Pith. sign in

REVIEW 4 major objections 5 minor 34 references

Kernel-learning parameter prediction and evaluation in algebraic multigrid method for several PDEs

T0 review · 4 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Gaussian process models trained on small grids predict near-optimal AMG parameters for larger grids.

desk verdict A plausible application of the authors' GPR pipeline to AMG strong-threshold tuning, but the evidence is too thin and the GP extrapolation to n=1024 likely rides on AMG's flat performance plateau rather than accurate theta prediction. read the letter →

arxiv 2504.14930 v3 pith:SCSPOTCQ submitted 2025-04-21 math.NA cs.NA

Junyue Luo , Xiaoqiang Yue , Fangfang Zhang , Juan Zhang This is my paper · ORCID
classification math.NAcs.NA MSC 52B1065D1868U0568U07
keywords algebraicmultigridGaussianprocessregressionstrongthresholdparameterkernellearningpredictionpartialdifferentialequationsHelmholtzequationPoisson
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

This paper tries to establish that Gaussian process regression (GPR), trained only on small algebraic multigrid (AMG) solves, can replace the expensive grid search over AMG's strong threshold parameter. The target is the threshold $\theta$ that controls which variables are treated as strongly connected and therefore how the coarse grid is built. Across Poisson, diffusion, and Helmholtz problems, the paper reports that GPR-predicted $\theta$ at $n=1024$ reproduces grid-search iteration counts while using far less computation, and that a default $\theta$ often performs worse or fails to converge. If true, this gives AMG users a cheap, data-driven way to set a parameter that is normally chosen by trial and error.

What carries the argument

The central object is a Gaussian process $f(n)$ with a kernel formed as a linear combination of a library of basic kernels $k(x,x') = \sum_{\xi=1}^N c_\xi k_\xi(x,x')$, with coefficients fit by maximizing the log marginal likelihood. The GP posterior mean $\mu_* = K(n_*,n)(K(n,n)+\sigma^2 I)^{-1}(\theta - \mu)$ converts observed optimal thresholds at small sizes into a predicted threshold at a new size. The target parameter is the AMG strong threshold $\theta$, defined by the condition $|a_{ij}| \geq \theta \max_{k \neq i} |a_{ik}|$, which controls C/F splitting and hence the whole multigrid hierarchy. The machinery's job is to make $\theta^*(n)$ extrapolable without running AMG at the target size.

What would settle it

For a held-out family of diffusion or Helmholtz matrices at $n=1024$, sweep $\theta$ finely near the GPR prediction, record the true optimal iteration count, and compare: if the GPR-chosen $\theta$ is beaten by more than a few percent on many instances, or if its posterior interval frequently misses the swept optimum, then the extrapolation claim is falsified.

Watch

Extended reading notes

Core claim

The central claim is that the mapping from problem size $n$ to grid-search-optimal threshold $\theta^*(n)$ can be learned by a Gaussian process from training data on $n \in [64,400]$ (plus targeted retraining points in $[200,600]$) and extrapolated to $n=1024$. At $n=1024$, the predicted $\theta$ yields the same iteration count as the grid-search optimum for the constant-coefficient Poisson equation (17 iterations at $\theta=0.368$ versus 17 at $\theta=0.276$), close counts for the Helmholtz equation (27 at $\theta=0.344$ versus 25 at $\theta=0.323$, while the default $\theta=0.25$ fails to converge), and a large improvement over the default for diffusion problems (489 average iterations versus 936, at 79 seconds versus 179). The paper interprets these matches as evidence that GPR-based parameter prediction is a practical substitute for parameter sweeping and that combined kernel functions improve reliability.

Load-bearing premise

The grid-search-optimal threshold, viewed as a function of problem size, changes smoothly enough that a model trained on sizes up to 400 can extrapolate it to size 1024.

Editorial extensions

If this is right

  • AMG setup for a new, larger problem can start from a GP prediction instead of a sweep over the threshold parameter, saving the dominant tuning cost.
  • Combined kernel functions, especially Gaussian+Laplacian and Rational Quadratic+Laplacian, can reach prediction-interval coverage at or above the single Gaussian kernel, making the reliability claim stronger than the accuracy claim alone.
  • The multi-metric evaluation template (R2, correlation, PICP, MSE/RMSE/MAE, MdAPE, BIC, LOO-SPE) gives a reusable way to compare parameter predictors for iterative solvers.
  • For repeated solves at many sizes, a trained GP can be reused at each new matrix size without recomputing the sweep, since the model maps size to parameter directly.

