Pith. sign in

REVIEW 3 major objections 6 minor 41 references

SplineNet: An Isogeometric Deep Learning Method for Complex Shells

T0 review · 3 major / 6 minor · reviewed 2026-07-08 · glm-5.2

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

Pith's one-line read Spline Bases Replace Neural Activations in Shell Analysis Network

desk verdict SplineNet embeds ASUTS basis functions into a neural network via Bézier extraction, enabling IGA-consistent shell analysis in both solution-learning and operator-learning modes. The construction is sound and the engineering motivation is real, but P-SNet's optimization scalability is untested and O-SNet lacks a baseline comparison. read the letter →

arxiv 2607.06026 v1 pith:NBXINQT7 submitted 2026-07-07 cs.LG cs.NAmath.NA

Shizhou Luo , Xiaodong Wei This is my paper · ORCID
classification cs.LGcs.NAmath.NA
keywords isogeometricanalysisneuralnetworksBézierextractionT-splinesKirchhoff–LoveshellsoperatorlearningDeepONetstructural
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

SplineNet is a neural network architecture whose weights and activation functions are fixed by the mathematics of CAD spline surfaces, rather than learned freely. The first hidden layer evaluates the 16 bicubic Bernstein polynomials at the input parametric coordinates; the second hidden layer multiplies those values by a precomputed Bézier extraction matrix to produce analysis-suitable unstructured T-spline (ASUTS) basis functions; the output layer multiplies the basis functions by trainable control variables to predict displacement. This exactly recovers the discretization used in isogeometric analysis (IGA), meaning the network's approximation space is identical to that of a conventional spline-based solver. The paper uses this architecture in two modes: P-SNet, which trains control variables by minimizing the total potential energy of a Kirchhoff–Love shell without any precomputed solution data, and O-SNet, which replaces the trunk net of a Deep Operator Network with SplineNet so that the learned load-to-displacement operator lives in the spline space and predictions for new loadings require no re-solving. The paper tests both modes on shell structures including a Scordelis–Lo roof, a B-pillar, and a geometrically nonlinear plane nose, all represented by ASUTS meshes with extraordinary points, reporting displacement errors below 1% for P-SNet and 1.7–3.9% for O-SNet relative to IGA reference solutions.

What carries the argument

Bernstein polynomials as activation functions; Bézier extraction matrix as fixed weights; trainable control variables as output-layer weights; energy-based loss (P-SNet) or data-driven MSE loss (O-SNet); ASUTS basis functions as the network's approximation space.

What would settle it

Construct a loading function class whose displacement response contains modes outside the ASUTS spline space used by O-SNet; if the operator-learning error does not decrease with additional training data or wider branch nets, the fixed spline prior is the bottleneck.

Watch

Extended reading notes

Core claim

The central object is the elementwise SplineNet: a two-hidden-layer fully connected network in which the first layer's activation functions are the 16 bicubic Bernstein polynomials, the second layer's weights are the fixed Bézier extraction coefficients C_e, and the output layer's weights are the trainable control variables U_e. The output u_e(ξ,η) = (U_e)^T C_e B(ξ,η) is algebraically identical to the per-element IGA discretization. The paper shows that this construction can serve as a drop-in neural surrogate for IGA: as a physics-informed solver (P-SNet) it minimizes shell energy with no training data, and as an operator-learning trunk net (O-SNet) it maps loading functions to control变量s,

Load-bearing premise

The paper assumes that using fixed spline basis functions as the trunk net of an operator network does not sacrifice approximation power relative to a network with learned basis functions, but provides no theoretical approximation result and no head-to-head comparison against a standard DeepONet with an MLP trunk on the same problems.

Editorial extensions

If this is right

  • Engineering design loops that currently require repeated IGA or FEM solves for each new loading or boundary condition could be replaced by a single trained O-SNet forward pass, with predictions staying in the CAD-native spline representation.
  • Shape optimization can treat control-point positions as differentiable parameters within the same network, since geometry and displacement share the spline basis, enabling end-to-end gradient-based optimization without mesh-to-CAD remapping.
  • The fixed-basis trunk net removes the dependence of operator-learning accuracy on query-point placement, a known sensitivity in point-sampling-based methods like standard DeepONet.
  • The architecture extends to geometrically nonlinear shell formulations, as shown by the plane-nose example, suggesting the spline-prior approach is not limited to linear elasticity.

