Pith. sign in

REVIEW 3 major objections 6 minor 54 references

Autoencoder-based non-intrusive model order reduction in continuum mechanics

T0 review · 3 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read The paper's core claim is that a three-stage autoencoder pipeline, trained purely on finite element snapshots, predicts accurate full-field solutions and boundary forces across nonlinear, anisotropic, and coupled thermo-mechanical problems

desk verdict Useful AE-MOR extensions, but the 'across all cases' claim overreaches; needs stronger validation. read the letter →

arxiv 2509.02237 v1 pith:3BWHYYC6 submitted 2025-09-02 cs.CE cs.AIcs.LG

Jannick Kehls , Ellen Kuhl , Tim Brepols , Kevin Linka , Hagen Holthusen This is my paper · ORCID
classification cs.CEcs.AIcs.LG
keywords autoencodermodelorderreductionnon-intrusivesurrogatefiniteelementmethodlatentspaceregressionreactionforcepredictionmulti-fieldthermo-mechanicsbifurcation
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 claims that a three-stage autoencoder pipeline can act as a fully non-intrusive surrogate for nonlinear continuum mechanics: first compress finite element solution snapshots into a low-dimensional latent space, then learn a mapping from problem parameters to latent codes, and finally decode directly to full-field predictions. It extends this base architecture in two ways: a force-augmented variant that learns a shared latent space for displacements and reaction forces at Dirichlet boundaries, and a multi-field variant with separate encoders per physical field feeding a joint decoder. The framework is validated on a heterogeneous unit cell, a fiber-reinforced plate with an elliptic hole, and a transient thermo-mechanical plate, reproducing high-fidelity displacement, force, and temperature fields for unseen parameters. The central promise is that accurate surrogate modeling can be built purely from simulation data, with no need to modify the finite element solver or inspect the governing equations.

What carries the argument

The central mechanism is the latent bottleneck: an unsupervised encoder-decoder compresses high-dimensional finite element snapshots into a few latent coordinates, then a feedforward regression network maps problem parameters to those coordinates, and the frozen decoder turns them back into full fields. Two extensions carry the paper's new claims: in the force-augmented variant the latent variable is the sum of separate displacement and force encoder outputs, forcing one shared code to serve both fields; in the multi-field variant each physical field gets its own encoder and the codes are concatenated before a shared decoder. Variance-normalized losses prevent large-magnitude fields from dom

What would settle it

Train the same multi-field architecture on the thermo-mechanical plate with no geometric imperfection, then evaluate at parameter points where the finite element reference produces two stable buckled shapes, one in the positive and one in the negative z-direction. If the surrogate frequently returns a mixed or averaged shape while the reference lands cleanly on one branch, the single-valued-manifold assumption is the cause; the paper's Figure 20 already shows qualitative deterioration, so the sharper test is to count branch-consistency across many such points.

Watch

Extended reading notes

Core claim

The central claim is that a nonlinear autoencoder can capture curved solution manifolds in continuum mechanics where a fixed linear subspace, such as POD, would struggle, and that the learned latent space can be shared or decomposed to cover additional quantities of interest. Concretely, the end-to-end surrogate is the composition of a regression network and a frozen decoder; the force-augmented variant replaces the single latent bottleneck with the sum of displacement and force encoder outputs and trains with variance-normalized losses; and the multi-field variant concatenates field-specific latent codes before a shared decoder. On the paper's benchmarks, this produces accurate reconstructi

Load-bearing premise

The surrogate treats the solution as a single-valued function of the input parameters, so when a physical system can take several valid answers for the same input (like which way a plate buckles), the training data mixes branches and the model averages them into an inaccurate prediction.

Editorial extensions

If this is right

  • For a new parameter value, the surrogate answers with one decoder evaluation, replacing a full nonlinear finite element solve in the many-query setting.
  • Neumann boundary reactions are available from the same end-to-end model rather than as a separate post-processing step, and including them during training also sharpens the displacement prediction.
  • Multi-physics surrogates can be assembled by adding an encoder per field and concatenating latent codes, so the architecture grows with the physics instead of being rebuilt from scratch.
  • The accuracy is conditional on the solution being a single-valued function of the input parameters: when multiple solution branches coexist, as in the unimperfected buckling plate, predictions degrade substantially.

