Pith. sign in

REVIEW 3 major objections 5 minor 55 references

Parametric Neural r-Adaptivity for Isogeometric Analysis via Residual Minimization

T0 review · 3 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read A mesh-size-weighted residual estimator, minimized over knot positions, is a valid and differentiable training loss for r-adaptive isogeometric analysis, reducing H1 error at fixed degrees of freedom.

desk verdict A solid, honestly-scoped neural r-adaptivity method whose main new trick is a level-independent knot density; the reliability proof has a genuine gap in its spline interpolation estimates, but it is not a show-stopper. read the letter →

arxiv 2607.21753 v1 pith:PB3FF3ON submitted 2026-07-23 math.NA cs.NA

Elías Caru , David Pardo , Judit Muñoz-Matute This is my paper · ORCID
classification math.NAcs.NA MSC 65N3065D0765N5068T07
keywords r-adaptivityisogeometricanalysisresidualminimizationaposteriorierrorestimationknotrelocationneuralmeshpredictionparametricPDEsautomaticdifferentiation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper argues that knot relocation in isogeometric analysis should be driven by the classical a posteriori residual estimator—element residuals weighted by local mesh size, plus interface flux jumps and Neumann boundary residuals—rather than by the strong-form residual used in physics-informed neural networks. In the coercive regime with admissible mesh families, this estimator is reliable and locally efficient up to oscillation, so minimizing it with respect to knot positions provably reduces the energy (H1) error while keeping the number of degrees of freedom fixed. Because the loss needs no minimization principle, the same construction extends differentiable r-adaptivity to indefinite and advection-dominated problems, where it acts as a mesh-quality functional rather than a certified bound. The parametric version trains one network to output a knot-density function sampled at any resolution, so a single trained network predicts admissible meshes at every refinement level. If true, this gives a parameter-free, error-driven way to place knots that recovers optimal convergence rates for singularities and cuts error constants in boundary layers.

What carries the argument

The load-bearing object is the residual estimator η(θ)² = Σ_E ρ_E²‖R_E‖² + Σ_I (h_I/σ_I)‖J_I‖² + Σ_B (h_B/σ_B)‖N_B‖², where ρ_E is a mesh- and coefficient-weighted element size, R_E is the strong residual inside an element, J_I is the normal-flux jump across interfaces, and N_B is the Neumann residual. This computable quantity is used as the differentiable loss L = η²/2, and minimizing it over knot parameters θ replaces a strong-form PINN loss that would control a norm stronger than H1. The second mechanism is the knot-density network: a coordinate-based fully connected network takes a parameter ν and a reference coordinate, and a fixed differentiable map (centering, tanh saturation, softmax

What would settle it

Compute the effectivity index I_eff = η/|u*−u_θ|_{H1} for a coercive problem with a known singular solution while refining; if I_eff grows without bound, or if for a convection-dominated family with ε→0 the adapted mesh's H1 error stops beating the uniform mesh at fixed degrees of freedom, the claim that this loss controls the energy error fails in that regime.

Watch

Extended reading notes

Core claim

The central claim is that the residual estimator (13) is the right training loss for r-adaptive isogeometric analysis. The paper proves (Proposition 3.2, Theorem 3.3) that under coercivity and admissible-mesh assumptions, the estimator bounds the energy error from above and locally from below up to oscillation; hence minimizing it over interior knots at fixed degrees of freedom reduces the H1 error. The method keeps the Galerkin solver untouched: a neural network relocates knots, the solution is always a conforming Galerkin solution, and gradients through the solve are obtained by the discrete adjoint, realized by reverse-mode automatic differentiation. In the parametric setting, the network

Load-bearing premise

The error-control guarantee collapses if the coefficient-weighted spline interpolation inequality or the local quasi-uniformity and admissibility conditions fail, since the paper relies on the former from the literature without proof and enforces the latter only partially.

Editorial extensions

If this is right

  • For coercive diffusion–reaction and advection–diffusion–reaction problems on admissible mesh families, minimizing the estimator with fixed degrees of freedom reduces the H1 error, with the reliability constant independent of the knot parameters.
  • One trained density network predicts an admissible mesh at any refinement level, including levels finer or coarser than those seen in training, enabling coarse-to-fine continuation without retraining.
  • The residual-based loss removes the symmetry/coercivity restriction of Ritz-based r-adaptivity, so the same pipeline applies to indefinite Helmholtz and convection-dominated problems, although without the reliability guarantee in those regimes.
  • Mesh gradients cost one adjoint solve, so the method is end-to-end differentiable and the memory cost does not grow with the internal depth of the linear solver.
  • In the benchmarks, the method recovers optimal N^{-p} rates for a one-dimensional singular power (error 349× smaller at N=64 for cubics), reduces constants for Helmholtz contrasts and arctangent layers, and improves fixed-DOF errors on an L-shaped corner and an advection–diffusion boundary layer.

