for the transported update, compared with second-order behavior for the additive update. Isentropic vortex convection confirms third-order convergence for the full nonlinear scheme, reduced error constants, and an enlarged empirical CFL range. Nonlinear Gaussian acoustic pulse evolution demonstrates preservation of radial symmetry and near-third-order decay of the symmetry error. Low-Mach shear layer tests show coherent vorticity evolution, ultra-low entropy dissipation, and absence of the coarse-grid secondary vortices seen in displayed DG/CG comparisons. Finally, a compressible under-resolved Kelvin-Helmholtz test demonstrates robust no-limiter evolution to late time with consistent entropy dissipation. Fourier diagnostics of the vertical-edge point operator support the observed improvements in acoustic phase and amplification behavior.

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 / 3 minor

Summary. The manuscript proposes a fully discrete Active Flux method for the 2D compressible Euler equations. It reconstructs the acoustic increment as a cellwise Q2 field and evaluates this field at the convective foot of the target point. For constant frozen coefficients, the resulting point update reduces to the transported composition, eliminating the additive split defect and yielding the exact unsplit frozen evolution when the acoustic and advective generators commute. The method preserves the exact locally linearized acoustic evolution operator of Barsukow (2025), the compact stencil, and the conservative one-stage average update. Numerical experiments include a mixed Fourier wave packet showing third-order point accuracy for the transported update (versus second-order for additive), third-order convergence for the isentropic vortex, symmetry preservation for the nonlinear Gaussian acoustic pulse, coherent vorticity in low-Mach shear layers, and robust evolution for the under-resolved Kelvin-Helmholtz instability.

Significance. If the central reduction holds, the work offers a targeted improvement to Active Flux schemes for hyperbolic conservation laws by removing the splitting inconsistency in the point evolution while retaining the exact linearized operator, compact stencil, and conservative update. The reported gains in convergence order, empirical CFL range, radial symmetry, low-Mach vorticity preservation, and absence of spurious secondary vortices relative to DG/CG comparisons indicate practical value. The Fourier diagnostics of the vertical-edge point operator and the direct derivation of the transported reduction for frozen coefficients are particular strengths that support broader adoption in compressible flow computations.

major comments (2)
  1. [Derivation of the transported point update] The reduction of the point update to the transported composition for constant frozen coefficients (described in the paragraph following the Q2 reconstruction definition) is load-bearing for the claim of eliminating the additive split defect. The manuscript should supply the explicit algebraic steps or a short lemma showing how the Q2 field evaluation at the convective foot produces exact cancellation of the split terms when the generators commute.
  2. [Numerical experiments, mixed Fourier test] In the mixed Fourier wave packet experiment, the reported third-order point accuracy for the transported update versus second-order for the additive update is key supporting evidence. The error norm (L2 or pointwise), wave-packet parameters, and quantitative convergence rates should be stated explicitly, ideally in a table, to confirm the isolation of the split error and allow independent verification.
minor comments (3)
  1. [Low-Mach shear layer tests] The phrase 'ultra-low entropy dissipation' in the low-Mach shear layer discussion would be strengthened by a specific quantitative measure, such as the observed entropy decay rate or a direct numerical comparison value.
  2. [Figure captions] Figure captions for the DG/CG comparisons in the shear layer test should specify the polynomial degree, mesh type, and time-stepping details of the reference schemes to facilitate fair assessment of the observed improvements.
  3. [Preservation statement] A brief parenthetical reference to the precise equation or property from Barsukow (2025) that is exactly preserved by the new reconstruction would improve clarity without lengthening the text.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the positive assessment and constructive comments. We address each major comment below and will incorporate the requested clarifications in the revised manuscript.

read point-by-point responses

  1. Referee: [Derivation of the transported point update] The reduction of the point update to the transported composition for constant frozen coefficients (described in the paragraph following the Q2 reconstruction definition) is load-bearing for the claim of eliminating the additive split defect. The manuscript should supply the explicit algebraic steps or a short lemma showing how the Q2 field evaluation at the convective foot produces exact cancellation of the split terms when the generators commute.