Reading between the lines

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

  • Because the pipeline treats the solver as a black box, the same latent-space recipe should transfer to non-standard outputs such as derived stress measures or localized damage fields, provided snapshot-parameter pairs can be generated; the paper does not demonstrate this.
  • The force-augmented shared latent space creates a natural physical-consistency probe: reaction forces from the decoder could be checked against internal equilibrium of the predicted displacement field, and enforcing that balance during training would address the residual inconsistency the paper lists as a limitation.
  • The bifurcation failure suggests a concrete preprocessing fix that the paper mentions only in passing: reflecting buckled snapshots to a common side before training would remove the mixed-branch ambiguity and might restore accuracy, which is directly testable on the same thermo-mechanical example.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes a three-stage, non-intrusive model order reduction framework for continuum mechanics: an unsupervised autoencoder compresses high-dimensional finite element snapshots into a low-dimensional latent space, a supervised regression network maps input parameters to latent codes, and an end-to-end surrogate reconstructs full-field solutions. Two extensions are introduced: a force-augmented variant with a shared latent space for displacements and reaction forces, and a multi-field architecture with field-specific encoders and a joint decoder. The method is validated on three benchmarks: a heterogeneous unit cell, a fiber-reinforced plate with an elliptical hole, and a transient thermo-mechanical plate. The authors claim accurate reconstructions across all cases and provide a publicly available implementation.

Significance. If the central claim were fully supported, the paper would make a useful contribution by extending non-intrusive autoencoder-based MOR to reaction-force prediction and coupled multi-field problems, areas that existing AE-based methods rarely address. The paper benefits from external high-fidelity FE references (so the benchmark is not circular), a clearly described architecture, and a public implementation. However, the validation is thin: each benchmark is assessed at a single in-domain parameter point, there are no error bars or test-set statistics, no comparison against a POD-based non-intrusive baseline, and the bifurcation failure in Section 3.3 directly contradicts the unqualified abstract claim. The significance is therefore contingent on a revision that narrows the claims and adds quantitative evidence.

major comments (3)
  1. [§3.3, Fig. 20; Abstract] The abstract's unqualified claim "Across all cases, it achieves accurate reconstructions of high-fidelity solutions" is contradicted by the paper's own no-imperfection thermo-mechanical experiment. Equations (5)-(6) assume a single-valued parameter-to-latent map; for the geometrically perfect plate, identical parameters produce two buckling branches (+z and -z). The MSE regression in Eq. (6) therefore fits a mean of incompatible branches. Figure 20 shows "a significant deterioration in the predictive accuracy" and a non-smooth predicted surface. The authors' remedy—introducing a 0.01 mm imperfection—changes the problem, and the training set omits (ξ,η)=(2/3,0) and (1,0) because of non-convergent FE solves. The Abstract and Section 4 should restrict the claim to problems with a unique solution or explicit symmetry breaking.
  2. [§3.1–§3.3 validation protocol] Each benchmark is evaluated at a single in-domain parameter point: (ξ, η) = (0.55, 0.35) for the unit cell, (0.75, 0.3) for the plate, and (ξ, η, t) = (0.5, 0.5, 0.8) for the thermo-mechanical problem. No test-set statistics, no repeated-training variability, and no error bars are reported. The introduction argues that POD's linear subspace is inadequate, but no POD-based non-intrusive baseline (e.g., POD plus regression) is quantitatively compared. The central claim of accurate reconstruction "across all cases" is not quantitatively established; a held-out parameter grid with relative L2 error statistics and one POD+regression baseline would be needed.
  3. [§3.1.2, Eq. (11)] The force-augmented variant's claimed benefit is supported only by qualitative force contours and a displacement-error plot (Fig. 10). No quantitative error metric for the predicted force vector f is reported, despite the variance-normalized MSE in Eq. (11) being explicitly designed to balance displacement and force losses. The superiority of summing encoder outputs into a shared latent (Eq. 8) over a simple concatenation is not tested. To support the force-prediction claim, the paper should report force errors (e.g., relative L2 error over the boundary) for both the force-augmented and force-reconstructed variants.
minor comments (6)
  1. [§3.2.2 caption of Fig. 15] The caption says "the displacements were then predicted by the end-to-end model of Section 3.1.2"; this appears to be a typo for Section 3.2.2.
  2. [Eqs. (12)–(15)] The notation for field-specific encoders and regression networks (1E, 2P) is visually awkward; superscripts or a different indexing scheme would improve readability.
  3. [§3.2] Typo: "hyperlelastic" should be "hyperelastic".
  4. [§2, End-to-End Surrogate Model] The sentence "To which is extent the prediction matches unseen states..." is ungrammatical and should be rewritten.
  5. [Figure 9] The labels "Prediction Force-augmentedForce-reconstructed" are run together; the figure would be clearer with separated labels.
  6. [§3.3] The missing training points (ξ,η)=(2/3,0) and (1,0) are mentioned only in passing; this should be noted as a data-coverage limitation in Section 4 as well.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: external FE validation anchors every claim; self-citations are motivational, not load-bearing.