Reading between the lines

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

  • If the ASUTS spline space is not rich enough to represent a particular load-response pattern, O-SNet cannot recover the missing modes by learning, because the trunk net is fixed—its expressiveness is bounded by the spline basis. This sets a hard accuracy ceiling that a vanilla DeepONet with a learned MLP trunk would not face.
  • The energy-based P-SNet mode could in principle be combined with the data-driven O-SNet mode in a hybrid loss, using sparse experimental data to regularize the physics-informed solve—neither mode alone exploits this possibility.
  • Because each Bézier element is an independent sub-network with no inter-element connections, the architecture is naturally parallelizable across elements, which could make large-scale shell problems tractable on GPUs in ways that monolithic PINN architectures are not.
  • Replacing ASUTS with other spline types (e.g., hierarchical B-splines or LR-splines) would require only changing the extraction matrix, suggesting the method generalizes beyond T-splines to any spline technology with a Bézier extraction operator.
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 / 6 minor

Summary. This paper introduces SplineNet, a neural network architecture that embeds analysis-suitable unstructured T-splines (ASUTS) via Bézier extraction, with Bernstein polynomials as activations and extraction coefficients as fixed weights. Two variants are proposed: P-SNet, a data-free energy-based solver for Kirchhoff–Love shells, and O-SNet, which replaces the trunk net of DeepONet with SplineNet for operator learning. The construction is verified on the Scordelis–Lo roof benchmark, and O-SNet is demonstrated on three shell problems of increasing complexity, including a geometrically nonlinear case. The core mathematical construction is sound: Eq. (16) recovers the IGA discretization by construction, and the energy-based loss (Eq. 19) is a standard variational formulation.

Significance. The paper addresses a genuine gap: integrating watertight CAD spline representations into deep learning for shell analysis, including geometries with extraordinary points that go beyond simple multi-patch settings. The SplineNet construction is parameter-free in its trunk (fixed extraction coefficients, Bernstein activations), which is a clean and interpretable design. The application to operator learning on complex real-world geometries (B-pillar, plane nose) with 1.7–3.9% test errors is a useful practical demonstration. The CAD/CAE integration motivation is well-articulated and the approach is technically coherent.

major comments (3)
  1. §5.1, Table 1: The P-SNet errors are non-monotonic across mesh refinement: 0.027% (Mesh 0, 124 CPs), 0.003% (Mesh 1, 432 CPs), then 0.679% (Mesh 2, 1624 CPs). For the linear KL shell, the total potential energy (Eq. 14) is convex and quadratic in U, so a direct solver achieves machine precision regardless of problem size. The two-order-of-magnitude degradation on the finest mesh indicates that Adam-based minimization of Eq. (19) struggles as the dimension of U grows. Since P-SNet as a data-free solver is a central contribution, the authors should (a) discuss this scaling behavior explicitly rather than presenting all three results as equally validating, and (b) report whether alternative optimizers (e.g., L-BFGS), preconditioning, or longer training schedules mitigate the degradation. Without this, the practical viability of P-SNet beyond modest mesh sizes remains unclear.
  2. §4.3 and §5.2: O-SNet is presented as replacing the DeepONet trunk net with fixed spline basis functions, but no comparison against a standard DeepONet with an MLP trunk net is provided on any of the three test problems. This makes it impossible to assess the accuracy trade-off introduced by the spline prior. While the absolute errors (1.7–3.9%) are reasonable, the reader cannot determine whether the spline trunk helps, hurts, or is neutral relative to a learned trunk. A single head-to-head comparison on, e.g., the roof problem would substantially strengthen the contribution claim.
  3. §4.2, Eq. (19): The boundary condition term uses a penalty formulation with equally spaced boundary points, but the weight on this term relative to the energy is not specified. For the linear problems this appears to work, but the sensitivity of the solution accuracy to the penalty weight and the number of boundary points M is not reported. Since essential boundary condition enforcement is mentioned as a motivation (§1, citing conflicting gradient issues in PINNs), a brief sensitivity study or at least a statement of the chosen weights for each example would be appropriate.
minor comments (6)
  1. §5.2.4: The plane nose problem uses the geometrically nonlinear KL shell formulation [3,48], but the energy functional in Eq. (14) and the P-SNet loss in Eq. (19) are stated for the linear case. Since O-SNet uses a data-driven loss (Eq. 20), this is internally consistent, but the paper should clarify that the nonlinear physics enters only through the reference data, not through the network loss.
  2. §5.2.1: The GRF length scale l=0.2 is stated, but the amplitude factor γ varies per problem (1.5×10^8, 1.5×10^7, 5×10^8). The rationale for these specific values is not given. A brief note on how these were chosen would help reproducibility.
  3. Table 2 lists shell parameters for the linear cases, but the plane nose uses different values (E=2.06×10^11, t=0.03) stated only in the text. Consolidating all problem parameters into a single table would improve clarity.
  4. Fig. 13b, 17b, 21b: The y-axis labels 'Relative L2 error' and 'Relative L2 error (%)' are inconsistent between subfigures. Standardizing the format would help.
  5. §4.1, Remark 1 mentions that training the second-layer weights yields r-adaptivity but is not pursued. A citation to where this has been explored (e.g., HiDeNN r-adaptivity references) would contextualize the remark.
  6. §5.1: The Scordelis–Lo roof verification uses meshes with 'intentionally introduced' extraordinary points, but the number and valence of these EPs are not specified. This information is relevant for assessing the ASUTS treatment.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for a careful and constructive report. The referee correctly identifies the core mathematical construction as sound and acknowledges the practical motivation. The three major comments are all legitimate: (1) the non-monotonic P-SNet errors on the finest mesh require explicit discussion and investigation of alternative optimizers; (2) the absence of a head-to-head comparison between O-SNet and a standard DeepONet with an MLP trunk is a gap that should be filled; and (3) the penalty weight and boundary point sensitivity for essential boundary condition enforcement should be reported. We agree with all three points and will revise the manuscript accordingly. We provide point-by-point responses below.