Reading between the lines

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

  • The paper stops at tensor-product knot lines; if the residual estimator is combined with hierarchical splines, the corner-singularity ceiling (effective order near 2) may be lifted to the optimal rate, a testable structural extension the paper itself names.
  • The level-independence of the density output suggests a time-dependent reading: adding time as an input coordinate would allow one network to track moving layers and fronts, and the fixed loss should transfer without architectural change.
  • The normalized loss (28) could serve as a cheap online error oracle for unseen parameters: one Galerkin solve plus estimator evaluation gives a mesh-quality check without any per-instance optimization, which the paper does not exploit.
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

3 major / 5 minor

Summary. The paper proposes a neural r-adaptive IGA method in which an MLP maps a PDE parameter to a knot-density field, the density is converted by a fixed differentiable map into an admissible knot vector, the discrete solution is obtained by a standard conforming Galerkin solve, and the training loss is a classical residual-based a posteriori estimator: element residuals weighted by local mesh size, interface flux jumps, and Neumann residuals. In the coercive, admissible-mesh setting the authors prove reliability and local efficiency of the estimator (Prop. 3.2, Thm. 3.3), and they derive the discrete-adjoint gradient for the knot parameters. Five parametric benchmarks (1D singular powers, 1D Helmholtz transmission, 2D arctangent layer, 2D L-shape, 2D advection-diffusion boundary layer) show fixed-DOF error reductions. The manuscript explicitly acknowledges that three of the five experiments operate outside the assumptions of the theory.

Significance. The central idea is attractive: rather than using a PINN strong residual or an approximated dual norm, the paper trains the mesh by minimizing a classical a posteriori error estimator, and the density-based parametrization makes the trained network independent of the refinement level. If the theoretical results are fully established, this would be a useful and principled contribution to differentiable r-adaptivity beyond Ritz formulations. The paper is also explicitly honest about its scope, reports errors against exact/manufactured/reference solutions, and gives a detailed training protocol. However, the proof of the central reliability bound rests on a weighted interpolation estimate for high-continuity spline spaces that is asserted but not proved, and the certified regime is validated numerically only in 1D; the 2D evidence is mostly outside the assumptions. The significance is therefore conditional on closing that gap and on strengthening the in-theory numerical validation.

major comments (3)
  1. [Section 3.1, proof of Prop. 3.2, Eq. (17)] The bound \|u−u_θ\|_E ≤ C_rel η(θ) is the central theoretical justification for the training loss. Its proof requires a quasi-interpolant I_h into the θ-dependent spline spaces V_h(θ) satisfying coefficient- and reaction-weighted L2 and trace estimates with constants uniform in θ. The manuscript cites the unweighted spline estimates to [16] and asserts that the weighted form follows 'by combining them patchwise with the scaling arguments of [54,55]'. References [54,55] treat standard finite element spaces, not globally C^{p−1} B-splines; the extension requires a Scott–Zhang-type projector that respects global smoothness and trace estimates with constants independent of θ. As written, this is a gap in the proof, not a scope limitation: if the weighted estimates fail or the constant depends on θ, the chain (17) breaks and minimizing η is not certified to control the H1 error. Please state
  2. [Section 5, Tables 1–5] Of the five experiments, only Experiment 1 (1D singular power) lies fully inside Assumptions I–II. Experiment 2 is indefinite Helmholtz (Remark 3.2), Experiment 4 is a non-conforming immersed discretization (Remark 3.3), and Experiment 5 is convection-dominated/anisotropic (Remarks 3.1 and 3.2); Experiment 3 is coercive but the reported meshes are anisotropic, so it is outside Assumption II's shape-regularity requirement. The paper is explicit about these departures, and I do not treat them as deceptive. Nevertheless, the central claim that the residual loss is a certified objective for r-adaptivity is empirically validated only in 1D. Since the method is intended for 2D parametric problems, please add at least one 2D benchmark satisfying Assumptions I and II (e.g., a coercive problem with a point singularity and shape-regular grading), or qualify the main claim so that the certified sta
  3. [Section 4.2, Prop. 4.1] The uniform-reliability statement inherits the interpolation-estimate gap from Prop. 3.2. In addition, the proof that sup_ν C_rel(ν) < ∞ assumes that the quasi-interpolant constants are uniform in ν and that Assumption II holds for all admissible meshes. The density parametrization of Section 4.1 enforces the h_min floor and the grading cap, but it does not enforce shape regularity in 2D; the paper itself concedes this in Remark 3.1. The proposition should therefore be stated with the caveat that its hypotheses include the unproved weighted interpolation property and the shape-regularity condition, or it should be restricted to the cases where those hypotheses are verified.
minor comments (5)
  1. [Section 5.2, paragraph 'Three implementation details'] The free-split treatment inserts the interface knot x_I by sorting. Sorting is a non-differentiable operation, and reverse-mode AD through it is not the same as differentiating the loss with respect to the pre-sort element sizes. Please clarify how the chain rule is defined when the relative order of breakpoints changes, or restrict the free-split treatment to configurations where the interface remains between the same two movable knots.
  2. [Equation (33)] There is a typographical comma in the numerator: the notation |u^* − u(·),ν|_{H^1} should presumably read |u^* − u(·;ν)|_{H^1}.
  3. [Section 5.3] The phrase 'on the parameter grid of [1]' should state the grid values explicitly; the reader should not need to consult [1] to know the training/testing parameter set.
  4. [Tables 1 and 4] The text describes effectivity indices as 'stable under refinement', but in Table 1 the adapted p=3 values decrease from about 25 to about 7, and in Table 4 the adapted values vary non-monotonically up to 48. 'Stable' is too strong; 'tracking the error trend with O(1) values' would be more accurate.
  5. [Section 5.4, final paragraph] The zero-shot evaluation at N=64 shows a visible drop in effective order for p=3 (local slope 1.14). This is an honest observation, but the general claim that 'one trained network serves every refinement level' should be tempered by this example.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the residual loss is a standard a posteriori estimator, reported errors use independent exact/reference solutions, and self-citations are not load-bearing.