full rationale

The paper's central predictions (displacement, reaction forces, temperature) are evaluated against full-order finite element solutions at held-out parameter points (e.g., ξ=0.55, η=0.35 in Sec. 3.1; ξ=0.75, η=0.3 in Sec. 3.2; ξ=η=0.5, t=0.8 in Sec. 3.3), so the benchmarks are not self-referential. The force-augmented model uses force labels to define the shared latent target (Eq. 8) and trains the regression on those latent codes; this is standard supervised surrogate construction, not a fitted input renamed as a prediction. The comparison with the staggered force-reconstructed model (A.2) explicitly shows that a force decoder trained on displacement-only latent codes does not reproduce the force-augmented result, ruling out a 'prediction forced by construction' pattern. The multi-field architecture is motivated by the authors' prior intrusive MOR papers [44,45], but the present numerical validation is independent of those papers; no uniqueness theorem or ansatz is imported. The acknowledged bifurcation failure (Sec. 3.3, Fig. 20) is a correctness and scope limitation of the single-valued latent-map assumption, not a circular derivation: the paper openly reports the deterioration rather than claiming the no-imperfection case succeeds. Overall, the derivation chain is self-contained and externally benchmarked; the self-citations are motivational, not load-bearing.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical entities, forces, or dimensions. The shared latent space and separate encoder-decoder pairs are architectural constructs, not invented physical objects. The main free parameters are the latent dimensions and training hyperparameters, all chosen heuristically without sensitivity studies.

free parameters (7)
  • Latent dimension (unit cell) = 4
    Hand-picked; compresses 18,580 active DOFs. No ablation study reported.
  • Latent dimension (mesh morphing) = 2
    Hand-picked for reference node positions; no sensitivity analysis.
  • Latent dimension (plate displacement) = 8
    Chosen because more information was anticipated; no ablation.
  • Latent dimensions (thermo-mechanics) = 4 per field (8 concatenated)
    Fixed for both encoders; no justification beyond simplicity.
  • Learning rate = 1e-3
    Uniform across all networks; heuristic (Section 3).
  • Elastic net penalty factor = 1e-7 (L1 and L2)
    Uniform regularization; heuristic, no hyperparameter study (Section 3).
  • Encoder/decoder layer sizes = e.g., 18580-1024-512-128-32-16-4
    Chosen heuristically; no architecture search.
assumptions (5)
  • domain assumption The FE solution manifold is approximately low-dimensional, so 2-8 latent neurons suffice.
    Section 2 describes the autoencoder as learning the 'core dynamics'; the adequacy is checked only by reconstruction loss, not a formal bound.
  • domain assumption The solution field is a single-valued function of the input parameters.
    Violated in the buckling example without imperfection (Section 3.3): training snapshots mix positive and negative buckling modes, and predictions deteriorate.
  • ad hoc to paper Summing the two encoder outputs yields a unified shared latent that preserves the complementary displacement and force information (Eq. 8).
    This is a new architectural choice with no theoretical justification; it is motivated by the physical intuition that both fields share the same governing laws.
  • domain assumption Elastic mesh morphing with a pseudo-isotropic material (µ=κ=1, K1=K2=0) produces valid nodal reference positions.
    Section 3.2.1; the morphed meshes are used as ground truth for the morphing autoencoder, but the material parameter choice is arbitrary and could affect the learned geometry.
  • standard math Neural networks with GELU/SiLU activations and MSE losses can represent the parameter-to-latent mapping accurately.
    Relies on standard approximation properties of MLPs; no error analysis is given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Autoencoder-based non-intrusive model order reduction in continuum mechanics." pith.science (2026). https://pith.science/paper/3BWHYYC6

@misc{pith2026250902237,
  author       = {Pith},
  title        = {Pith review of: Autoencoder-based non-intrusive model order reduction in continuum mechanics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3BWHYYC6}},
  note         = {Machine review of arXiv:2509.02237}
}
read the original abstract