read point-by-point responses

  1. Referee: §5.1, Table 1: The P-SNet errors are non-monotonic across mesh refinement: 0.027% (Mesh 0, 124 CPs), 0.003% (Mesh 1, 432 CPs), then 0.679% (Mesh 2, 1624 CPs). For the linear KL shell, the total potential energy (Eq. 14) is convex and quadratic in U, so a direct solver achieves machine precision regardless of problem size. The two-order-of-magnitude degradation on the finest mesh indicates that Adam-based minimization of Eq. (19) struggles as the dimension of U grows. Since P-SNet as a data-free solver is a central contribution, the authors should (a) discuss this scaling behavior explicitly rather than presenting all three results as equally validating, and (b) report whether alternative optimizers (e.g., L-BFGS), preconditioning, or longer training schedules mitigate the degradation. Without this, the practical viability of P-SNet beyond modest mesh sizes remains unclear.

    Authors: The referee is correct on both counts. The non-monotonicity in Table 1 is a real phenomenon, not a typographical error, and we agree that presenting all three results as equally validating without discussion is misleading. The degradation from 0.003% to 0.679% on Mesh 2 is attributable to the scaling behavior of Adam as the dimension of the control variable vector U grows from 432 to 1624. As the referee notes, the energy functional for the linear KL shell is convex and quadratic in U, so a direct solver would achieve machine precision; the observed degradation is entirely an optimizer limitation, not a discretization issue. We will revise the manuscript in two ways. First, we will add an explicit discussion of this scaling behavior in §5.1, acknowledging that Adam-based minimization of Eq. (19) becomes less effective as the problem dimension grows and that the Mesh 2 result should not be interpreted as evidence of discretization error but rather as an indicator of optimization difficulty. Second, we will conduct additional experiments using L-BFGS (both as a standalone optimizer and as a second-stage refinement after Adam) and report whether the finest-mesh accuracy improves. We expect L-BFGS to perform better given the convex quadratic structure of the loss landscape, but we will report the actual results honestly. If L-BFGS does not fully resolve the issue, we will state this plainly and discuss preconditioning and longer training schedules as further remedies to be explored in future work. We will not claim that P-SNet in its current form is practically viable for large-scale problems without these additional results. revision: yes

  2. Referee: §4.3 and §5.2: O-SNet is presented as replacing the DeepONet trunk net with fixed spline basis functions, but no comparison against a standard DeepONet with an MLP trunk net is provided on any of the three test problems. This makes it impossible to assess the accuracy trade-off introduced by the spline prior. While the absolute errors (1.7–3.9%) are reasonable, the reader cannot determine whether the spline trunk helps, hurts, or is neutral relative to a learned trunk. A single head-to-head comparison on, e.g., the roof problem would substantially strengthen the contribution claim.

    Authors: We agree that a head-to-head comparison against a standard DeepONet with an MLP trunk is essential to substantiate the claim that the spline prior is beneficial. The current manuscript presents only absolute errors for O-SNet without a baseline, which is insufficient. We will add a comparison on the roof problem (§5.2.2) between O-SNet and a standard DeepONet with an MLP trunk net of comparable parameter count, trained on the same dataset with the same branch net and optimization schedule. We will report relative L2 errors at Gaussian quadrature points and element corners for both architectures. We note that the comparison may not be purely about raw accuracy: the spline trunk provides interpretability (predictions live in the ASUTS spline space, enabling direct CAD/CAE integration) and independence from query point locations, which an MLP trunk does not. However, if the spline trunk also achieves comparable or better accuracy, this strengthens the contribution; if it is worse, we will report this honestly and frame the contribution in terms of interpretability and CAD integration rather than accuracy. We will not selectively report only favorable outcomes. revision: yes

  3. Referee: §4.2, Eq. (19): The boundary condition term uses a penalty formulation with equally spaced boundary points, but the weight on this term relative to the energy is not specified. For the linear problems this appears to work, but the sensitivity of the solution accuracy to the penalty weight and the number of boundary points M is not reported. Since essential boundary condition enforcement is mentioned as a motivation (§1, citing conflicting gradient issues in PINNs), a brief sensitivity study or at least a statement of the chosen weights for each example would be appropriate.