full rationale

The central loop is not circular. The training loss L(θ)=η²(θ)/2 (Eq. 14) is a classical residual-type a posteriori estimator (Eq. 13); its reliability and local efficiency (Prop. 3.2, Thm. 3.3) are argued in the paper from coercivity, Galerkin orthogonality, and interpolation/bubble estimates attributed to external references [16,54,55], none of which are by the present authors. The reported H¹ errors are computed against exact solutions (Experiments 1,2,3,5), a high-degree independent reference with a self-convergence study (Experiment 4), or, for the effectivity index, the same estimator used only as a diagnostic; the improvement factors use Eq. 33, not the training loss. The normalized loss (Eq. 28) divides by η(θ_unif) only to make terms dimensionless; the paper explicitly states it is 'not a reference error and needs no precomputed optimal meshes.' No constants are fitted to H¹ errors, and the held-out test protocol separates training from evaluation. Self-citations [1,35,43,42,52,53] are contextual or comparative and do not carry the proof burden of the central claim. The unproved coefficient-weighted spline Scott–Zhang estimates invoked in the proof of Prop. 3.2—[54,55] are standard finite-element results, not B-spline results—are a rigor/correctness concern, not circularity, since the paper does not define those estimates in terms of the target result. The paper also concedes the key restrictions on admissible meshes, coercivity, and conforming discretizations in Remarks 3.1–3.3 and Section 5.4, further confirming that the claim is scoped rather than definitionally forced.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central method rests on classical a posteriori estimates plus standard coercivity and mesh-regularity assumptions. The main input-side free choices are the grading cap T, the minimum element size hmin, and a few experiment-specific safeguards (water-filling cap, resonance exclusion, normalization ε). No new physical entities are introduced; the knot-density function is a representation, not an entity.

free parameters (5)
  • T (saturation/grading cap) = 2–7 depending on experiment (Exp1: 2→5 for p=2, 6→7 for p=3; Exp2–5: T=5)
    Caps the maximal element-size ratio at e^{2T}; chosen by hand per experiment and refinement level; larger T permits stronger grading and improves reported error reductions, so the headline improvement factors partly reflect this choice.
  • hmin (minimum element size) = 1e-7 or 1e-8
    Floor in the size map (27) that prevents element collapse; also determines how fine clustering can become near singularities; set ad hoc per experiment.
  • water-filling cap h_e ≤ 2π/(2.5 k_loc) = depends on local wavenumber k_loc
    Added safeguard in the indefinite Helmholtz experiment (Sec. 5.2) to enforce Nyquist sampling on coarse meshes at high contrast; not part of the generic construction and affects Helmholtz results.
  • resonance exclusion half-width = 0.02
    Training/evaluation parameter set excludes contrasts within ±0.02 of resonances of the transmission problem; this selection shapes the indefinite-regime experiment.
  • ε (normalization guard) = 1e-12
    Prevents division by zero in the normalized loss (28); inconsequential but an arbitrary hyperparameter.
assumptions (5)
  • standard math Coefficient- and reaction-weighted Clement/Scott-Zhang interpolation estimates hold for spline spaces on admissible meshes (proof cited to [16,54,55])
    Used in the proof of Prop. 3.2 to bound the three residual terms by ∥v∥_E; not proved in the paper.
  • domain assumption Assumption I: coercivity — μ=α−½∇·β ≥ 0 and Γ_in ⊆ Γ_D
    Restricts the certified theory to coercive diffusion–reaction and advection–diffusion–reaction problems; violated by Helmholtz and indefinite cases.
  • domain assumption Assumption II: admissible meshes — uniform shape regularity, local quasi-uniformity h_E ≤ γ_loc h_E', bi-Lipschitz geometry map
    Required for the interpolation and trace estimates in Prop. 3.2; the softmax/hmin construction ensures h_E ≥ h_min and γ_loc ≤ e^{2T}, but shape regularity in 2D tensor-product meshes is not guaranteed (Remark 3.1).
  • domain assumption Exact Dirichlet data imposition (u_{D,h}=u_D on Γ_D)
    Assumed in Prop. 3.2; the paper notes an oscillation term appears otherwise; the immersed L-shape experiment enforces the boundary only approximately via a C0 cut, so the theory does not cover it.
  • domain assumption Reference solution in L-shaped experiment (p=5 immersed solution with same C0 cut) is accurate to ≲2e-5 relative H1
    Used to compute reported errors and convergence rates in Sec. 5.4; supported by a self-convergence study said to be included in the code, but no code link is given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Parametric Neural r-Adaptivity for Isogeometric Analysis via Residual Minimization." pith.science (2026). https://pith.science/paper/PB3FF3ON

@misc{pith2026260721753,
  author       = {Pith},
  title        = {Pith review of: Parametric Neural r-Adaptivity for Isogeometric Analysis via Residual Minimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PB3FF3ON}},
  note         = {Machine review of arXiv:2607.21753}
}
read the original abstract

We propose an r-adaptive neural algorithm for Isogeometric Analysis (IGA) based on residual minimization. The boundary-value problem is solved using a standard conforming Galerkin formulation, while a neural network relocates the interior knots. A strong-form residual in the sense of physics-informed neural networks (PINNs) controls a norm stronger than the energy (H^1) error. We therefore weight it by classical a posteriori theory: element residuals scaled by the local mesh size, interface flux jumps, and Neumann boundary residuals yield a computable estimator of the energy error, which we minimize with respect to the knots. For coercive problems on admissible mesh families, this estimator is reliable and locally efficient up to oscillation terms; beyond that regime, the same loss remains well-defined and extends differentiable r-adaptivity to indefinite and advection-dominated problems. In the parametric setting, the network maps each parameter to a knot-density function in a single evaluation; since it outputs a density rather than a fixed-dimensional vector of knot locations, one trained network produces an admissible mesh at any refinement level. Mesh gradients are obtained by reverse-mode automatic differentiation through the discrete solution equation. Numerical experiments in one and two dimensions illustrate that the method concentrates degrees of freedom near singularities, material interfaces, and boundary layers, improving accuracy for a fixed number of degrees of freedom.