    Authors: We agree that an explicit derivation strengthens the presentation. In the revised manuscript we will insert a short lemma right after the Q2 reconstruction definition. The lemma will derive the algebraic cancellation: when the frozen coefficients are constant and the acoustic and advective generators commute, evaluating the cellwise Q2 acoustic-increment polynomial at the convective foot x - u Δt recovers exactly the transported composition, with the additive split terms cancelling identically. This leaves the stencil, the exact linearized acoustic operator, and the conservative average update unchanged. revision: yes

  2. Referee: [Numerical experiments, mixed Fourier test] In the mixed Fourier wave packet experiment, the reported third-order point accuracy for the transported update versus second-order for the additive update is key supporting evidence. The error norm (L2 or pointwise), wave-packet parameters, and quantitative convergence rates should be stated explicitly, ideally in a table, to confirm the isolation of the split error and allow independent verification.

    Authors: We will add a dedicated table to the mixed Fourier wave-packet subsection. The table will report the L2 error norms on the point values, the precise wave-packet parameters (wave numbers kx, ky, amplitudes, and phase), the sequence of grid resolutions, and the measured convergence rates (approximately 3.0 for the transported update and 2.0 for the additive update). This will make the isolation of the split error fully verifiable. revision: yes

Circularity Check

0 steps flagged

Minor self-citation of prior formulation; central reconstruction step is independent

full rationale

The paper introduces a new reconstruction of the acoustic increment as a cellwise Q2 field evaluated at the convective foot. For constant frozen coefficients this directly yields the transported composition by explicit construction from the reconstruction procedure, eliminating the additive split defect when generators commute. This step is self-contained and does not reduce to a fitted parameter or self-defined input. The method references Roe's evolution-operator formulation and preserves Barsukow's linearized operator, but these are contextual citations rather than load-bearing reductions; the mixed Fourier test and nonlinear experiments supply independent numerical confirmation of the accuracy improvement. No derivation step collapses to its own inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard conservation properties of the Euler equations and the prior evolution-operator framework; no new free parameters or invented entities are introduced in the abstract.

axioms (2)
  • domain assumption The compressible Euler equations are the governing system.
    Standard assumption for the target problem.
  • domain assumption Acoustic and advective generators commute under frozen coefficients.
    Invoked to obtain exact unsplit evolution for the frozen case.

pith-pipeline@v0.9.0 · 5592 in / 1223 out tokens · 42561 ms · 2026-05-14T19:03:33.792462+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

Reference graph

Works this paper leans on

29 extracted references · 29 canonical work pages

  1. [1]

    , title =

    Roe, Philip L. , title =. Proceedings of the. 2025 , doi =

    work page 2025

  2. [2]

    , title =

    Roe, Philip L. , title =. Computers & Fluids , volume =

    work page

  3. [3]

    , title =

    Roe, Philip L. , title =. Handbook of Numerical Analysis , volume =

    work page

  4. [4]

    Journal of Computational Physics , volume =

    van Leer, Bram , title =. Journal of Computational Physics , volume =

    work page

  5. [5]

    and Roe, Philip L

    Eymann, Timothy A. and Roe, Philip L. , title =. 20th

    work page

  6. [6]

    and Roe, Philip L

    Eymann, Timothy A. and Roe, Philip L. , title =. 21st

    work page

  7. [7]

    , title =

    Barsukow, Wasilij and Hohm, Jonathan and Klingenberg, Christian and Roe, Philip L. , title =. Journal of Scientific Computing , volume =

    work page

  8. [8]

    Journal of Computational Physics , volume =

    Barsukow, Wasilij , title =. Journal of Computational Physics , volume =

    work page

  9. [9]

    Journal of Scientific Computing , volume =

    Barsukow, Wasilij , title =. Journal of Scientific Computing , volume =

    work page

  10. [10]