    Authors: The referee is correct that the penalty weight and the number of boundary points M are not specified in the manuscript, and this omission makes it difficult for readers to assess the robustness of the boundary condition enforcement. We will address this in the revision. First, we will state the specific penalty weights and values of M used in each example. In the current implementation, the boundary condition term in Eq. (19) has a coefficient of 1/M (as written), and the prescribed displacements are additionally imposed directly on the corresponding boundary control variables, which provides a strong enforcement mechanism beyond the penalty term alone. We will clarify this dual enforcement strategy in the revised text. Second, we will add a brief sensitivity study on the roof problem showing how the solution accuracy varies with the penalty weight (e.g., scaling the boundary loss by factors of 10, 1, 0.1 relative to the energy term) and with M. This will allow readers to gauge whether the method is robust to these hyperparameters or requires careful tuning. If the sensitivity study reveals that accuracy depends strongly on these choices, we will state this as a limitation. revision: yes

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: SplineNet recovers IGA discretization by construction (intended), energy loss is standard KL shell, O-SNet validated against external IGA solvers

full rationale

The paper's central construction is SplineNet, which embeds ASUTS basis functions into a neural network via Bézier extraction. Eq. 16 shows u_e(ξ,η) = (U_e)^T C_e B(ξ,η), which the paper explicitly states 'recovers the discretization of IGA.' This is the intended result of the construction, not a circular derivation: the paper is not claiming to derive IGA from neural networks, but rather showing that the neural network architecture is equivalent to the known IGA discretization. The Bézier extraction coefficients C_e come from the ASUTS framework [37, 42], which are external, independently developed mathematical objects, not fitted parameters. The P-SNet energy loss (Eq. 19) uses the standard Kirchhoff-Love total potential energy (Eq. 14), which is a well-known variational formulation, not a fitted functional. The O-SNet predictions are validated against external IGA solvers (in-house ASUTS solver [45] and tIGAr [49]), providing independent benchmarks. While the paper cites [37] (co-authored by X. Wei) for ASUTS, this is a legitimate reference to an independently published spline framework with its own mathematical properties (watertightness, C1-continuity, convergence), not a self-citation that smuggles in an unverified ansatz. The ASUTS framework is externally verifiable and has been developed with stated mathematical properties that do not depend on the present paper's results. No step in the derivation chain reduces to its own inputs by construction.

Assumptions & free parameters 7 free parameters · 5 assumptions · 3 invented entities

The axiom ledger reveals that SplineNet introduces no new physical entities or postulated phenomena — it is a computational method. The free parameters are standard neural network hyperparameters and problem-specific physical constants. The most consequential axiom is the last one (ASUTS spline space sufficiency for operator learning), which is ad hoc to this paper and only empirically supported without theoretical backing.

free parameters (7)
  • Control variables U (P-SNet) = trained via gradient descent on energy loss
    These are the primary unknowns of the shell problem, analogous to nodal DOFs in FEM. They are problem-specific and determined by training, not by the method.
  • Branch net parameters θ (O-SNet) = trained via MSE loss on IGA solver data
    MLP weights mapping loading functions to control variables. Standard neural network training parameters.
  • Branch net width (O-SNet) = 100/300/500 (roof), 250/500/750 (B-pillar, plane nose)
    Hyperparameter selected by performance comparison, not derived from theory.
  • GRF length scale l = 0.2
    Controls smoothness of generated loading functions. Stated in §5.2.1 without sensitivity analysis.
  • GRF amplitude factor γ = 1.5e8 (roof), 1.5e7 (B-pillar), 5e8 (plane nose)
    Problem-specific loading magnitude, chosen per example.
  • Number of training epochs = 3000
    Fixed across all examples, with staged learning rate schedule.
  • Number of training/test samples = 1000/250
    Fixed split across all O-SNet examples without justification for sufficiency.
assumptions (5)
  • domain assumption Analysis-suitable unstructured T-splines (ASUTS) provide watertight, C1-continuous, linearly independent basis functions with optimal convergence on complex geometries with extraordinary points.
    Stated in §2 and cited to [37]. This is the foundational spline theory the method depends on. The paper does not re-derive ASUTS properties but relies on them entirely.
  • domain assumption The Kirchhoff-Love shell model with small-deformation assumption (linear KL) accurately represents the mechanical behavior of thin shells for the roof and B-pillar problems.
    Invoked in §3 and §5. The linear KL model [43] is adopted as the model problem. Validity depends on thin-shell and small-deformation assumptions being satisfied.
  • standard math Bézier extraction operators C_e from ASUTS can be directly used as fixed neural network weights without loss of the spline approximation properties.
    §4.1, Remark 1. The paper states this is equivalent to IGA discretization by construction (Eq. 16). Mathematically sound but assumed without formal proof in the neural network context.
  • domain assumption Gaussian quadrature points used as training input coordinates provide sufficient integration accuracy for the energy loss evaluation.
    §4.2: 'the input parametric coordinates are chosen to be the Gaussian points during training.' Standard in IGA but the interaction with stochastic gradient descent training is not analyzed.
  • ad hoc to paper The ASUTS spline space is sufficiently rich to serve as the trunk net basis for DeepONet operator learning of the load-to-displacement mapping.
    §4.3: O-SNet 'introduces a prior to fulfill the same task' as a learned trunk net. No theoretical justification is provided; only empirical evidence from three examples supports this.
invented entities (3)
  • SplineNet (the neural network architecture) independent evidence
    purpose: Embeds ASUTS basis functions into a neural network via Bézier extraction with Bernstein polynomial activations
    The architecture is validated against IGA reference solutions on four benchmark problems with quantified relative errors (Tables 1–5). The construction is falsifiable: if Eq. 16 did not recover IGA discretization, the P-SNet results would not match IGA solutions.
  • P-SNet (Physics-informed SplineNet) independent evidence
    purpose: Data-free energy-based solver using SplineNet with KL shell energy loss
    Validated on Scordelis-Lo roof benchmark with <1% error across three meshes (Table 1). The benchmark has known reference solutions.
  • O-SNet (Operator SplineNet) independent evidence
    purpose: Operator learning variant where SplineNet serves as DeepONet trunk net
    Validated on three shell problems with 1.7–3.9% test errors against IGA solver outputs (Tables 3–5). However, no comparison against vanilla DeepONet is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SplineNet: An Isogeometric Deep Learning Method for Complex Shells." pith.science (2026). https://pith.science/paper/NBXINQT7