Reading between the lines

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

  • A sharper test of the method would report the width of the iteration plateau around the predicted $\theta$; if AMG's iteration count barely changes over a wide $\theta$ range, the reported $n=1024$ matches may mostly reflect robustness rather than prediction skill.
  • The same small-to-large GP transfer could be tried on other AMG controls (smoother sweeps, coarsening aggressiveness), where the smoothness of the optimal parameter as a function of size may differ.
  • Because the paper retrains separately for each PDE family, the method as presented does not yet show cross-equation transfer; testing one GP on an unseen equation type would be a natural extension.
  • Predicting on $n=1024$ from training on $n \leq 600$ relies on extrapolation; a Bayesian treatment that also reports the posterior variance at the target size could flag cases where the prediction should not be trusted.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. This manuscript proposes to select the strong threshold parameter theta in algebraic multigrid (AMG) by Gaussian process regression (GPR) trained on small-scale systems. The training data are grid-search-optimal theta values for n=64,...,400 (with additional retraining points up to n=600), and the method is evaluated by predicting theta at n=1024 for the constant-coefficient Poisson equation, a variable-coefficient diffusion equation, and the Helmholtz equation. A kernel library with linear combinations is compared using MSE, RMSE, MAE, R2, BIC, correlation, MdAPE, LOO-SPE, and PICP. The paper reports that GPR-predicted theta gives iteration counts close to grid-search values at n=1024 while reducing the cost of parameter sweeping.

Significance. The use of GPR to avoid parameter sweeps in AMG is a practical and timely idea, and the paper makes an effort to compare several kernel combinations with multiple evaluation metrics. If the extrapolation from small n to large n were demonstrated robustly, the method would be useful. However, the current evidence is thin: only one extrapolation point per equation, no repeated trials, no error bars, and the reported iteration counts are consistent with the flat performance plateau of AMG with respect to theta rather than with accurate theta prediction. The contribution is also incremental with respect to the authors' prior kernel-learning frameworks [10,11]. The conclusions are therefore plausible but not yet substantiated.

major comments (4)
  1. [§3.1, Tables 1 and 3] The central claim is tested at a single extrapolation point n=1024, with no intermediate extrapolation points or repeated trials. Table 1 shows that the optimal theta*(n) is highly non-smooth (0.269 at n=80, 0.387 at n=96, 0.324 at n=112), so the GP's stationarity/smoothness assumption used in §2.2 is not satisfied; the fact that theta_pred=0.368 and theta_opt=0.276 give the same iteration count is likely a consequence of the flat iteration plateau visible in Figure 1, not of accurate parameter prediction.
  2. [§3.2, Table 6] For the diffusion equation, the GPR prediction yields 489.34 iterations versus 432.67 for the grid-search optimum and 78.89 s versus 51.17 s; this is roughly 13% more iterations and 54% more time, so the statement that GPR results 'closely approximate the optimal parameters' efficiency' is not supported by the numbers in the table. The table also omits the predicted and optimal theta values, the number of test matrices, and any standard deviation, so the comparison cannot be assessed.
  3. [§3.3, Table 8] The table lists a test set with n in [200,600] and 10 randomly selected points, but Table 9 reports results at n=1024; if the 10 points were used only for model selection or retraining, the distinction between test and retraining sets must be clarified, and if they were used as a test set, the extrapolation claim at n=1024 is not the one being tested. Either way, the internal inconsistency prevents the reader from verifying the experimental design.
  4. [§3.1, Tables 4–5 and §3.3, Tables 10–11] The evaluation metrics are computed on an implied test set of only 7 points (PICP fractions such as 4/7, 3/7), and the reported R2 values are low or negative (e.g., R2=0.3965 for Gaussian+Laplacian in Table 4, and R2=0.0957 for Rational Quadratic+Laplacian in Table 10). These values do not support the abstract's claim that a 'comprehensive analysis ... confirms GPR's efficiency and reliability'; the number of test points and the exact n values used for the metrics need to be stated.
minor comments (5)
  1. [§3.2] Section 3.2 states that the training set consists of 23 matrices, but n=64,80,...,400 with increment 16 gives 22 values; please reconcile the count.
  2. [§3.2 and §3.3] The text refers to 'Table reft2dks' and 'Table 3.3' instead of stable table numbers; the cross-references need to be corrected.
  3. [§3.2] Figure 2 is referenced in Section 3.2 for the block partition, but the caption of Figure 2 in the text actually describes regression curves in Section 3.1; the figure numbering and cross-references need correction.
  4. [§3] The manuscript does not state the GPR hyperparameters, the full kernel-library elements, the random seeds for the diffusion experiments, or the number of Monte Carlo runs; without these, the numerical results are not reproducible.
  5. [Abstract and §5] The abstract and Section 5 claim 'theoretical support for AMG's practical application,' but the paper provides no theoretical analysis; please rephrase to match the actual contribution.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular reduction in the derivation chain; the GPR result at n=1024 is an out-of-sample extrapolation, with only minor non-load-bearing self-citations to the authors' earlier GPR and kernel-learning framework.