We propose a non-intrusive, Autoencoder-based framework for reduced-order modeling in continuum mechanics. Our method integrates three stages: (i) an unsupervised Autoencoder compresses high-dimensional finite element solutions into a compact latent space, (ii) a supervised regression network maps problem parameters to latent codes, and (iii) an end-to-end surrogate reconstructs full-field solutions directly from input parameters. To overcome limitations of existing approaches, we propose two key extensions: a force-augmented variant that jointly predicts displacement fields and reaction forces at Neumann boundaries, and a multi-field architecture that enables coupled field predictions, such as in thermo-mechanical systems. The framework is validated on nonlinear benchmark problems involving heterogeneous composites, anisotropic elasticity with geometric variation, and thermo-mechanical coupling. Across all cases, it achieves accurate reconstructions of high-fidelity solutions while remaining fully non-intrusive. These results highlight the potential of combining deep learning with dimensionality reduction to build efficient and extensible surrogate models. Our publicly available implementation provides a foundation for integrating data-driven model order reduction into uncertainty quantification, optimization, and digital twin applications.

Figures

Figures reproduced from arXiv: 2509.02237 by the authors.

Figure 1
Figure 1. Unsupervised Autoencoder Discovery. Unlabeled snapshots (•) s from the input space ϕ are encoded into a latent representation and subsequently reconstructed by the decoder. The encoder’s hidden layers are shown in blue, the latent space neurons in green, and the decoder’s hidden layers in orange. The architectures of the encoder and decoder are not necessarily the same. extension are presented. For interested reader… view at source ↗
Figure 2
Figure 2. Supervised Latent Space Prediction. The regression network (yellow) is trained to infer the mapping from input parameters θ to the solution field ϕ via a latent representation. To this end, the encoder (dashed, frozen during training) provides the latent space for each snapshot. generally approximates the latent space representation zˆ = P(θ; wP ), (5) with the network’s weights denoted by wP . For the time being, t… view at source ↗
Figure 3
Figure 3. End-to-End Surrogate Model. The latent representation generated by the encoder is replaced by the output of the trained regression network, enabling a direct and efficient pipeline for predicting the solution field from input parameters. To this end, we augment the latent space of the previous model by incorporating the force information at these non-active degrees. Recognizing the distinct structural characteristic… view at source ↗
Figures from the paper (20 more)
Figure 4
Figure 4. Figure 4: For comparison, we also investigated a simplified variant predicting force vectors from the solution latent space alone (cf. A.2). However, as shown in Section 3, this approach failed to achieve satisfactory accuracy. Multi-field extension. Thus far, we have outlined a…
Figure 4
Figure 4. Figure 4: End-to-End Force-augmented Model. Top: The solution space ϕ and the corresponding force terms f are provided to individual encoders. Subsequently, the encoders learn a shared latent space. Thereafter, individual decoder networks reconstruct the solution field and force…
Figure 5
Figure 5. Figure 5: Multi-field extension. Top: The input space ϕ is provided to individual encoders in a semi-labeled fashion. Subsequently, each encoder learns its own latent space representation, which are then concatenated to an unified latent representation. A fully-connected decoder…
Figure 6
Figure 6. Figure 6: Sketch of the three-dimensional unit cell with inclusions. The dimensions of the unit cell are [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Unit cell. Training loss over the course of training for both neural networks, namely the Autoencoder and the Latent Space Prediction network, each trained for a maximum of 5000 epochs. Left: The end-to-end model is trained solely on solution field data, i.e., the disp…
Figure 8
Figure 8. Figure 8: Gradient of the y-component of the solution field ϕ in y-direction, computed using ParaView’s Gradient filter [49]. The natural coordinates are evaluated at ξ = 0.55 and η = 0.35. On the left, the reference solution obtained from the full-order model is shown; the righ…
Figure 9
Figure 9. Figure 9: Force components at the Dirichlet boundary in the [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]
Figure 10
Figure 10. Figure 10: Prediction error between the full-order reference solution and the non-intrusive prediction for networks [PITH_FULL_IMAGE:figures/full_fig_p017_10.png]
Figure 11
Figure 11. Figure 11: Sketch of the three-dimensional plate with an elliptic hole. The plate dimensions are [PITH_FULL_IMAGE:figures/full_fig_p018_11.png]
Figure 12
Figure 12. Figure 12: Snapshots illustrating the mesh morphing of the inner elliptical hole. The black lines represent the original [PITH_FULL_IMAGE:figures/full_fig_p019_12.png]
Figure 13
Figure 13. Figure 13: Plate with elliptic hole. Training loss over the course of training for both neural networks, namely the Autoencoder and the Latent Space Prediction network, each trained for a maximum of 5000 epochs. Left: The end-to-end model is trained to predict the reference node…
Figure 14
Figure 14. Figure 14: Reference node positions resulting from elastic mesh morphing. The green lines indicate the nodal positions [PITH_FULL_IMAGE:figures/full_fig_p020_14.png]
Figure 15
Figure 15. Figure 15: Deformed mesh of the plate with an elliptic hole. Green lines illustrate the high-fidelity reference solution, [PITH_FULL_IMAGE:figures/full_fig_p022_15.png]
Figure 16
Figure 16. Figure 16: Sketch of the three-dimensional plate with an elliptic hole. The plate dimensions are [PITH_FULL_IMAGE:figures/full_fig_p023_16.png]
Figure 20
Figure 20. Figure 20: The original purpose of introducing an inclined imperfection was to avoid the bifurcation issue, where it is [PITH_FULL_IMAGE:figures/full_fig_p024_20.png]
Figure 17
Figure 17. Figure 17: Thermo-mechanics. Training loss during the minimization procedure for all three networks: the Autoen￾coder, the latent space predictor for the displacement field u, and the latent space predictor for the temperature field T. In all training sessions, the number of epo…
Figure 18
Figure 18. Figure 18: Temperature field across the specimen evaluated at [PITH_FULL_IMAGE:figures/full_fig_p025_18.png]
Figure 19
Figure 19. Figure 19: Comparison between the predicted displacement fields and the high-fidelity reference solution obtained [PITH_FULL_IMAGE:figures/full_fig_p026_19.png]
Figure 20
Figure 20. Figure 20: Computation and prediction of the deformation of a plate with an elliptic hole subjected to thermal loading, [PITH_FULL_IMAGE:figures/full_fig_p027_20.png]
Figure 21
Figure 21. Figure 21: Staggered Surrogate Force Model. The decoder network for the force terms is trained on latent representa￾tions generated by the regression network, which itself was trained solely on solution field data. Hence, the regression network has never been exposed to force in…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