Figures

Figures reproduced from arXiv: 2607.21753 by the authors.

Figure 1
Figure 1. Knot redistribution for the one-dimensional contrast Helmholtz problem (free-split treat [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Residual-informed neural mesh flowchart. [PITH_FULL_IMAGE:figures/full_fig_p013_2.png] view at source ↗
Figure 3
Figure 3. Block decomposition on the L-shaped domain. The interface lines at [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figures from the paper (18 more)
Figure 4
Figure 4. Figure 4: Experiment 1. Relative H1 -seminorm error versus the number of elements N for the uniform and r-adaptive meshes. Curves are medians and shaded bands the interquartile range, over seeds and the held-out test exponents. 5.2 One-dimensional Helmholtz regime We now conside…
Figure 5
Figure 5. Figure 5: Experiment 1. Log–log profiles of the exact second derivative and that of the numerical approximation near x = 0, for the uniform (blue) and adapted (orange) meshes, at ν = 1.6038 and N = 8. 0 0.2 0.4 0.6 0.8 1 0 0.5 1 u(x) u ∗ uh uθ 0 0.2 0.4 0.6 0.8 1 0 0.5 1 x 0 0.2…
Figure 6
Figure 6. Figure 6: Experiment 1. Exact solutions and their uniform (uh, blue) and r-adaptive (uθ, orange) approximations at N = 8, for the representative exponents ν ∈ {1.6038, 1.6974, 1.8037}. Markers indicate Greville abscissae. 20 [PITH_FULL_IMAGE:figures/full_fig_p020_6.png]
Figure 7
Figure 7. Figure 7: Experiment 1. Training and validation losses Jres versus Adam epoch at the fixed refinement level N = 4, for (a) p = 2 and (b) p = 3. with u(0) = 0, the Neumann condition σu′ (1) = 10π, and continuity of u and of the flux σu′ at the interface xI = 1/2. The local wavenu…
Figure 8
Figure 8. Figure 8: Experiment 2. Relative H1 -seminorm error versus N for the uniform and r-adaptive meshes. Curves are medians and shaded bands the interquartile range, over the four seeds and the held-out test contrasts. 22 [PITH_FULL_IMAGE:figures/full_fig_p022_8.png]
Figure 9
Figure 9. Figure 9: Experiment 2. Exact solution and its uniform (uh, blue) and r-adaptive (uθ, orange) approximations near resonance, at c = 5 and N = 32. 100 101 102 10−0.8 10−0.6 10−0.4 Epoch Jres = η 2 (θφ; ν)/η 2 (θ0; ν) (a) Quadratic B-splines (p = 2). 100 101 102 10−1 10−0.5 Epoch …
Figure 10
Figure 10. Figure 10: Experiment 2. Training and validation losses Jres versus Adam epoch at the fixed refinement level N = 64, for (a) p = 2 and (b) p = 3. 23 [PITH_FULL_IMAGE:figures/full_fig_p023_10.png]
Figure 11
Figure 11. Figure 11: Experiment 3. Relative weighted H1 -seminorm error versus N for the uniform and r-adaptive meshes. Curves are medians and shaded bands the interquartile range, over seeds and the held-out test parameters. 0 0.5 1 0 0.5 1 0 0.5 1 0 0.5 1 0 0.5 1 0 0.5 1 0 2 4 0 0.5 1 0…
Figure 12
Figure 12. Figure 12: Experiment 3. Predicted meshes and solutions at N = 16 for two representative values of ν = (t, s1, s2), one per row: upper panels (a)–(c) correspond to ν = (3.92, 0.54, 0.54) and lower panels (a)–(c) to (19.12, 0.46, 0.46). 25 [PITH_FULL_IMAGE:figures/full_fig_p025_…
Figure 13
Figure 13. Figure 13: Experiment 3. Training and validation losses Jres versus Adam epoch at the fixed refinement level N = 4, for (a) p = 2 and (b) p = 3. the embedding square (0, 1)2 , in the spirit of the Finite Cell Method [38, 45, 46]: a single tensor￾product spline space in which the…
Figure 14
Figure 14. Figure 14: Experiment 4. Relative error in the σ-weighted H1 seminorm (40) versus the number of elements per axis N for the uniform and r-adaptive meshes (median over seeds and the held-out test parameters, measured against the high-degree immersed reference), trained with batch…
Figure 15
Figure 15. Figure 15: Experiment 4. Predicted meshes and solutions at N = 16 for two representative parameter pairs, one per row: upper panels (a)–(c) correspond to (σ1, σ2) = (0.89, 0.89) and lower panels (a)–(c) to (10, 0.1). 28 [PITH_FULL_IMAGE:figures/full_fig_p028_15.png]
Figure 16
Figure 16. Figure 16: Experiment 4. Detail of the solution gradient near the re-entrant corner at (σ1, σ2) = (0.89, 0.89) and N = 16, with the knot lines superimposed. Panels (a)–(c) share a common color scale. 100 101 102 10−0.5 10−0.4 10−0.3 Epoch Jres = η 2 (θφ; ν)/η 2 (θ0; ν) (a) Quadr…
Figure 17
Figure 17. Figure 17: Experiment 4. Training and validation losses Jres versus Adam epoch at the fixed refinement level N = 8, for (a) p = 2 and (b) p = 3. 29 [PITH_FULL_IMAGE:figures/full_fig_p029_17.png]
Figure 18
Figure 18. Figure 18: Experiment 5. Relative H1 -seminorm error versus N for the uniform and r-adaptive meshes (median over seeds and the held-out test parameters), trained with batch size 16. 0 0.5 1 0 0.2 0.4 0.6 0 0.5 1 0 0.5 1 (a) Uniform mesh, p = 2 0 0.5 1 (b) r-adaptive mesh, p = 2 …
Figure 19
Figure 19. Figure 19: Experiment 5. Predicted meshes at N = 16 over the solution field, for two represen￾tative values of ν (one per row): upper panels (a)–(c) correspond to ν = (−1.58, 0.5) and lower panels (a)–(c) to ν = (−1.95, 2.0). 31 [PITH_FULL_IMAGE:figures/full_fig_p031_19.png]
Figure 20
Figure 20. Figure 20: Experiment 5. Cross-section of the solution at y = 0.5 and N = 16 for p = 2, for three representative values of ν, shown in panels (a)–(c). Each panel compares the exact solution with its uniform and r-adaptive approximations; markers indicate the Greville abscissae o…
Figure 21
Figure 21. Figure 21: Experiment 5. Training and validation losses Jres versus Adam epoch at the fixed refinement level N = 4, for (a) p = 2 and (b) p = 3. 32 [PITH_FULL_IMAGE:figures/full_fig_p032_21.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 15 canonical work pages