@misc{pith2026260706026,
  author       = {Pith},
  title        = {Pith review of: SplineNet: An Isogeometric Deep Learning Method for Complex Shells},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NBXINQT7}},
  note         = {Machine review of arXiv:2607.06026}
}
read the original abstract

We present a novel isogeometric deep learning method, termed SplineNet, for the seamless design and analysis of shell structures with complex geometries. The proposed approach is built upon watertight spline representations, e.g., analysis-suitable unstructured T-splines, and features exact geometric descriptions of Computer-Aided Design (CAD) models in neural networks. B\'ezier extraction is used to build the network architecture, where Bernstein polynomials serve as the nonlinear activation functions. SplineNet can be applied in a data-free or data-driven way. In the data-free case, energy-based formulations can be naturally incorporated as loss terms, which fulfill the need of Computer-Aided Engineering (CAE) and can be accurately calculated. In particular, the Kirchhoff--Love (KL) model is adopted to solve for the mechanical behaviors of shell structures. This way, CAD and CAE can be tightly integrated in a deep neural network without the time-consuming model/data exchange process. In the data-driven case, SplineNet can be used as the trunk net of Deep Operator Networks (DeepONet) to provide interpretability. Given such a trained network and unseen input data, results can be immediately obtained without retraining the network or repeatedly performing the traditional workflow for analysis. In the end, a variety of numerical examples are studied to demonstrate the effectiveness of the proposed method, especially when real-world complex geometries are involved.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 41 canonical work pages

  1. [1]

    John Wiley & Sons, Chich- ester, UK (2009)

    Cottrell, J.A., Hughes, T.J., Bazilevs, Y.: Isogeometric Analysis: Toward Integration of CAD and FEA. John Wiley & Sons, Chich- ester, UK (2009). https://doi.org/10.1002/ 9780470749081

    work page 2009

  2. [2]

    McGraw-Hill, New York (1993)

    Reddy, J.N.: An Introduction to the Finite Element Method, 2nd edn. McGraw-Hill, New York (1993)

    work page 1993

  3. [3]

    Computer methods in applied mechanics and engineering198(49- 52), 3902–3914 (2009) https://doi.org/10.1 016/j.cma.2009.08.013

    Kiendl, J., Bletzinger, K.-U., Linhard, J., W¨ uchner, R.: Isogeometric shell analysis with Kirchhoff–Love elements. Computer methods in applied mechanics and engineering198(49- 52), 3902–3914 (2009) https://doi.org/10.1 016/j.cma.2009.08.013

    work page 2009

  4. [4]

    Computer Methods in Applied Mechanics and Engineering199(17- 20), 1216–1230 (2010) https://doi.org/10.1 016/j.cma.2009.12.010

    Nagy, A.P., Abdalla, M.M., G¨ urdal, Z.: Isogeometric sizing and shape optimisation of beam structures. Computer Methods in Applied Mechanics and Engineering199(17- 20), 1216–1230 (2010) https://doi.org/10.1 016/j.cma.2009.12.010

    work page 2010

  5. [5]

    Com- puter Methods in Applied Mechanics and Engineering351, 330–350 (2019) https://do i.org/10.1016/j.cma.2019.03.042

    Beck, J., Tamellini, L., Tempone, R.: IGA- based multi-index stochastic collocation for random PDEs on arbitrary domains. Com- puter Methods in Applied Mechanics and Engineering351, 330–350 (2019) https://do i.org/10.1016/j.cma.2019.03.042

    work page doi:10.1016/j.cma.2019.03.042 2019

  6. [6]

    Applied Mechanics Reviews, 1–81 (2026) https://doi.org/10.1115/1.4071710

    Wang, Y., Bai, J., Lin, Z., Wang, Q., Anitescu, C., Sun, J., Eshaghi, M.S., Gu, Y., Feng, X.-Q., Zhuang, X., et al.: Arti- ficial intelligence for partial differential equations in computational mechanics: a review. Applied Mechanics Reviews, 1–81 (2026) https://doi.org/10.1115/1.4071710

    work page doi:10.1115/1.4071710 2026

  7. [7]

