Pith. sign in

REVIEW 3 major objections 2 minor 37 references

The explicit generalized forcing method generates data that yields accurate stable reduced models for linear transport PDE closures when the model form matches the underlying relation.

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 ↗

Generalized Forcing Method generates diverse data for training linear transport PDE closure models via compatible body forces in zero-initial-condition simulations, with explicit and implicit variants applied to shear and inhomogeneous flows.

T0 review reviewed 2026-06-28 challenge →

load-bearing objection GFM gives a forcing-based way to make tailored training data for linear transport closures, but the abstract leaves the accuracy claims unquantified. the 3 major comments →

arxiv 2606.05141 v1 pith:IKRGAB3N submitted 2026-06-03 physics.flu-dyn

Generalized Forcing Method: Generation of Diverse Data for Training Linear Transport PDE Closure Models

Wenyuan Xue , Ali Mani This is my paper
classification physics.flu-dyn
keywords generalized forcing methoddata generationclosure modelinglinear transport PDEreduced modelsexplicit GFMfluid dynamicstraining data
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.

The reading

The paper introduces the Generalized Forcing Method as a framework for producing accurate, affordable, and diverse training data specifically suited to linear transport closure models. It achieves this by running simulations from a zero initial condition while adding an extra body force constructed to remain compatible with the target reduced dynamics. The method splits into an implicit variant that prescribes resolved trajectories and an explicit variant that builds a basis of admissible forcings. Tests on homogeneous shear flows, spatially inhomogeneous flows, and random-coefficient shear flows show that the explicit version recovers accurate and stable reduced models precisely when the chosen reduced variables and model form align with the true closure relation.

Core claim

The Generalized Forcing Method generates training data by running simulations with zero initial condition and an extra body force constructed compatibly with the reduced dynamics. This produces the explicit GFM variant that constructs a basis of admissible forcings, and application to three linear transport closure problems demonstrates that it identifies accurate and stable reduced models when the reduced variables and model form are consistent with the underlying closure relation.

What carries the argument

Generalized Forcing Method (GFM), a data-generation procedure that adds an extra body force compatible with reduced dynamics to zero-initial-condition simulations in order to produce tailored training data.

Load-bearing premise

The extra body force can be constructed compatibly with the reduced dynamics without introducing inconsistencies that invalidate the generated data for training.

What would settle it

Apply eGFM to a homogeneous shear flow where the reduced variables and model form are chosen inconsistently with the true closure and check whether the identified model is inaccurate or unstable on independent test trajectories.

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

If this is right

  • eGFM produces training data that recovers accurate reduced models for homogeneous shear flows when consistency holds.
  • The same data-generation procedure extends to spatially inhomogeneous flows and homogeneous shear flows with random coefficients.
  • The resulting reduced models remain stable when the model form matches the closure relation.
  • Implicit GFM prescribes resolved trajectories while explicit GFM builds a basis of admissible forcings to ensure diversity.

Where Pith is reading between the lines

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

  • The compatibility requirement on the body force could be generalized to nonlinear closures if a suitable projection or constraint is defined.
  • This structured data-generation approach may reduce reliance on expensive ensembles of full-order simulations for training scientific machine-learning models.
  • The method points toward a broader principle that training data for reduced models should be generated under constraints that mirror the target model structure.
Share X Bluesky LinkedIn Reddit HN

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

3 major / 2 minor

Summary. The paper introduces the Generalized Forcing Method (GFM) as a data-generation framework for training linear transport PDE closure models. It defines implicit GFM (iGFM), which prescribes resolved trajectories, and explicit GFM (eGFM), which constructs a basis of admissible forcings added as extra body forces to zero-initial-condition simulations. The framework is applied to three linear transport closure problems (homogeneous shear flows, spatially inhomogeneous flows, and homogeneous shear flows with random coefficients), with the central claim that eGFM identifies accurate and stable reduced models when the reduced variables and model form are consistent with the underlying closure relation.