full rationale

The load-bearing chain is an empirical extrapolation rather than a tautology. The training targets are grid-search-optimal AMG strong-threshold values from Tables 1 and 7 for n=64,...,400; the GPR posterior in Section 2.2 is built from these values and from retraining points in [200,600] as described in Tables 2 and 8; the final reported test is n=1024 in Tables 3 and 9, which lies outside every training and retraining range. Thus the claimed iteration-count match is a genuine out-of-sample check, not a fitted quantity renamed as a prediction. No equation defines the grid-search optimum in terms of the GPR output, and no n=1024 target value is fed back into the posterior before the comparison is made. The paper does cite the authors' earlier GPR parameter-prediction and kernel-learning framework, [10] and [11], in Section 1 ('we adopt the approach outlined in [10]' and 'we incorporate the kernel learning technique described in [4, 11]'), and those citations share authors with the present paper; however, they are methodological provenance for a standard Gaussian-process technique rather than load-bearing evidence for the numerical claims. Concerns about non-smooth theta*(n) shown in Table 1 and the flat performance plateau visible in Figure 1 are correctness and robustness risks, not circular reductions, because the paper does not force the predicted theta to equal the grid-search optimum; in fact the predicted theta differs from the optimal theta by 0.092 for the Poisson case at n=1024 while still matching the iteration count. Overall, no step reduces by construction to its own input, so any circularity is limited to minor, non-load-bearing self-citation.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The central claim rests on the smoothness and transferability of the optimal AMG parameter across problem sizes, plus the standard AMG framework. Three free parameters (kernel hyperparameters, combination coefficients, and noise variance) are fitted or hand-chosen and are not independently verified. No new physical or mathematical entities are introduced.

free parameters (3)
  • GPR kernel hyperparameters = not reported
    Length scales and variances for each base kernel are obtained by maximizing the log-likelihood (Section 2.2); these are data-fitted and control the prediction.
  • Kernel combination coefficients c_xi = not reported
    The final kernel is a linear combination of base kernels; coefficients are determined by 'TDLS training' (Introduction), a procedure not defined in the paper, so the fitted values are not reproducible.
  • Noise variance sigma = 1e-4
    Set by hand in Section 2.2 as 'the common range ... we have set sigma=1e-4'; this choice affects the posterior and hence predictions.
assumptions (3)
  • ad hoc to paper The optimal strong threshold theta*(n) is a smooth function of n, so a GP with a stationary kernel can interpolate and extrapolate it.
    This is the core transfer assumption; Section 2.2 models f(n) as a GP, but Table 1 shows the grid-search optimum is highly oscillatory in n.
  • domain assumption The Ruge-Stuben/PMIS coarsening and hybrid Gauss-Seidel/Jacobi smoothing in hypre behave as standard AMG theory describes, and theta as defined in Definition 2.1 controls convergence as expected.
    The paper relies on hypre's implementation and the classical AMG framework without independently verifying it.
  • domain assumption Iteration count to reach relative residual tolerance 1e-8 is a sufficient proxy for computational efficiency.
    Used throughout Section 3 to define 'optimal' theta; wall-clock time is reported only for the final test cases.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Kernel-learning parameter prediction and evaluation in algebraic multigrid method for several PDEs." pith.science (2026). https://pith.science/paper/SCSPOTCQ

@misc{pith2026250414930,
  author       = {Pith},
  title        = {Pith review of: Kernel-learning parameter prediction and evaluation in algebraic multigrid method for several PDEs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SCSPOTCQ}},
  note         = {Machine review of arXiv:2504.14930}
}
read the original abstract