  1. [16]

    Mathematical Founda- tions of Adaptive Isogeometric Analysis

    A. Buffa, G. Gantner, C. Giannelli, D. Praetorius, and R. V´ azquez. “Mathematical Founda- tions of Adaptive Isogeometric Analysis”. In: Archives of Computational Methods in Engi- neering 29.7 (2022), pp. 4479–4555. doi: 10.1007/s11831-022-09752-5

    doi:10.1007/s11831-022-09752-5 2022

  2. [1]

    An r-adaptive finite element method using neural networks for parametric self- adjoint elliptic problems

    D. Aballay, F. Fuentes, V. Iligaray, ´A. J. Omella, D. Pardo, M. A. S´ anchez, I. Tapia, and C. Uriarte. “An r-adaptive finite element method using neural networks for parametric self- adjoint elliptic problems”. In: Journal of Computational Physics 545 (2026), p. 114447. doi: 10.1016/j.jcp.2025.114447. arXiv: 2504.21160 [math.NA]

    arXiv 2026

  3. [2]

    Ainsworth and J

    M. Ainsworth and J. T. Oden. A Posteriori Error Estimation in Finite Element Analysis . John Wiley & Sons, 2000. isbn: 9780471294115

    2000

  4. [3]

    T. Apel. Anisotropic Finite Elements: Local Estimates and Applications. Advances in Numer- ical Mathematics. Stuttgart: Teubner, 1999

    1999

  5. [4]

    A posteriori error estimation of residual type for anisotropic diffusion-convection-reaction problems

    T. Apel, S. Nicaise, and D. Sirch. “A posteriori error estimation of residual type for anisotropic diffusion-convection-reaction problems”. In: Journal of Computational and Applied Mathe- matics 235.8 (2011), pp. 2805–2820. doi: 10.1016/j.cam.2010.11.032

    work page doi:10.1016/j.cam.2010.11.032 2011

  6. [5]

    Direct and inverse error estimates for finite elements with mesh refinements

    I. Babuˇ ska, R. B. Kellogg, and J. Pitk¨ aranta. “Direct and inverse error estimates for finite elements with mesh refinements”. In: Numerische Mathematik 33.4 (1979), pp. 447–471. doi: 10.1007/BF01399326

    doi:10.1007/bf01399326 1979

  7. [6]

    Is the pollution effect of the FEM avoidable for the Helmholtz equation considering high wave numbers?

    I. M. Babuˇ ska and S. A. Sauter. “Is the pollution effect of the FEM avoidable for the Helmholtz equation considering high wave numbers?” In: SIAM Journal on Numerical Analysis 34.6 (1997), pp. 2392–2423. doi: 10.1137/S0036142994269186

    doi:10.1137/s0036142994269186 1997

  8. [7]

    Bangerth and R

    W. Bangerth and R. Rannacher. Adaptive Finite Element Methods for Differential Equations . Lectures in Mathematics. ETH Z¨ urich. Basel: Birkh¨ auser, 2003.doi: 10.1007/978-3-0348- 7605-6

    doi:10.1007/978-3-0348- 2003

Show all 55 references
  1. [8]

    Automatic differentiation in machine learning: A survey

    A. G. Baydin, B. A. Pearlmutter, A. A. Radul, and J. M. Siskind. “Automatic differentiation in machine learning: A survey”. In: Journal of Machine Learning Research 18.153 (2018), pp. 1–43. url: http://jmlr.org/papers/v18/17-468.html

    2018

  2. [9]

    Iso- geometric analysis: Approximation, stability and error estimates for h-refined meshes

    Y. Bazilevs, L. Beir˜ ao da Veiga, J. A. Cottrell, T. J. R. Hughes, and G. Sangalli. “Iso- geometric analysis: Approximation, stability and error estimates for h-refined meshes”. In: Mathematical Models and Methods in Applied Sciences 16.07 (2006), pp. 1031–1090. doi: 10.1142/...