Significance. If the compatibility of the extra body force with reduced dynamics can be shown to leave the target closure relation unchanged, the method would offer a systematic route to generating diverse, tailored training data for data-driven closure modeling in transport PDEs, addressing a key bottleneck in reduced-order modeling for fluid dynamics applications.

major comments (3)
  1. [Abstract] Abstract: the claim of 'accurate and stable reduced models' is unsupported by any quantitative metrics, error norms, baseline comparisons, or validation details; this absence makes it impossible to evaluate whether the generated data actually support the stability and accuracy assertions.
  2. [Method] eGFM construction (likely §3): the admissibility condition for the basis of forcings must be shown explicitly not to correlate with or embed the closure coefficients being learned; otherwise the generated trajectories risk containing artifacts that make the learned model appear accurate only due to the forcing construction rather than capturing the original PDE closure.
  3. [Results] Application sections (likely §4): without reported quantitative error metrics, cross-validation against full-order simulations, or ablation on the forcing basis size, the assertion that eGFM succeeds on the three test cases cannot be assessed for load-bearing consistency with the reduced dynamics.
minor comments (2)
  1. [Introduction] Clarify notation for the extra body force term and its projection onto the reduced variables early in the manuscript to avoid ambiguity when reading the compatibility condition.
  2. [Results] Add a short table summarizing the three test cases, their reduced variables, and the chosen model forms for easier comparison across applications.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive comments. We address each major point below and will revise the manuscript accordingly to provide stronger quantitative support.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim of 'accurate and stable reduced models' is unsupported by any quantitative metrics, error norms, baseline comparisons, or validation details; this absence makes it impossible to evaluate whether the generated data actually support the stability and accuracy assertions.

    Authors: We agree the abstract is high-level and lacks supporting numbers. In revision we will add concise quantitative indicators (e.g., L2 error norms and stability indicators drawn from the results) to make the claim evaluable. revision: yes

  2. Referee: [Method] eGFM construction (likely §3): the admissibility condition for the basis of forcings must be shown explicitly not to correlate with or embed the closure coefficients being learned; otherwise the generated trajectories risk containing artifacts that make the learned model appear accurate only due to the forcing construction rather than capturing the original PDE closure.

    Authors: The admissibility condition is formulated from the linear structure of the transport operator and the chosen reduced variables, independent of the specific closure coefficients. We will insert an explicit algebraic demonstration in the revised §3 proving that the forcing basis does not embed or correlate with the coefficients to be learned. revision: yes

  3. Referee: [Results] Application sections (likely §4): without reported quantitative error metrics, cross-validation against full-order simulations, or ablation on the forcing basis size, the assertion that eGFM succeeds on the three test cases cannot be assessed for load-bearing consistency with the reduced dynamics.

    Authors: We accept that additional quantitative evidence is needed. The revised §4 will report L2 error norms, cross-validation against full-order runs, and an ablation on basis size for each of the three cases to confirm consistency with the reduced dynamics. revision: yes

Circularity Check

0 steps flagged

No circularity: new data-generation construction presented without reduction to inputs or self-citations

full rationale

The paper introduces GFM/eGFM as a framework for generating training data via extra body forces constructed compatibly with reduced dynamics. The abstract and reader's summary present this as an original construction for producing diverse data tailored to linear transport closures. No equations or steps are shown that define a quantity in terms of itself, rename a fitted parameter as a prediction, or rely on load-bearing self-citations. The central claim (accurate reduced models when variables match the closure) is framed as an empirical outcome of the method rather than a definitional tautology. This is the common case of a self-contained methodological contribution.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides insufficient detail to identify specific free parameters, axioms, or invented entities; no explicit fitting procedures or new postulated quantities are described.

reviewed 2026-06-28 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Generalized Forcing Method: Generation of Diverse Data for Training Linear Transport PDE Closure Models." pith.science (2026). https://pith.science/paper/IKRGAB3N

@misc{pith2026260605141,
  author       = {Pith},
  title        = {Pith review of: Generalized Forcing Method: Generation of Diverse Data for Training Linear Transport PDE Closure Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IKRGAB3N}},
  note         = {Machine review of arXiv:2606.05141}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Data-driven closure modeling for transport partial differential equations requires training data that are accurate, affordable, diverse, and directly tailored to the target closure fields. We develop the Generalized Forcing Method (GFM), a data-generation framework for training linear transport closure models. GFM generates such data by running simulations with a zero initial condition and an extra body force that is constructed compatibly with the reduced dynamics. This framework leads to implicit GFM (iGFM), which prescribes resolved trajectories, and explicit GFM (eGFM), which constructs a basis of admissible forcings. We apply eGFM to three linear transport closure problems: homogeneous shear flows, spatially inhomogeneous flows, and homogeneous shear flows with random coefficients. The results show that eGFM can identify accurate and stable reduced models when the reduced variables and model form are consistent with the underlying closure relation.