54 extracted references · 51 canonical work pages

  1. [1]

    An introduction to the proper orthogonal decomposition

    Anindya Chatterjee. An introduction to the proper orthogonal decomposition. Current Science, 78(7):808–817, 2000

    work page 2000

  2. [2]

    Proper orthogonal decomposition surrogate models for nonlinear dynamical systems: Error estimates and suboptimal control

    Michael Hinze and Stefan V olkwein. Proper orthogonal decomposition surrogate models for nonlinear dynamical systems: Error estimates and suboptimal control. In Peter Benner, Danny C. Sorensen, and V olker Mehrmann, editors, Dimension Reduction of Large-Scale Systems, pages 261–306, Berlin, Heidelberg, 2005. Springer. 31 Autoencoder-based non-intrusive mo...

    work page 2005

  3. [3]

    Vakakis, and Lawrence A

    Gaetan Kerschen, Jean-claude Golinval, Alexander F. Vakakis, and Lawrence A. Bergman. The method of proper orthogonal decomposition for dynamical characterization and order reduction of mechanical systems: An overview. Nonlinear Dynamics, 41(1):147–169, 2005

    work page 2005

  4. [4]

    The applications of pod method in dual rotor-bearing systems with coupling misalignment.Mechanical Systems and Signal Processing, 150:107236, 2021

    Kuan Lu, Yulin Jin, Panfeng Huang, Fan Zhang, Haopeng Zhang, Chao Fu, and Yushu Chen. The applications of pod method in dual rotor-bearing systems with coupling misalignment.Mechanical Systems and Signal Processing, 150:107236, 2021

    work page 2021

  5. [5]

    Berkooz, P

    G. Berkooz, P. Holmes, and J. L. Lumley. The proper orthogonal decomposition in the analysis of turbulent flows. Annual Review of Fluid Mechanics, 25:539–575, 1993

    work page 1993

  6. [6]

    Smith, Jeff Moehlis, and Philip Holmes

    Troy R. Smith, Jeff Moehlis, and Philip Holmes. Low-dimensional modelling of turbulence using the proper orthogonal decomposition: A tutorial. Nonlinear Dynamics, 41(1):275–307, 2005

    work page 2005

  7. [7]

    Data-driven pod-galerkin reduced order model for turbulent flows

    Saddam Hijazi, Giovanni Stabile, Andrea Mola, and Gianluigi Rozza. Data-driven pod-galerkin reduced order model for turbulent flows. Journal of Computational Physics, 416:109513, 2020

    work page 2020

  8. [8]

    Proper orthogonal decomposition-based modeling, analysis, and simulation of dynamic wind load effects on structures

    Xinzhong Chen and Ahsan Kareem. Proper orthogonal decomposition-based modeling, analysis, and simulation of dynamic wind load effects on structures. Journal of Engineering Mechanics, 131(4):325–339, 2005

    work page 2005

Show all 54 references
  1. [9]