    2006 doi

  3. [10]

    Efficient and modular implicit differentiation

    M. Blondel, Q. Berthet, M. Cuturi, R. Frostig, S. Hoyer, F. Llinares-L´ opez, F. Pedregosa, and J.-P. Vert. “Efficient and modular implicit differentiation”. In: Advances in Neural In- formation Processing Systems . Vol. 35. 2022, pp. 5230–5242. url: https://proceedings. neuri...

    2022

  4. [12]

    Good Approximation by Splines with Variable Knots. II

    C. de Boor. “Good Approximation by Splines with Variable Knots. II”. In: Conference on the Numerical Solution of Differential Equations . Vol. 363. Lecture Notes in Mathematics. Springer, 1974, pp. 12–20. doi: 10.1007/BFb0069121

    1974 doi

  5. [13]

    C. de Boor. A Practical Guide to Splines . Revised. Vol. 27. Applied Mathematical Sciences. Springer, 2001. doi: 10.1007/978-1-4612-6333-3

    2001 doi

  6. [14]

    Bradbury, R

    J. Bradbury, R. Frostig, P. Hawkins, M. J. Johnson, C. Leary, D. Maclaurin, G. Necula, A. Paszke, J. VanderPlas, S. Wanderman-Milne, and Q. Zhang. JAX: composable transforma- tions of Python+NumPy programs . 2018. url: https://github.com/google/jax

    2018

  7. [15]

    Adaptivity with moving grids

    C. J. Budd, W. Huang, and R. D. Russell. “Adaptivity with moving grids”. In: Acta Numerica 18 (2009), pp. 111–241. doi: 10.1017/S0962492906400015

    2009 doi

  8. [17]

    Fictitious domain finite element methods using cut elements: II. A stabilized Nitsche method

    E. Burman and P. Hansbo. “Fictitious domain finite element methods using cut elements: II. A stabilized Nitsche method”. In: Applied Numerical Mathematics 62.4 (2012), pp. 328–341. doi: 10.1016/j.apnum.2011.01.008

    2012 doi

  9. [18]

    Approximation by finite element functions using local regularization

    P. Cl´ ement. “Approximation by finite element functions using local regularization”. In:RAIRO Analyse Num´ erique9.R2 (1975), pp. 77–84. doi: 10.1051/m2an/197509R200771

    1975 doi

  10. [19]

    J. A. Cottrell, T. J. R. Hughes, and Y. Bazilevs. Isogeometric Analysis: Toward Integra- tion of CAD and FEA . John Wiley & Sons, 2009. isbn: 9780470748732. doi: 10 . 1002 / 9780470749081

    2009

  11. [20]

    A Posteriori Error Estimation for Highly Indefinite Helmholtz Problems

    W. D¨ orfler and S. A. Sauter. “A Posteriori Error Estimation for Highly Indefinite Helmholtz Problems”. In: Computational Methods in Applied Mathematics 13.3 (2013), pp. 333–347. doi: 10.1515/cmam-2013-0008

    2013 doi

  12. [21]

    Imposing Dirichlet boundary conditions with Nitsche’s method and spline-based finite elements

    A. Embar, J. Dolbow, and I. Harari. “Imposing Dirichlet boundary conditions with Nitsche’s method and spline-based finite elements”. In: International Journal for Numerical Methods in Engineering 83.7 (2010), pp. 877–898. doi: 10.1002/nme.2863

    2010 doi

  13. [22]

    New anisotropic a priori error estimates

    L. Formaggia and S. Perotto. “New anisotropic a priori error estimates”. In: Numerische Mathematik 89.4 (2001), pp. 641–667. doi: 10.1007/s002110100273

    2001 doi

  14. [23]

    Deep reinforcement learning for adaptive mesh refinement

    C. Foucart, A. Charous, and P. F. J. Lermusiaux. “Deep reinforcement learning for adaptive mesh refinement”. In: Journal of Computational Physics 491 (2023), p. 112381. doi: 10 . 1016/j.jcp.2023.112381

    2023

  15. [25]

    Grisvard

    P. Grisvard. Elliptic Problems in Nonsmooth Domains . Vol. 24. Monographs and Studies in Mathematics. Boston: Pitman, 1985

    1985

  16. [26]

    Hinze, R

    M. Hinze, R. Pinnau, M. Ulbrich, and S. Ulbrich. Optimization with PDE Constraints. Vol. 23. Mathematical Modelling: Theory and Applications. Springer, 2009. doi: 10.1007/978- 1- 4020-8839-1. 35

    2009 doi

  17. [27]

    Huang and R

    W. Huang and R. D. Russell. Adaptive Moving Mesh Methods. Vol. 174. Applied Mathematical Sciences. Springer, 2010. doi: 10.1007/978-1-4419-7916-2

    2010 doi

  18. [28]

    Isogeometric analysis: CAD, finite elements, NURBS, exact geometry and mesh refinement