Figures

Figures reproduced from arXiv: 2606.05141 by Ali Mani, Wenyuan Xue.

Figure 1
Figure 1. Figure 1: FIG. 1. Dispersion of passive scalar in parallel flow with ¯c [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Contours of [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Convergence of the analytical model family for the problem in Eq. (1) with increasing [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Comparison of the RMSE of ¯c [PITH_FULL_IMAGE:figures/full_fig_p023_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Time histories of ¯c [PITH_FULL_IMAGE:figures/full_fig_p023_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. Time histories of ¯c [PITH_FULL_IMAGE:figures/full_fig_p024_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7. Time histories of ¯c [PITH_FULL_IMAGE:figures/full_fig_p024_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIG. 8. RMSE of eGFM spectral-based reduced models over the space of forcing subsets for [PITH_FULL_IMAGE:figures/full_fig_p026_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: FIG. 9. RMSE of eGFM spectral-based reduced models over the space of forcing subsets for [PITH_FULL_IMAGE:figures/full_fig_p027_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: FIG. 10. Dispersion of passive scalar in the complex asymmetric parallel flow in Eq. (78) for the [PITH_FULL_IMAGE:figures/full_fig_p027_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: FIG. 11. Contours of [PITH_FULL_IMAGE:figures/full_fig_p028_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: FIG. 12. Comparison of the RMSE of ¯c [PITH_FULL_IMAGE:figures/full_fig_p029_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: FIG. 13. Comparison of the RMSE of ¯c [PITH_FULL_IMAGE:figures/full_fig_p029_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: FIG. 14. RMSE of eGFM spectral-based reduced models over the space of forcing subsets for [PITH_FULL_IMAGE:figures/full_fig_p030_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: FIG. 15. RMSE of eGFM spectral-based reduced models over the space of forcing subsets for [PITH_FULL_IMAGE:figures/full_fig_p031_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: FIG. 16. RMSE of eGFM velocity-based reduced models over the space of forcing subsets for [PITH_FULL_IMAGE:figures/full_fig_p032_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: FIG. 17. Passive scalar dispersion in inhomogeneous flow I, Eq. (80), with initial condition ¯c [PITH_FULL_IMAGE:figures/full_fig_p033_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: FIG. 18. Passive scalar dispersion in inhomogeneous flow II, Eq. (81), with initial condition ¯c [PITH_FULL_IMAGE:figures/full_fig_p033_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: FIG. 19. Comparison of the RMSE of ¯c [PITH_FULL_IMAGE:figures/full_fig_p036_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: FIG. 20. Comparison of the RMSE of ¯c [PITH_FULL_IMAGE:figures/full_fig_p036_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: FIG. 21. RMSE of eGFM spectral-based reduced models over the space of forcing subsets for [PITH_FULL_IMAGE:figures/full_fig_p037_21.png] view at source ↗
Figure 22
Figure 22. Figure 22: FIG. 22. RMSE of eGFM velocity-based reduced models over the space of forcing subsets for [PITH_FULL_IMAGE:figures/full_fig_p037_22.png] view at source ↗
Figure 23
Figure 23. Figure 23: FIG. 23. Comparison of the RMSE of ¯c [PITH_FULL_IMAGE:figures/full_fig_p038_23.png] view at source ↗
Figure 24
Figure 24. Figure 24: FIG. 24. Comparison of the RMSE of ¯c [PITH_FULL_IMAGE:figures/full_fig_p038_24.png] view at source ↗
Figure 25
Figure 25. Figure 25: FIG. 25. RMSE of eGFM spectral-based reduced models over the space of forcing subsets for [PITH_FULL_IMAGE:figures/full_fig_p039_25.png] view at source ↗
Figure 26
Figure 26. Figure 26: FIG. 26. RMSE of eGFM velocity-based reduced models over the space of forcing subsets for [PITH_FULL_IMAGE:figures/full_fig_p039_26.png] view at source ↗
Figure 27
Figure 27. Figure 27: FIG. 27. Passive scalar dispersion for the problem in Eq. (87) with initial condition ¯c [PITH_FULL_IMAGE:figures/full_fig_p040_27.png] view at source ↗
Figure 28
Figure 28. Figure 28: For n ≤ 3, the eGFM-based and direct-fit models attain comparable RMSE. For n ≥ 4, the direct-fit approach fails, while the eGFM-based models remain stable but show no further improvement; the error reaches a plateau beyond n = 3. The dependence on the training dataset is shown in Figs. 29 and 30. In contrast to the previous examples, the present case exhibits qualitatively different trends: increasing nx… view at source ↗
Figure 29
Figure 29. Figure 29: FIG. 29. RMSE of eGFM spectral-based reduced models over the space of forcing subsets for [PITH_FULL_IMAGE:figures/full_fig_p043_29.png] view at source ↗
Figure 30
Figure 30. Figure 30: FIG. 30. RMSE of eGFM spectral-based reduced models over the space of forcing subsets for [PITH_FULL_IMAGE:figures/full_fig_p044_30.png] view at source ↗
Figure 31
Figure 31. Figure 31: FIG. 31. Comparison of the RMSE of ¯c [PITH_FULL_IMAGE:figures/full_fig_p045_31.png] view at source ↗
Figure 32
Figure 32. Figure 32: FIG. 32. RMSE of eGFM velocity-based reduced models over the space of forcing subsets for [PITH_FULL_IMAGE:figures/full_fig_p045_32.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