    Raissi and P

    Raissi, M., Perdikaris, P., Karniadakis, G.E.: Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computa- tional physics378, 686–707 (2019) https: //doi.org/10.1016/j.jcp.2018.10.045

    work page doi:10.1016/j.jcp.2018.10.045 2019

  8. [8]

    Communications in Mathematics and Statistics6(1), 1–12 (2018) https://doi.org/10.1007/s40304-018-0127-z

    E, W., Yu, B.: The deep Ritz method: a deep learning-based numerical algorithm for solv- ing variational problems. Communications in Mathematics and Statistics6(1), 1–12 (2018) https://doi.org/10.1007/s40304-018-0127-z

    work page doi:10.1007/s40304-018-0127-z 2018

Show all 41 references
  1. [10]

    Computer Methods in Applied Mechan- ics and Engineering374, 113547 (2021) https://doi.org/10.1016/j.cma.2020.113547

    Kharazmi, E., Zhang, Z., Karniadakis, G.E.: hp-VPINNs: variational physics-informed neural networks with domain decomposition. Computer Methods in Applied Mechan- ics and Engineering374, 113547 (2021) https://doi.org/10.1016/j.cma.2020.113547

    2021 doi

  2. [11]

    Communications in Computational Physics33(5), 1240–1269 (2023) https://do i.org/10.4208/cicp.OA-2022-0239

    Basir, S.: Investigating and mitigating failure modes in physics-informed neural networks (PINNs). Communications in Computational Physics33(5), 1240–1269 (2023) https://do i.org/10.4208/cicp.OA-2022-0239

    2023 doi

  3. [12]

    arXiv preprint arXiv:2412.09752 (2024) https://doi.org/10 .48550/arXiv.2412.09752 arXiv:2412.09752

    Chickering, K.R.: A quasilinear algorithm for computing higher-order derivatives of deep feed-forward neural networks. arXiv preprint arXiv:2412.09752 (2024) https://doi.org/10 .48550/arXiv.2412.09752 arXiv:2412.09752

    2024 arXiv

  4. [13]

    Jour- nal of Computational Physics451, 110839 (2022) https://doi.org/10.1016/j.jcp.2021.1 10839

    Fuhg, J.N., Bouklas, N.: The mixed deep energy method for resolving concentration features in finite strain hyperelasticity. Jour- nal of Computational Physics451, 110839 (2022) https://doi.org/10.1016/j.jcp.2021.1 10839

    2022 doi

  5. [14]

    Computer Methods in Applied Mechanics and Engineering437, 117787 (2025) https://doi.org/10.1016/j.cma.2025.1 17787

    Bai, J., Lin, Z., Wang, Y., Wen, J., Liu, Y., Rabczuk, T., Gu, Y., Feng, X.-Q.: Energy- based physics-informed neural network for frictionless contact problems under large deformation. Computer Methods in Applied Mechanics and Engineering437, 117787 (2025) https://doi.org/10.1...

    2025 doi

  6. [15]

    SIAM Journal on Scientific Computing43(5), 3055–3081 (2021) https://doi.org/10.1137/20m1318043

    Wang, S., Teng, Y., Perdikaris, P.: Under- standing and mitigating gradient flow pathologies in physics-informed neural networks. SIAM Journal on Scientific Computing43(5), 3055–3081 (2021) https://doi.org/10.1137/20m1318043

    2021 doi

  7. [16]

    Heliyon9(8), 18820 (2023) ht t p s : //doi.org/10.1016/j.heliyon.2023.e18820

    Berrone, S., Canuto, C., Pintore, M., Suku- mar, N.: Enforcing Dirichlet boundary con- ditions in physics-informed neural networks and variational physics-informed neural net- works. Heliyon9(8), 18820 (2023) ht t p s : //doi.org/10.1016/j.heliyon.2023.e18820

    2023 doi

  8. [18]

    Computational Mechanics67(1), 207–230 (2021) https://do i.org/10.1007/s00466-020-01928-9

    Zhang, L., Cheng, L., Li, H., Gao, J., Yu, C., Domel, R., Yang, Y., Tang, S., Liu, W.K.: Hierarchical deep-learning neural networks: finite elements and beyond. Computational Mechanics67(1), 207–230 (2021) https://do i.org/10.1007/s00466-020-01928-9

    2021 doi

  9. [19]

    Computational mechanics 72(1), 173–194 (2023) https://doi.org/10.1 007/s00466-023-02293-z

    Liu, Y., Park, C., Lu, Y., Mojumder, S., Liu, W.K., Qian, D.: HiDeNN-FEM: a seamless machine learning approach to nonlinear finite element analysis. Computational mechanics 72(1), 173–194 (2023) https://doi.org/10.1 007/s00466-023-02293-z

    2023

  10. [20]

    Computational Mechanics72(2), 333–362 (2023) https://doi.org/10.1007/s0 0466-023-02336-5

    Lu, Y., Li, H., Zhang, L., Park, C., Mojumder, S., Knapik, S., Sang, Z., Tang, S., Apley, D.W., Wagner, G.J.,et al.: Convo- lution hierarchical deep-learning neural net- works (C-HiDeNN): finite elements, isogeo- metric analysis, tensor decomposition, and beyond. Computational...