    T. J. R. Hughes, J. A. Cottrell, and Y. Bazilevs. “Isogeometric analysis: CAD, finite elements, NURBS, exact geometry and mesh refinement”. In: Computer Methods in Applied Mechanics and Engineering 194.39–41 (2005), pp. 4135–4195. doi: 10.1016/j.cma.2004.10.008

    2005 doi

  19. [29]

    Ihlenburg

    F. Ihlenburg. Finite Element Analysis of Acoustic Scattering. Vol. 132. Applied Mathematical Sciences. Springer, 1998. doi: 10.1007/978-1-4757-2969-6

    1998 doi

  20. [30]

    Finite element solution of the Helmholtz equation with high wave number. Part I: The h-version of the FEM

    F. Ihlenburg and I. Babuˇ ska. “Finite element solution of the Helmholtz equation with high wave number. Part I: The h-version of the FEM”. In: Computers & Mathematics with Appli- cations 30.9 (1995), pp. 9–37. doi: 10.1016/0898-1221(95)00144-N

    1995 doi

  21. [31]

    hp-VPINNs: Variational physics-informed neural networks with domain decomposition

    E. Kharazmi, Z. Zhang, and G. E. M. Karniadakis. “hp-VPINNs: Variational physics-informed neural networks with domain decomposition”. In: Computer Methods in Applied Mechanics and Engineering 374 (2021), p. 113547. doi: 10.1016/j.cma.2020.113547

    2021

  22. [32]

    Adam: A method for stochastic optimization

    D. P. Kingma and J. Ba. “Adam: A method for stochastic optimization”. In: International Conference on Learning Representations. 2015. url: https://arxiv.org/abs/1412.6980

    2015 arXiv

  23. [33]

    Robust local problem error estimation for a singularly perturbed problem on anisotropic finite element meshes

    G. Kunert. “Robust local problem error estimation for a singularly perturbed problem on anisotropic finite element meshes”. In: ESAIM: Mathematical Modelling and Numerical Anal- ysis 35.6 (2001), pp. 1079–1109. doi: 10.1051/m2an:2001148

    2001 doi

  24. [34]

    J.-L. Lions. Optimal Control of Systems Governed by Partial Differential Equations . Springer- Verlag, 1971

    1971

  25. [35]

    Isogeometric Residual Minimization Method (iGRM) with direction splitting for non-stationary advection–diffusion problems

    M. Lo´ s, J. Mu˜ noz-Matute, I. Muga, and M. Paszy´ nski. “Isogeometric Residual Minimization Method (iGRM) with direction splitting for non-stationary advection–diffusion problems”. In: Computers & Mathematics with Applications 79.2 (2021), pp. 213–229. doi: 10.1016/j. camwa....

    2021 doi

  26. [36]

    Magueresse and S

    A. Magueresse and S. Badia. Energy minimisation using overlapping tensor-product free-knot B-splines. 2025. doi: 10.48550/arXiv.2508.17705 . arXiv: 2508.17705 [math.NA]. url: https://arxiv.org/abs/2508.17705

    2025 doi arXiv

  27. [37]

    r-Adaptive deep learning method for solving partial differential equations

    ´A. J. Omella and D. Pardo. “ r-Adaptive deep learning method for solving partial differential equations”. In: Computers & Mathematics with Applications 153 (2024), pp. 33–42. doi: 10.1016/j.camwa.2023.11.005

    2024 doi

  28. [38]

    Finite cell method: h- and p-extension for embedded domain problems in solid mechanics

    J. Parvizian, A. D¨ uster, and E. Rank. “Finite cell method: h- and p-extension for embedded domain problems in solid mechanics”. In: Computational Mechanics 41.1 (2007), pp. 121–133. doi: 10.1007/s00466-007-0173-y

    2007 doi

  29. [39]

    A posteriori error estimators for elliptic equations with discontinuous coeffi- cients

    M. Petzoldt. “A posteriori error estimators for elliptic equations with discontinuous coeffi- cients”. In: Advances in Computational Mathematics 16.1 (2002), pp. 47–75. doi: 10.1023/A: 1014221125034

    2002 doi

  30. [40]

    Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial dif- ferential equations

    M. Raissi, P. Perdikaris, and G. E. Karniadakis. “Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial dif- ferential equations”. In: Journal of Computational Physics 378 (2019), pp. 686–707. doi: 10.1...

    2019 doi

  31. [41]

    Adaptive optimization of isogeometric multi-patch dis- cretizations using artificial neural networks

    D. R ´ ıos, F. Scholz, and T. Takacs. “Adaptive optimization of isogeometric multi-patch dis- cretizations using artificial neural networks”. In: Computer Methods in Applied Mechanics and Engineering 432 (2024), p. 117400. doi: 10.1016/j.cma.2024.117400. 36

    2024

  32. [42]

    Robust Variational Physics-Informed Neural Networks

    S. Rojas, P. Maczuga, J. Mu˜ noz-Matute, D. Pardo, and M. Paszy´ nski. “Robust Variational Physics-Informed Neural Networks”. In: Computer Methods in Applied Mechanics and Engi- neering 425 (2024), p. 116904. doi: 10.1016/j.cma.2024.116904

    2024

  33. [43]

    Goal-oriented adaptivity for a con- forming residual minimization method in a dual discontinuous Galerkin norm

    S. Rojas, D. Pardo, P. Behnoudfar, and V. M. Calo. “Goal-oriented adaptivity for a con- forming residual minimization method in a dual discontinuous Galerkin norm”. In: Computer Methods in Applied Mechanics and Engineering 377 (2021), p. 113686. doi: 10.1016/j.cma. 2021.113686