37 extracted references · 2 canonical work pages

  1. [1]

    (59) for training and test data generation, and for the reduced model in Eq

    Simulation settings We briefly summarize the numerical solvers used for the full model in Eq. (59) for training and test data generation, and for the reduced model in Eq. (60) for model evaluation. We truncate the domain inxto a finite periodic domain [0, L x]. a. Full model.We discretize Eq. (59) using a Fourier pseudo-spectral method in (x, y) and advan...

  2. [2]

    (60), we generate training data using eGFM forcings

    GFM forcing construction To identify the reduced model in Eq. (60), we generate training data using eGFM forcings. As derived in Sec. II B, the admissible forcings are determined by the weight functions associated with the chosen reduced variables. Thus, we discuss the forcing construction for the two types of resolved variables separately. 20 TABLE I. Ke...

  3. [3]

    For this flow, the velocity-based variables are linear combinations of the spectral-based variables; thus, the two reduced model families are equiv- alent

    Example 1: Simple parallel flow We return to the caseu= cos(y). For this flow, the velocity-based variables are linear combinations of the spectral-based variables; thus, the two reduced model families are equiv- alent. Accordingly, we report only the results from spectral-based models. The evolution of the mean field ¯cfor this case has been shown in Fig...

  4. [4]

    Example 2: Complex asymmetric parallel flow We next consider the more complex asymmetric shear flow profile in Eq. (78). The velocity profile and the time history of ¯cfor the test case are shown in Fig. 10. Further details of the full model solution are shown in Fig. 11. In contrast to the simple cosine shear flow case, ¯cis no longer symmetric inxdue to...

  5. [5]

    (59) following Ref

    Problem statement We consider a two-dimensional unsteady inhomogeneous advection-diffusion problem, adapted from Eq. (59) following Ref. [11]: ∂c ∂t +u 1(x, y) ∂c ∂x +u 2(x, y) ∂c ∂y = 0.05 ∂2c ∂x2 + ∂2c ∂y 2 ,Ω = [−π, π]×[−π, π].(79) The system is periodic in bothxandy, with initial conditionc(x, y,0) = ¯c 0(x). The averaging operator ¯•denotes the cross...

  6. [6]

    Reduced variables and GFM forcing construction.Following the discussion in Sec

    Numerical discovery of the reduced model a. Reduced variables and GFM forcing construction.Following the discussion in Sec. III B, we consider both spectral-based and velocity-based reduced variables. For the spectral-based variables, we adopt the cosine-mode definition in Eq. (69), motivated by the symmetry of the velocity field iny. For the velocity-bas...