    2023 doi

  11. [21]

    Computa- tional Mechanics, 1–21 (2025) https://doi.or g/10.1007/s00466-025-02677-3

    ˇSkardov´ a, K., Daby-Seesaram, A., Genet, M.: Finite element neural network interpo- lation: Part I—interpretable and adaptive discretization for solving PDEs. Computa- tional Mechanics, 1–21 (2025) https://doi.or g/10.1007/s00466-025-02677-3

    2025 doi

  12. [22]

    Computational Mechanics, 1–26 (2025) https://doi.org/10.1007/s00466-025-0 2676-4

    Daby-Seesaram, A., ˇSkardov´ a, K., Genet, M.: Finite element neural network interpolation: Part II—hybridisation with the proper gener- alised decomposition for non-linear surrogate 23 modelling. Computational Mechanics, 1–26 (2025) https://doi.org/10.1007/s00466-025-0 2676-4

    2025 doi

  13. [24]

    Journal of Machine Learning Research 24(388), 1–26 (2023)

    Li, Z., Huang, D.Z., Liu, B., Anandkumar, A.: Fourier neural operator with learned deformations for PDEs on general geome- tries. Journal of Machine Learning Research 24(388), 1–26 (2023)

    2023

  14. [25]

    arXiv preprint arXiv:2003.03485 (2020) https:// d o i

    Li, Z., Kovachki, N., Azizzadenesheli, K., Liu, B., Bhattacharya, K., Stuart, A., Anandku- mar, A.: Neural operator: graph kernel net- work for partial differential equations. arXiv preprint arXiv:2003.03485 (2020) https:// d o i . o r g / 1 0 . 4 8 5 5 0 / a r X i v . 2 0 0 3...

    2003 arXiv

  15. [26]

    Nature machine intelligence3(3), 218–229 (2021) ht tps://doi.org/10.1038/s42256-021-00302-5

    Lu, L., Jin, P., Pang, G., Zhang, Z., Kar- niadakis, G.E.: Learning nonlinear opera- tors via DeepONet based on the universal approximation theorem of operators. Nature machine intelligence3(3), 218–229 (2021) ht tps://doi.org/10.1038/s42256-021-00302-5

    2021 doi

  16. [28]

    Neural Networks198, 108560 (2026) https: //doi.org/10.1016/j.neunet.2026.108560

    Park, J., Kang, N.: Point-DeepONet: pre- dicting nonlinear fields on non-parametric geometries under variable load conditions. Neural Networks198, 108560 (2026) https: //doi.org/10.1016/j.neunet.2026.108560

    2026 doi

  17. [29]

    National Science Open 3(6), 20240001 (2024) https://doi.org/10.1 360/nso/20240001

    Chen, G., Liu, X., Meng, Q., Chen, L., Liu, C., Li, Y.: Learning neural operators on Rie- mannian manifolds. National Science Open 3(6), 20240001 (2024) https://doi.org/10.1 360/nso/20240001

    2024

  18. [30]

    Computer Meth- ods in Applied Mechanics and Engineering 403, 115671 (2023) https://doi.org/10.1016/ j.cma.2022.115671

    Wu, C., Zhu, M., Tan, Q., Kartha, Y., Lu, L.: A comprehensive study of non-adaptive and residual-based adaptive sampling for physics- informed neural networks. Computer Meth- ods in Applied Mechanics and Engineering 403, 115671 (2023) https://doi.org/10.1016/ j.cma.2022.115671

    2023

  19. [31]

    International Journal of Solids and Structures276, 112321 (2023) ht tps://doi.org/10.1016/j.ijsolstr.2023.112321

    Gu, Y., Zhang, C., Zhang, P., Golub, M.V., Yu, B.: Enriched physics-informed neural net- works for 2D in-plane crack analysis: theory and MATLAB code. International Journal of Solids and Structures276, 112321 (2023) ht tps://doi.org/10.1016/j.ijsolstr.2023.112321

    2023 doi

  20. [32]

    Computer Methods in Applied Mechanics and Engineer- ing452, 118723 (2026) https://doi.org/10.1 016/j.cma.2025.118723

    Visser, C., Heinlein, A., Giovanardi, B.: PAC- MANN: point adaptive collocation method for artificial neural networks. Computer Methods in Applied Mechanics and Engineer- ing452, 118723 (2026) https://doi.org/10.1 016/j.cma.2025.118723

    2026

  21. [33]

    Computer Methods in Applied Mechanics and Engineering405, 115839 (2023) https://doi.org/10.1016/j.cma.2022.1 15839