    2021

  34. [44]

    Robust a-posteriori estimator for advection-diffusion-reaction problems

    G. Sangalli. “Robust a-posteriori estimator for advection-diffusion-reaction problems”. In: Mathematics of Computation 77.261 (2008), pp. 41–70. doi: 10 . 1090 / S0025 - 5718 - 07 - 02018-2

    2008

  35. [45]

    Isogeometric colloca- tion: Cost comparison with Galerkin methods and extension to adaptive hierarchical NURBS discretizations

    D. Schillinger, J. A. Evans, A. Reali, M. A. Scott, and T. J. R. Hughes. “Isogeometric colloca- tion: Cost comparison with Galerkin methods and extension to adaptive hierarchical NURBS discretizations”. In: Computer Methods in Applied Mechanics and Engineering 267 (2013), pp. ...

    2013 doi

  36. [46]

    The Finite Cell Method: A review in the context of higher-order structural analysis of CAD and image-based geometric models

    D. Schillinger and M. Ruess. “The Finite Cell Method: A review in the context of higher-order structural analysis of CAD and image-based geometric models”. In: Archives of Computa- tional Methods in Engineering 22.3 (2015), pp. 391–455. doi: 10.1007/s11831-014-9115-y

    2015 doi

  37. [47]

    C. Schwab. p- and hp-Finite Element Methods: Theory and Applications in Solid and Fluid Mechanics. Oxford: Oxford University Press, 1998

    1998

  38. [48]

    Finite element interpolation of nonsmooth functions satisfying boundary conditions

    L. R. Scott and S. Zhang. “Finite element interpolation of nonsmooth functions satisfying boundary conditions”. In: Mathematics of Computation 54.190 (1990), pp. 483–493. doi: 10.1090/S0025-5718-1990-1011446-7

    1990 doi

  39. [49]

    Implicit Neu- ral Representations with Periodic Activation Functions

    V. Sitzmann, J. N. P. Martel, A. W. Bergman, D. B. Lindell, and G. Wetzstein. “Implicit Neu- ral Representations with Periodic Activation Functions”. In: Advances in Neural Information Processing Systems. Vol. 33. 2020, pp. 7462–7473

    2020

  40. [50]

    M2N: Mesh Movement Networks for PDE Solvers

    W. Song, M. Zhang, J. G. Wallwork, J. Gao, Z. Tian, F. Sun, M. D. Piggott, J. Chen, Z. Shi, X. Chen, and J. Wang. “M2N: Mesh Movement Networks for PDE Solvers”. In: Advances in Neural Information Processing Systems . Vol. 35. 2022. arXiv: 2204.11188 . url: https:// proceedings...

    2022 arXiv

  41. [51]

    Deep Fourier Residual method for solv- ing time-harmonic Maxwell’s equations

    J. M. Taylor, M. Bastidas, D. Pardo, and I. Muga. “Deep Fourier Residual method for solv- ing time-harmonic Maxwell’s equations”. In: Journal of Computational Physics 523 (2025), p. 113623. doi: 10.1016/j.jcp.2024.113623

    2025

  42. [52]

    A Deep Fourier Residual method for solving PDEs using neural networks

    J. M. Taylor, D. Pardo, and I. Muga. “A Deep Fourier Residual method for solving PDEs using neural networks”. In: Computer Methods in Applied Mechanics and Engineering 405 (2023), p. 115850. doi: 10.1016/j.cma.2022.115850

    2023

  43. [53]

    A Deep Double Ritz method (D 2RM) for solving partial differential equations using neural networks

    C. Uriarte, D. Pardo, I. Muga, and J. Mu˜ noz-Matute. “A Deep Double Ritz method (D 2RM) for solving partial differential equations using neural networks”. In: Computer Methods in Applied Mechanics and Engineering 405 (2023), p. 115892. doi: 10 . 1016 / j . cma . 2023 . 115892

    2023

  44. [54]

    Robust a posteriori error estimates for stationary convection-diffusion equa- tions

    R. Verf¨ urth. “Robust a posteriori error estimates for stationary convection-diffusion equa- tions”. In: SIAM Journal on Numerical Analysis 43.4 (2005), pp. 1766–1782. doi: 10.1137/ 040604261. 37

    2005

  45. [55]

    Robust a posteriori error estimators for a singularly perturbed reaction-diffusion equation

    R. Verf¨ urth. “Robust a posteriori error estimators for a singularly perturbed reaction-diffusion equation”. In: Numerische Mathematik 101.4 (2005), pp. 645–684. doi: 10.1007/s00211- 005-0662-x

    2005 doi

  46. [56]

    Verf¨ urth

    R. Verf¨ urth. A Posteriori Error Estimation Techniques for Finite Element Methods . Oxford University Press, 2013. isbn: 9780199679423. doi: 10.1093/acprof:oso/9780199679423. 001.0001

    2013

  47. [57]

    Efficient r-adaptive isogeometric anal- ysis with Winslow’s mapping and monitor function approach

    G. Xu, B. Li, L. Shu, L. Chen, J. Xu, and T. Khajah. “Efficient r-adaptive isogeometric anal- ysis with Winslow’s mapping and monitor function approach”. In: Journal of Computational and Applied Mathematics 351 (2019), pp. 186–197. doi: 10.1016/j.cam.2018.11.003. 38

    2019 doi

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.