  7. [7]

    (80) and (81), the eGFM training data are generated on a coarse grid withN x = 32,N y = 32, ∆t= 8×10 −3,T max = 8, and zero initial condition ¯c0 ≡0

    Results and discussion For both inhomogeneous flow cases, Eqs. (80) and (81), the eGFM training data are generated on a coarse grid withN x = 32,N y = 32, ∆t= 8×10 −3,T max = 8, and zero initial condition ¯c0 ≡0. Compared to the example in Sec. III, more temporal snapshots are taken to improve the stability of the pointwise fit. For modeling, we consider ...

  8. [8]

    Problem statement We consider a modified version of Eq. (59) where the velocity carries a random temporal fluctuation: ∂c ∂t +u(y, t) ∂c ∂x = ∂2c ∂y 2 ,Ω = (−∞,∞)×[−π, π],(87) with periodic boundary conditions inyand initial conditionc(x, y,0) = ¯c 0(x). The velocity is given by u(y, t) = cos(y) 1 + 0.5β(t) ,(88) whereβ(t) is a zero-mean Gaussian process ...

  9. [9]

    Reduced variables and GFM forcing construction.As in Sec

    Numerical discovery of the reduced model a. Reduced variables and GFM forcing construction.As in Sec. III B, we consider both spectral-based and velocity-based reduced variables. Sinceu(y, t) is symmetric iny, the spectral-based variables follow the cosine-mode definition in Eq. (69). The velocity-based variables follow Eq. (72). As discussed in Sec. III ...

  10. [11]

    Smagorinsky, General circulation experiments with the primitive equations: I

    J. Smagorinsky, General circulation experiments with the primitive equations: I. the basic experiment, Monthly weather review91, 99 (1963)

    1963

  11. [12]

    S. B. Pope, Turbulent flows, Measurement Science and Technology12, 2020 (2001)

    2020

  12. [13]

    Sanderse, P

    B. Sanderse, P. Stinis, R. Maulik, and S. E. Ahmed, Scientific machine learning for closure models in multiscale problems: A review, arXiv preprint arXiv:2403.02913 (2024)

    work page arXiv 2024

  13. [14]

    Hamba, An analysis of nonlocal scalar transport in the convective boundary layer using the green’s function, Journal of Atmospheric Sciences52, 1084 (1995)

    F. Hamba, An analysis of nonlocal scalar transport in the convective boundary layer using the green’s function, Journal of Atmospheric Sciences52, 1084 (1995)

    1995

  14. [15]

    Hamba, Nonlocal expression for scalar flux in turbulent shear flow, Physics of Fluids16, 1493 (2004)

    F. Hamba, Nonlocal expression for scalar flux in turbulent shear flow, Physics of Fluids16, 1493 (2004)

    2004

  15. [16]

    Hamba, Nonlocal analysis of the reynolds stress in turbulent shear flow, Physics of Fluids 17(2005)

    F. Hamba, Nonlocal analysis of the reynolds stress in turbulent shear flow, Physics of Fluids 17(2005)

    2005

  16. [17]

    Hamba, Analysis and modelling of non-local eddy diffusivity for turbulent scalar flux, Journal of Fluid Mechanics950, A38 (2022)

    F. Hamba, Analysis and modelling of non-local eddy diffusivity for turbulent scalar flux, Journal of Fluid Mechanics950, A38 (2022)

    2022

  17. [18]

    Hamba, Analysis and modelling of non-local eddy diffusivity in turbulent channel flow, Journal of Fluid Mechanics1012, A21 (2025)

    F. Hamba, Analysis and modelling of non-local eddy diffusivity in turbulent channel flow, Journal of Fluid Mechanics1012, A21 (2025)

    2025

  18. [19]

    Hamba, Nonlocal eddy viscosity for reynolds stress and passive vector flux in turbulence, Physical Review Fluids10, 124601 (2025)

    F. Hamba, Nonlocal eddy viscosity for reynolds stress and passive vector flux in turbulence, Physical Review Fluids10, 124601 (2025)

    2025

  19. [20]

    Mani and D

    A. Mani and D. Park, Macroscopic forcing method: a tool for turbulence modeling and analysis of closures, Physical Review Fluids6, 054607 (2021)

    2021

  20. [21]