    A comparison of projection-based model reduction concepts in the context of nonlinear biomechanics

    Annika Radermacher and Stefanie Reese. A comparison of projection-based model reduction concepts in the context of nonlinear biomechanics. Archive of Applied Mechanics, 83(8):1193–1213, 2013

    2013

  2. [10]

    An adaptive model order reduction technique for parameter-dependent modular structures

    Stephan Ritzert, Domen Macek, Jaan-Willem Simon, and Stefanie Reese. An adaptive model order reduction technique for parameter-dependent modular structures. Computational Mechanics, 73(5):1147–1163, 2024

    2024

  3. [11]

    Component-based model-order reduction with mortar tied contact for nonlinear quasi-static mechanical problems

    Stephan Ritzert, Jannick Kehls, Stefanie Reese, and Tim Brepols. Component-based model-order reduction with mortar tied contact for nonlinear quasi-static mechanical problems. International Journal for Numerical Methods in Engineering, 126(8):e70041, 2025

    2025

  4. [12]

    Ryckelynck

    D. Ryckelynck. A priori hyperreduction method: an adaptive approach. Journal of Computational Physics , 202(1):346–366, 2005

    2005

  5. [13]

    Ryckelynck

    D. Ryckelynck. Hyper-reduction of mechanical models involving internal variables. International Journal for Numerical Methods in Engineering, 77(1):75–89, 2009

    2009

  6. [14]

    Miled, D

    B. Miled, D. Ryckelynck, and S. Cantournet. A priori hyper-reduction method for coupled viscoelastic–viscoplastic composites. Computers & Structures, 119:95–103, 2013

    2013

  7. [15]

    Sorensen

    Saifon Chaturantabut and Danny C. Sorensen. Nonlinear model reduction via discrete empirical interpolation. SIAM Journal on Scientific Computing, 32(5):2737–2764, 2010

    2010

  8. [16]

    Pod-based model reduction with empirical interpolation applied to nonlinear elasticity

    Annika Radermacher and Stefanie Reese. Pod-based model reduction with empirical interpolation applied to nonlinear elasticity. International Journal for Numerical Methods in Engineering, 107(6):477–495, 2016

    2016

  9. [17]

    An, Theodore Kim, and Doug L

    Steven S. An, Theodore Kim, and Doug L. James. Optimizing cubature for efficient integration of subspace deformations. ACM Transactions on Graphics, 27(5), 2008

    2008

  10. [18]

    Dimensional reduction of nonlinear finite element dynamic models with finite rotations and energy-based mesh sampling and weighting for computational efficiency

    Charbel Farhat, Philip Avery, Todd Chapman, and Julien Cortial. Dimensional reduction of nonlinear finite element dynamic models with finite rotations and energy-based mesh sampling and weighting for computational efficiency. International Journal for Numerical Methods in Engi...

    2014

  11. [19]

    Structure-preserving, stability, and accuracy properties of the energy-conserving sampling and weighting method for the hyper reduction of nonlinear finite element dynamic models

    Charbel Farhat, Todd Chapman, and Philip Avery. Structure-preserving, stability, and accuracy properties of the energy-conserving sampling and weighting method for the hyper reduction of nonlinear finite element dynamic models. International Journal for Numerical Methods in En...

    2015

  12. [20]

    Francesco Trainotti, Jure Marinko, Johannes Maierhofer, and Daniel J. Rixen. Ecsw hyperreduction of hyper- viscoelastic components via co-simulation with abaqus. Finite Elements in Analysis and Design, 241:104222, 2024. 32 Autoencoder-based non-intrusive model order reduction ...

    2024

  13. [21]

    J. A. Hernández, M. A. Caicedo, and A. Ferrer. Dimensional hyper-reduction of nonlinear finite element models via empirical cubature. Computer Methods in Applied Mechanics and Engineering, 313:687–722, 2017

    2017

  14. [22]

    Bravo, S

    J.R. Bravo, S. Ares de Parga, J.A. Hernández, and R. Rossi. Local-ecm: An empirical cubature hyper-reduction method adapted to local reduced order models, 2023

    2023

  15. [23]

    Hernández, J.R

    J.A. Hernández, J.R. Bravo, and S. Ares de Parga. Cecm: A continuous empirical cubature method with application to the dimensional hyperreduction of parameterized finite element models. Computer Methods in Applied Mechanics and Engineering, 418:116552, 2024

    2024

  16. [24]