This paper explores the application of kernel learning methods for parameter prediction and evaluation in the Algebraic Multigrid Method (AMG), focusing on several Partial Differential Equation (PDE) problems. AMG is an efficient iterative solver for large-scale sparse linear systems, particularly those derived from elliptic and parabolic PDE discretizations. However, its performance heavily relies on numerous parameters, which are often set empirically and are highly sensitive to AMG's effectiveness. Traditional parameter optimization methods are either computationally expensive or lack theoretical support. To address this, we propose a Gaussian Process Regression (GPR)-based strategy to optimize AMG parameters and introduce evaluation metrics to assess their effectiveness. Trained on small-scale datasets, GPR predicts nearly optimal parameters, bypassing the time-consuming parameter sweeping process. We also use kernel learning techniques to build a kernel function library and determine the optimal kernel function through linear combination, enhancing prediction accuracy. In numerical experiments, we tested typical PDEs such as the constant-coefficient Poisson equation, variable-coefficient Poisson equation, diffusion equation, and Helmholtz equation. Results show that GPR-predicted parameters match grid search results in iteration counts while significantly reducing computational time. A comprehensive analysis using metrics like mean squared error, prediction interval coverage, and Bayesian information criterion confirms GPR's efficiency and reliability. These findings validate GPR's effectiveness in AMG parameter optimization and provide theoretical support for AMG's practical application.

Figures

Figures reproduced from arXiv: 2504.14930 by the authors.

Figure 1
Figure 1. The variation of the number of iterations for AMG solving the constant coefficient Poisson equation [PITH_FULL_IMAGE:figures/full_fig_p014_1.png] view at source ↗
Figure 2
Figure 2. Regression curves for predicting θ with respect to n using GPR for AMG solving the constant coefficient Poisson equation The following table contrasts the predicted outcomes with those acquired through the traversal method. It is clear that the difference in the number of iterations between the predicted and actual values is negligible, indicating that the predicted parameters are at least suboptimal for this specif… view at source ↗
Figure 3
Figure 3. When T = 2, the computational domain is uniformly partitioned into four blocks (Bi ,i = 1,2,3,4). Diffusion coefficient κ is the same in each block, when Bi ̸=Bj , ki ̸=kj . Specifically, we generate the training set starting from n = 64 with a step size of 16 until n=400. For each matrix generation, T randomly varies within the interval (10,20), and the random seed seed is equal to the index of the matrix. We will … view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Traversal plots for three sets of test matrices, along with iteration steps corresponding to GPR [PITH_FULL_IMAGE:figures/full_fig_p017_4.png]
Figure 5
Figure 5. Figure 5: Variation of the number of iterations for AMG solving the Helmholtz equation with coefficient [PITH_FULL_IMAGE:figures/full_fig_p019_5.png]
Figure 6
Figure 6. Figure 6: Regression curves of θ with respect to n predicted by GPR for AMG solving the Helmholtz equation with coefficient 2π [PITH_FULL_IMAGE:figures/full_fig_p019_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 34 canonical work pages

  1. [1]

    Accelerating algebraic multigrid methods via artificial neural networks

    Antonietti P F, Caldana M, Dede’ L. Accelerating algebraic multigrid methods via artificial neural networks. Vietnam Journal of Mathematics, 51 (2023), pp. 1-36

    work page 2023

  2. [2]

    Algebraic multigrid schemes for high-order nodal discontinuous Galerkin methods

    Antonietti P F, Melas L. Algebraic multigrid schemes for high-order nodal discontinuous Galerkin methods. SIAM Journal on Scientific Computing, 42 (2020), pp. A1147-A1173

    work page 2020

  3. [3]

    iFEM: an innovative finite element methods package in MATLAB

    Chen L. iFEM: an innovative finite element methods package in MATLAB. Preprint, Univer- sity of Maryland, 20, (2008)

    work page 2008

  4. [4]

    Gaussian processes for machine learning

    C K I Williams, C E Rasmussen. Gaussian processes for machine learning. Cambridge, MA: MIT press, (2006) 22

    work page 2006

  5. [5]

    hypre: A library of high performance preconditioners

    Falgout R D, Yang U M. hypre: A library of high performance preconditioners. Interna- tional Conference on computational science. Berlin, Heidelberg: Springer Berlin Heidelberg, (2002),pp. 632-641

    work page 2002

  6. [6]