    J. Liu, H. H. Williams, and A. Mani, Systematic approach for modeling a nonlocal eddy diffusivity, Physical Review Fluids8, 124501 (2023)

    2023

  21. [22]

    Park and A

    D. Park and A. Mani, Direct calculation of the eddy viscosity operator in turbulent channel flow at reτ= 180, Journal of Fluid Mechanics998, A33 (2024). 48

    2024

  22. [23]

    Lavacot, J

    D.-L. Lavacot, J. Liu, H. Williams, B. E. Morgan, and A. Mani, Non-locality of mean scalar transport in two-dimensional rayleigh–taylor instability using the macroscopic forcing method, Journal of Fluid Mechanics985, A47 (2024)

    2024

  23. [24]

    van der Oord, M

    C. van der Oord, M. Sachs, D. P. Kov´ acs, C. Ortner, and G. Cs´ anyi, Hyperactive learning for data-driven interatomic potentials, npj Computational Materials9, 168 (2023)

    2023

  24. [25]

    G. I. Taylor, Dispersion of soluble matter in solvent flowing slowly through a tube, Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences219, 186 (1953)

    1953

  25. [26]

    Aris, On the dispersion of a solute in a fluid flowing through a tube, Proceedings of the Royal Society of London

    R. Aris, On the dispersion of a solute in a fluid flowing through a tube, Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences235, 67 (1956)

    1956

  26. [27]

    G. I. Taylor, Conditions under which dispersion of a solute in a stream of solvent can be used to measure molecular diffusion, Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences225, 473 (1954)

    1954

  27. [28]

    Gill and R

    W. Gill and R. Sankarasubramanian, Exact analysis of unsteady convective diffusion, Pro- ceedings of the Royal Society of London. A. Mathematical and Physical Sciences316, 341 (1970)

    1970

  28. [29]

    Barton, On the method of moments for solute dispersion, Journal of Fluid Mechanics126, 205 (1983)

    N. Barton, On the method of moments for solute dispersion, Journal of Fluid Mechanics126, 205 (1983)

    1983

  29. [30]

    Frankel and H

    I. Frankel and H. Brenner, On the foundations of generalized taylor dispersion theory, Journal of Fluid Mechanics204, 97 (1989)

    1989

  30. [31]

    Moyal, Stochastic processes and statistical physics, Journal of the Royal Statistical Society

    J. Moyal, Stochastic processes and statistical physics, Journal of the Royal Statistical Society. Series B (Methodological)11, 150 (1949)

    1949

  31. [32]

    Virtanen, R

    P. Virtanen, R. Gommers, T. E. Oliphant, M. Haberland, T. Reddy, D. Cournapeau, E. Burovski, P. Peterson, W. Weckesser, J. Bright,et al., Scipy 1.0: fundamental algorithms for scientific computing in python, Nature methods17, 261 (2020)

    2020

  32. [33]

    Frigo and S

    M. Frigo and S. G. Johnson, The design and implementation of fftw3, Proceedings of the IEEE 93, 216 (2005)

    2005

  33. [34]

    Canuto, M

    C. Canuto, M. Y. Hussaini, A. Quarteroni, and T. A. Zang,Spectral methods: fundamentals in single domains(Springer, 2006)

    2006

  34. [35]

    C. L. Lawson and R. J. Hanson,Solving least squares problems(SIAM, 1995)

    1995

  35. [36]

    Kubo, Stochastic liouville equations, Journal of Mathematical Physics4, 174 (1963)

    R. Kubo, Stochastic liouville equations, Journal of Mathematical Physics4, 174 (1963). 49

    1963

  36. [37]

    Mori, Transport, collective motion, and brownian motion, Progress of theoretical physics 33, 423 (1965)

    H. Mori, Transport, collective motion, and brownian motion, Progress of theoretical physics 33, 423 (1965)

    1965

  37. [38]

    Generalized Forcing Method: Generation of Diverse Data for Training Linear Transport PDE Closure Models

    W. Xue and A. Mani, Reproduction data and scripts for “Generalized Forcing Method: Generation of Diverse Data for Training Linear Transport PDE Closure Models” (2026), https://doi.org/10.5281/zenodo.20424795. 50

    work page doi:10.5281/zenodo.20424795 2026

This paper was first reviewed by grok-4.3 on June 28, 2026.