    D. Xiao, F. Fang, C. Pain, and G. Hu. Non-intrusive reduced-order modelling of the navier–stokes equations based on rbf interpolation. International Journal for Numerical Methods in Fluids, 79(11):580–595, 2015

    2015

  17. [25]

    J. S. Hesthaven and S. Ubbiali. Non-intrusive reduced order modeling of nonlinear problems using neural networks. Journal of Computational Physics, 363:55–78, 2018

    2018

  18. [26]

    Hesthaven, and Deep Ray

    Qian Wang, Jan S. Hesthaven, and Deep Ray. Non-intrusive reduced order modeling of unsteady flows using artificial neural networks with application to a combustion problem. Journal of Computational Physics, 384:289– 307, 2019

    2019

  19. [27]

    Projection-based model reduction: Formulations for physics-based machine learning

    Renee Swischuk, Laura Mainini, Benjamin Peherstorfer, and Karen Willcox. Projection-based model reduction: Formulations for physics-based machine learning. Computers & Fluids, 179:704–717, 2019

    2019

  20. [28]

    A nonintrusive nonlinear model reduction method for structural dynamical problems based on machine learning

    Jonas Kneifl, Dennis Grunert, and Joerg Fehr. A nonintrusive nonlinear model reduction method for structural dynamical problems based on machine learning. International Journal for Numerical Methods in Engineering, 122(17):4774–4786, 2021

    2021

  21. [29]

    Salvador, L

    M. Salvador, L. Dedè, and A. Manzoni. Non intrusive reduced order modeling of parametrized PDEs by kernel POD and neural networks. Computers & Mathematics with Applications, 104:1–13, 2021

    2021

  22. [30]

    Testing the manifold hypothesis

    Charles Fefferman, Sanjoy Mitter, and Hariharan Narayanan. Testing the manifold hypothesis. Journal of the American Mathematical Society, 29(4):983–1049, 2016

    2016

  23. [31]

    Bourlard and Y

    H. Bourlard and Y . Kamp. Auto-association by multilayer perceptrons and singular value decomposition. Biological Cybernetics, 59(4):291–294, 1988

    1988

  24. [32]

    Neural networks and principal component analysis: Learning from examples without local minima

    Pierre Baldi and Kurt Hornik. Neural networks and principal component analysis: Learning from examples without local minima. Neural Networks, 2(1):53–58, 1989

    1989

  25. [33]

    From principal subspaces to principal components with linear autoencoders, 2018

    Elad Plaut. From principal subspaces to principal components with linear autoencoders, 2018

    2018

  26. [34]

    Gonzalez and Maciej Balajewicz

    Francisco J. Gonzalez and Maciej Balajewicz. Deep convolutional recurrent autoencoders for learning low- dimensional feature dynamics of fluid systems, 2018

    2018

  27. [35]

    Deep learning of thermodynamics-aware reduced-order models from data

    Quercus Hernandez, Alberto Badías, David González, Francisco Chinesta, and Elías Cueto. Deep learning of thermodynamics-aware reduced-order models from data. Computer Methods in Applied Mechanics and Engineering, 379:113763, 2021

    2021

  28. [36]

    A comprehensive deep learning-based approach to reduced order modeling of nonlinear time-dependent parametrized PDEs

    Stefania Fresca, Luca Dedè, and Andrea Manzoni. A comprehensive deep learning-based approach to reduced order modeling of nonlinear time-dependent parametrized PDEs. Journal of Scientific Computing, 87(2):61, Apr 2021

    2021

  29. [37]

    Machine learning approach to model order reduction of nonlinear systems via autoencoder and lstm networks

    Thomas Simpson, Nikolaos Dervilis, and Eleni Chatzi. Machine learning approach to model order reduction of nonlinear systems via autoencoder and lstm networks. Journal of Engineering Mechanics, 147(10):04021061, 2021

    2021

  30. [38]

    Dimensionality reduction through convolutional autoencoders for fracture patterns prediction

    Krushna Shinde, Vincent Itier, José Mennesson, Dmytro Vasiukov, and Modesar Shakoor. Dimensionality reduction through convolutional autoencoders for fracture patterns prediction. Applied Mathematical Modelling, 114:94–113, 2023. 33 Autoencoder-based non-intrusive model order r...