    Barsukow, Wasilij and Edelmann, Philipp V. F. and Klingenberg, Christian and Miczek, Fabian and R. A Numerical Scheme for the Compressible Low-. Journal of Scientific Computing , volume =

    work page

  11. [11]

    Fan, Duoming , title =

    work page

  12. [12]

    Maeng, Jungyeoul , title =

    work page

  13. [13]

    ESAIM: Mathematical Modelling and Numerical Analysis (M2AN) , volume =

    Barsukow, Wasilij and Klingenberg, Christian , title =. ESAIM: Mathematical Modelling and Numerical Analysis (M2AN) , volume =

    work page

  14. [14]

    2025 , eprint =

    Barsukow, Wasilij , title =. 2025 , eprint =

    work page 2025

  15. [15]

    Extensions of

    Abgrall, R. Extensions of. ESAIM: Mathematical Modelling and Numerical Analysis , volume =

    work page

  16. [16]

    A Semi-Discrete

    Abgrall, R. A Semi-Discrete. Journal of Scientific Computing , volume =

    work page

  17. [17]

    Chudzik, Erik and Helzel, Christiane and Luk. Active. Journal of Scientific Computing , volume =

    work page

  18. [18]

    On the Behaviour of Upwind Schemes in the Low

    Guillard, Herv. On the Behaviour of Upwind Schemes in the Low. Computers & Fluids , volume =

    work page

  19. [19]

    Analysis of

    Dellacherie, St. Analysis of. Journal of Computational Physics , volume =

    work page

  20. [20]

    International Journal for Numerical Methods in Fluids , volume=

    High-order CFD methods: current status and perspective , author=. International Journal for Numerical Methods in Fluids , volume=. 2013 , publisher=

    work page 2013

  21. [21]

    Computer Physics Communications , volume=

    : A high-order discontinuous Galerkin solver for flow simulations and multi-physics applications , author=. Computer Physics Communications , volume=. 2023 , publisher=

    work page 2023

  22. [22]

    Annual Review of Fluid Mechanics , volume=

    A perspective on the state of aerospace computational fluid dynamics technology , author=. Annual Review of Fluid Mechanics , volume=. 2023 , publisher=

    work page 2023

  23. [23]

    Journal of Computational Physics , volume=

    Generalized upwind summation-by-parts operators and their application to nodal discontinuous Galerkin methods , author=. Journal of Computational Physics , volume=. 2025 , publisher=

    work page 2025

  24. [24]

    Journal of Computational Physics , volume=

    On the robustness of high-order upwind summation-by-parts methods for nonlinear conservation laws , author=. Journal of Computational Physics , volume=. 2025 , publisher=

    work page 2025

  25. [25]

    Journal of Scientific Computing , volume=

    Stable volume dissipation for high-order finite-difference and spectral-element methods with the summation-by-parts property , author=. Journal of Scientific Computing , volume=. 2026 , publisher=

    work page 2026

  26. [26]

    2018 AIAA aerospace sciences meeting , pages=

    Comparing active flux and discontinuous Galerkin methods for compressble flow , author=. 2018 AIAA aerospace sciences meeting , pages=

    work page 2018

  27. [27]

    arXiv preprint arXiv:2603.18160 , year=

    On the equivalence of semi-discrete Active Flux and Discontinuous Galerkin methods and a comparison of their performance , author=. arXiv preprint arXiv:2603.18160 , year=

    work page arXiv

  28. [28]

    arXiv preprint arXiv:2508.15017 , year=

    Semi-discrete Active Flux as a Petrov-Galerkin method , author=. arXiv preprint arXiv:2508.15017 , year=

    work page arXiv

  29. [29]

    arXiv preprint arXiv:2508.06273 , year=

    A Fully Discrete Truly Multidimensional Active Flux Method For The Two-Dimensional Euler Equations , author=. arXiv preprint arXiv:2508.06273 , year=

    work page arXiv