    Gasick, J., Qian, X.: Isogeometric neural networks: a new deep learning approach for solving parameterized partial differen- tial equations. Computer Methods in Applied Mechanics and Engineering405, 115839 (2023) https://doi.org/10.1016/j.cma.2022.1 15839

    2023 doi

  22. [35]

    Com- puter Methods in Applied Mechanics and Engineering434, 117582 (2025) https://doi

    Zhang, L., Park, C., Hughes, T.J., Liu, W.K.: Multi-patch isogeometric convolution hierar- chical deep-learning neural network. Com- puter Methods in Applied Mechanics and Engineering434, 117582 (2025) https://doi. org/10.1016/j.cma.2024.117582

    2025 doi

  23. [36]

    Engineering with Computers42(4), 116 (2026) ht t p s : //doi.org/10.1007/s00366-026-02351-z

    Tresckow, M., Ion, I.G., Loukrezis, D.: Multi- patch neural solver with isogeometric map- pings for partial differential equations on 24 computer-aided design domains. Engineering with Computers42(4), 116 (2026) ht t p s : //doi.org/10.1007/s00366-026-02351-z

    2026 doi

  24. [38]

    Morgan Kauf- mann, San Francisco, CA, USA (2001)

    Rogers, D.F.: An Introduction to NURBS: with Historical Perspective. Morgan Kauf- mann, San Francisco, CA, USA (2001)

    2001

  25. [39]

    ACM transactions on graphics (TOG)22(3), 477– 484 (2003) https://doi.org/10.1145/1201775

    Sederberg, T.W., Zheng, J., Bakenov, A., Nasri, A.: T-splines and T-NURCCs. ACM transactions on graphics (TOG)22(3), 477– 484 (2003) https://doi.org/10.1145/1201775. 882295

    2003 doi

  26. [40]

    Jour- nal of Approximation Theory90(2), 174–199 (1997) https://doi.org/10.1006/jath.1996.30 79

    Reif, U.: A refineable space of smooth spline surfaces of arbitrary topological genus. Jour- nal of Approximation Theory90(2), 174–199 (1997) https://doi.org/10.1006/jath.1996.30 79

    1997 doi

  27. [42]

    International Journal for Numerical Methods in Engineering88(2), 126–156 (2011) https://doi.org/10.1002/nm e.3167

    Scott, M.A., Borden, M.J., Verhoosel, C.V., Sederberg, T.W., Hughes, T.J.: Isogeometric finite element data structures based on B´ ezier extraction of T-splines. International Journal for Numerical Methods in Engineering88(2), 126–156 (2011) https://doi.org/10.1002/nm e.3167

    2011 doi

  28. [44]

    In: Advances in Neural Infor- mation Processing Systems, vol

    Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L.,et al.: PyTorch: an imperative style, high-performance deep learning library. In: Advances in Neural Infor- mation Processing Systems, vol. 32, pp. 8024– 8035 (2019)

    2019

  29. [45]

    Computer Methods in Applied Mechanics and Engineering444, 118114 (2025) https://doi.org/10.1016/j.cm a.2025.118114

    Sheng, J., Wei, X.: Isogeometric topology optimization of thin-walled structures with complex design domains. Computer Methods in Applied Mechanics and Engineering444, 118114 (2025) https://doi.org/10.1016/j.cm a.2025.118114

    2025 doi

  30. [46]

    In: Interna- tional Conference on Learning Representa- tions (2015)

    Kingma, D.P., Ba, J.: Adam: a method for stochastic optimization. In: Interna- tional Conference on Learning Representa- tions (2015). https://doi.org/10.48550/arXiv .1412.6980 .https://arxiv.org/abs/1412.6980

    2015 doi arXiv

  31. [47]

    In: Proceedings of the IEEE Conference on Com- puter Vision and Pattern Recognition, pp

    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: Proceedings of the IEEE Conference on Com- puter Vision and Pattern Recognition, pp. 770–778 (2016). https://doi.org/10.1109/CV PR.2016.90

    2016 doi

  32. [48]

    Com- puter Methods in Applied Mechanics and Engineering291, 280–303 (2015) https://do i.org/10.1016/j.cma.2015.03.010

    Kiendl, J., Hsu, M.-C., Wu, M.C., Reali, A.: Isogeometric Kirchhoff–Love shell formula- tions for general hyperelastic materials. Com- puter Methods in Applied Mechanics and Engineering291, 280–303 (2015) https://do i.org/10.1016/j.cma.2015.03.010

    2015 doi

  33. [49]

    Com- puter Methods in Applied Mechanics and Engineering344, 477–498 (2019) https://do i.org/10.1016/j.cma.2018.10.002 25

    Kamensky, D., Bazilevs, Y.: tIGAr: automat- ing isogeometric analysis with FEniCS. Com- puter Methods in Applied Mechanics and Engineering344, 477–498 (2019) https://do i.org/10.1016/j.cma.2018.10.002 25

    2019 doi

Pith tools

Reviewed July 8, 2026 · model on record in the stance chip above.