    2023

  31. [39]

    Hesthaven

    Federico Pichi, Beatriz Moya, and Jan S. Hesthaven. A graph convolutional autoencoder approach to model order reduction for parametrized pdes. Journal of Computational Physics, 501:112762, 2024

    2024

  32. [40]

    Fidkowski, and Kevin J

    Rakesh Halder, Krzysztof J. Fidkowski, and Kevin J. Maki. Non-intrusive reduced-order modeling using convolutional autoencoders. International Journal for Numerical Methods in Engineering, 123(21):5369–5390, 2022

    2022

  33. [41]

    Low-dimensional data-based surrogate model of a continuum-mechanical musculoskeletal system based on non-intrusive model order reduction

    Jonas Kneifl, David Rosin, Okan Avci, Oliver Röhrle, and Jörg Fehr. Low-dimensional data-based surrogate model of a continuum-mechanical musculoskeletal system based on non-intrusive model order reduction. Archive of Applied Mechanics, 93(9):3637–3663, Sep 2023

    2023

  34. [42]

    Bordas, and Jakub Lengiewicz

    Saurabh Deshpande, Hussein Rappel, Mark Hobbs, Stéphane P.A. Bordas, and Jakub Lengiewicz. Gaussian process regression + deep neural network autoencoder for probabilistic surrogate modeling in nonlinear mechanics of solids. Computer Methods in Applied Mechanics and Engineering...

    2025

  35. [43]

    A length scale insensitive phase-field damage model for brittle fracture

    Jian-Ying Wu and Vinh Phu Nguyen. A length scale insensitive phase-field damage model for brittle fracture. Journal of the Mechanics and Physics of Solids, 119:20–42, 2018

    2018

  36. [44]

    A multi- field decomposed model order reduction approach for thermo-mechanically coupled gradient-extended damage simulations

    Qinghua Zhang, Stephan Ritzert, Jian Zhang, Jannick Kehls, Stefanie Reese, and Tim Brepols. A multi- field decomposed model order reduction approach for thermo-mechanically coupled gradient-extended damage simulations. Computer Methods in Applied Mechanics and Engineering, 434...

    2025

  37. [45]

    Multi-field decomposed hyper-reduced order modeling of damage-plasticity simulations, 2025

    Jannick Kehls, Stephan Ritzert, Lars Breuer, Qinghua Zhang, Stefanie Reese, and Tim Brepols. Multi-field decomposed hyper-reduced order modeling of damage-plasticity simulations, 2025

    2025

  38. [46]

    Polyconvexity of generalized polynomial-type hyperelastic strain energy functions for near-incompressibility

    Stefan Hartmann and Patrizio Neff. Polyconvexity of generalized polynomial-type hyperelastic strain energy functions for near-incompressibility. International Journal of Solids and Structures, 40(11):2767–2791, 2003

    2003

  39. [47]

    A. N. Gent. A new constitutive relation for rubber. Rubber Chemistry and Technology, 69(1):59–61, 03 1996

    1996

  40. [48]

    Gaussian error linear units (gelus), 2023

    Dan Hendrycks and Kevin Gimpel. Gaussian error linear units (gelus), 2023

    2023

  41. [49]

    ParaView: An end-user tool for large data visualization

    James Ahrens, Berk Geveci, and Charles Law. ParaView: An end-user tool for large data visualization. In Visualization Handbook. Elesvier, 2005. ISBN 978-0123875822

    2005

  42. [50]

    Invariant formulation of hyperelastic transverse isotropy based on polyconvex free energy functions

    Jörg Schröder and Patrizio Neff. Invariant formulation of hyperelastic transverse isotropy based on polyconvex free energy functions. International Journal of Solids and Structures, 40(2):401–445, 2003

    2003

  43. [51]

    On the stability of POD basis interpolation on grassmann manifolds for parametric model order reduction

    Orestis Friderikos, Emmanuel Baranger, Marc Olive, and David Neron. On the stability of POD basis interpolation on grassmann manifolds for parametric model order reduction. Computational Mechanics, 70(1):181–204, Jul 2022

    2022

  44. [52]

    On finite thermal deformations

    R Stojanovic, S Djuric, and L Vujosevic. On finite thermal deformations. Archiwum Mechaniki Stosowanej, 16(103-108):51, 1964

    1964

  45. [53]

    Finite-strain thermoelasticity based on multiplicative decomposition of deformation gradient

    L Vujoševi´c and V A Lubarda. Finite-strain thermoelasticity based on multiplicative decomposition of deformation gradient. Theoretical and applied mechanics, (28-29):379–399, 2002

    2002

  46. [54]

    Felder, N

    S. Felder, N. Kopic-Osmanovic, H. Holthusen, T. Brepols, and S. Reese. Thermo-mechanically coupled gradient- extended damage-plasticity modeling of metallic materials at finite strains. International Journal of Plasticity, 148:103142, 2022. 34

    2022

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.