    Limitations of Bayesian leave-one-out cross-validation for model selection

    Gronau Q F, Wagenmakers E J. Limitations of Bayesian leave-one-out cross-validation for model selection. Computational Brain & Behavior, 2 (2019), pp. 1-11

    work page 2019

  7. [7]

    Coarsening in algebraic multigrid using Gaussian processes

    Gottschalk H, Kahl K. Coarsening in algebraic multigrid using Gaussian processes. Elec- tronic Transactions on Numerical Analysis, 54 (2021), pp. 514-533

    work page 2021

  8. [8]

    FP-AMG: FPGA-based acceleration frame- work for algebraic multigrid solvers

    Haghi P , Geng T, Guo A, Wang T, Herbordt M. FP-AMG: FPGA-based acceleration frame- work for algebraic multigrid solvers. In: 2020 IEEE 28th Annual International Symposium on Field-Programmable Custom Computing Machines (FCCM), 2020, pp. 148-156

    work page 2020

Show all 34 references
  1. [9]

    Root mean square error (RMSE) or mean absolute error (MAE): When to use them or not

    Hodson T O. Root mean square error (RMSE) or mean absolute error (MAE): When to use them or not. Geoscientific Model Development Discussions, 2022, pp. 1-10

    2022

  2. [10]

    A general alternating-direction implicit framework with Gaussian process regression parameter prediction for large sparse linear systems

    Jiang K, Su X, Zhang J. A general alternating-direction implicit framework with Gaussian process regression parameter prediction for large sparse linear systems. SIAM Journal on Scientific Computing, 44 (2022), pp. A1960-A1988

    2022

  3. [11]

    Multitask kernel-learning parameter prediction method for solv- ing time-dependent linear systems

    Jiang K, Zhang J, Zhou Q. Multitask kernel-learning parameter prediction method for solv- ing time-dependent linear systems. CSIAM Transactions on Applied Mathematics, 4 (2023), pp. 672-695

    2023

  4. [12]

    Algebraic multi-grid for discrete elliptic second-order problems

    Kickinger F. Algebraic multi-grid for discrete elliptic second-order problems. In: Multigrid Methods V: Proceedings of the Fifth European Multigrid Conference held in Stuttgart, 1998, pp. 157-172

    1998

  5. [13]

    Algebraic multigrid

    Ruge J W, St ¨uben K. Algebraic multigrid. In: Multigrid Methods, 1987, pp. 73-130

    1987

  6. [14]

    GPU acceleration of algebraic multigrid preconditioners for discrete elliptic field problems

    Richter C, Sch ¨ops S, Clemens M. GPU acceleration of algebraic multigrid preconditioners for discrete elliptic field problems. IEEE Transactions on Magnetics, 50 (2014), pp. 461-464

    2014

  7. [15]

    Learning algebraic multigrid using graph neu- ral networks

    Luz I, Galun M, Maron H, Basri R, Yavneh I. Learning algebraic multigrid using graph neu- ral networks. In: International Conference on Machine Learning, 2020, pp. 6489-6499

    2020

  8. [16]

    Adaptive algebraic multigrid

    Brezina M, Falgout R, MacLachlan S, Mccormick T, Mccormick S, Ruge J. Adaptive algebraic multigrid. SIAM Journal on Scientific Computing, 27 (2006), pp. 1261-1286

    2006

  9. [17]

    The Bayesian information criterion: background, derivation, and applications

    Neath A A, Cavanaugh J E. The Bayesian information criterion: background, derivation, and applications. Wiley Interdisciplinary Reviews: Computational Statistics, 4 (2012), pp. 199-203

    2012

  10. [18]

    Algebraic multigrid for discontinuous Galerkin discretizations of heterogeneous elliptic problems

    Bastian P , Blatt M, Scheichl R. Algebraic multigrid for discontinuous Galerkin discretizations of heterogeneous elliptic problems. Numerical Linear Algebra with Applications, 19 (2012), pp. 367-388

    2012

  11. [19]

    Parameter selection in Gaussian process interpolation: an empirical study of selection criteria

    Petit S J, Bect J, Feliot P , Vazquez E. Parameter selection in Gaussian process interpolation: an empirical study of selection criteria. SIAM/ASA Journal on Uncertainty Quantification, 11 (2023), pp. 1308-1328

    2023

  12. [20]

    A scalable estimate of the out-of-sample prediction error via approxi- mate leave-one-out cross-validation

    Rad K R, Maleki A. A scalable estimate of the out-of-sample prediction error via approxi- mate leave-one-out cross-validation. Journal of the Royal Statistical Society Series B: Statis- tical Methodology, 82 (2020), pp. 965-996

    2020

  13. [21]

    & Bui D T

    Rahmati O, Choubin B, Fathabadi A, Coulon F, Soltani E, Shahabi H, ... & Bui D T. Predicting uncertainty of machine learning models for modelling nitrate pollution of groundwater us- ing quantile regression and UNEEC methods. Science of the Total Environment, 688 (2019), pp. 855-866

    2019

  14. [22]

    Robust Gaussian process modeling using EM algorithm

    Ranjan R, Huang B, Fatehi A. Robust Gaussian process modeling using EM algorithm. Jour- nal of Process Control, 42 (2016), pp. 125-136

    2016

  15. [23]

    On solving groundwater flow and transport models with algebraic 23 multigrid preconditioning

    Sbai M A, Larabi A. On solving groundwater flow and transport models with algebraic 23 multigrid preconditioning. Groundwater, 59 (2021), pp. 100-108

    2021

  16. [24]

    A tutorial on Gaussian process regression: Modelling, exploring, and exploiting functions

    Schulz E, Speekenbrink M, Krause A. A tutorial on Gaussian process regression: Modelling, exploring, and exploiting functions. Journal of Mathematical Psychology, 85 (2018), pp. 1-16

    2018

  17. [25]

    Machine learning approaches for estimation of prediction interval for the model output

    Shrestha D L, Solomatine D P . Machine learning approaches for estimation of prediction interval for the model output. Neural Networks, 19 (2006), pp. 225-235

    2006

  18. [26]

    Optimization-based algebraic multigrid coarsening using reinforcement learning

    Taghibakhshi A, MacLachlan S, Olson L, West M. Optimization-based algebraic multigrid coarsening using reinforcement learning. Advances in Neural Information Processing Sys- tems, 34 (2021), pp. 12129-12140

    2021

  19. [27]

    Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC

    Vehtari A, Gelman A, Gabry J. Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC. Statistics and Computing, 27 (2017), pp. 1413-1432

    2017

  20. [28]

    A widely applicable Bayesian information criterion

    Watanabe S. A widely applicable Bayesian information criterion. The Journal of Machine Learning Research, 14 (2013), pp. 867-897

    2013

  21. [29]

    Algebraic multigrid methods for saddle point systems arising from mortar contact formulations

    Wiesner T A, Mayr M, Popp A, Gee M W, Wall W A. Algebraic multigrid methods for saddle point systems arising from mortar contact formulations. International Journal for Numerical Methods in Engineering, 122 (2021), pp. 3749-3779

    2021

  22. [30]

    & Huang M

    Wan X, Li X, Wang X, Yi X, Zhao Y, He X, ... & Huang M. Water quality prediction model using Gaussian process regression based on deep learning for carbon neutrality in paper- making wastewater treatment system. Environmental Research, 211 (2022), pp. 112942

    2022

  23. [31]

    An aggregation-based algebraic multigrid method

    Notay Y. An aggregation-based algebraic multigrid method. Electronic Transactions on Nu- merical Analysis, 37 (2010), pp. 123-146

    2010

  24. [32]

    A supplementary strategy for coarsening in algebraic multigrid

    Ye S, Xu X, An H, Yang X. A supplementary strategy for coarsening in algebraic multigrid. Applied Mathematics and Computation, 394 (2021), pp. 125795

    2021

  25. [33]

    Gaussian process regression-based load forecasting model

    Yadav A, Bareth R, Kochar M, Pazoki M, Sehiemy R A E. Gaussian process regression-based load forecasting model. IET Generation, Transmission & Distribution, 18 (2024), pp. 899-910

    2024

  26. [34]

    AutoAMG (θ): An Auto-tuned AMG Method Based on Deep Learning for Strong Threshold

    Zou H, Xu X, Zhang C S, Mo Z Y. AutoAMG (θ): An Auto-tuned AMG Method Based on Deep Learning for Strong Threshold. Communications in Computational Physics, 36 (2024),pp. 200–220

    2024

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.