pith. machine review for the scientific record. sign in

arxiv: 2604.05698 · v1 · submitted 2026-04-07 · 💻 cs.CE

Recognition: 2 theorem links

· Lean Theorem

Adaptive Material Fingerprinting for the fast discovery of polyconvex feature combinations in isotropic and anisotropic hyperelasticity

Moritz Flaschel , Hagen Holthusen , Denisa Martonov\'a , Ellen Kuhl
Authors on Pith no claims yet

Pith reviewed 2026-05-10 18:32 UTC · model grok-4.3

classification 💻 cs.CE
keywords material fingerprintinghyperelasticitypolyconvexitymodel discoveryOgden modelanisotropic invariantsstrain energy densityiterative refinement
0
0 comments X

The pith

Adaptive fingerprinting discovers complex hyperelastic models as linear feature combinations

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

The paper introduces an adaptive extension to Material Fingerprinting that builds material models as sums of strain energy features drawn from a database. Starting from basic isotropic and anisotropic invariants, an iterative algorithm refines which features to include and with what coefficients until the combined fingerprint matches experimental data. This removes the restriction to predefined single models and permits discovery of complex forms such as multi-term Ogden models or the Holzapfel-Gasser-Ogden model for anisotropic tissues. The features are constructed to obey physical constraints, with an optional switch to enforce polyconvexity. The approach is demonstrated on mechanical test data from isotropic rubber and anisotropic animal skin.

Core claim

We propose an adaptive model database coupled with an iterative pattern recognition algorithm that refines the material model in each step. This strategy enables Material Fingerprinting to discover arbitrary linear combinations of material models from the database, rather than being restricted to selecting a single model from a predefined set. In comparison to previous works on Material Fingerprinting, this enables the discovery of more complex models, such as multi-term Ogden models or the anisotropic Holzapfel-Gasser-Ogden model. To design the adaptive database, we leverage sums of strain energy density feature functions that depend on isotropic and anisotropic invariants. All modeling of

What carries the argument

An adaptive database of sums of strain energy density feature functions based on isotropic and anisotropic invariants, refined by iterative pattern recognition on mechanical fingerprints

If this is right

  • Enables discovery of arbitrary linear combinations rather than single models from the database
  • Allows identification of complex models such as multi-term Ogden and anisotropic Holzapfel-Gasser-Ogden
  • Maintains physical constraints with optional enforcement of polyconvexity
  • Supports real-time discovery for experimental measurements of rubber and tissue
  • Avoids continuous optimization by using precomputed fingerprints

Where Pith is reading between the lines

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

  • The iterative approach may allow systematic exploration of admissible model spaces for hyperelastic materials
  • Application to other material classes could be possible if appropriate fingerprint experiments are defined
  • Potential non-uniqueness in feature combinations for similar materials may require additional validation steps

Load-bearing premise

That each material produces a sufficiently unique fingerprint under the standardized experimental setup and iterative refinement converges without introducing non-uniqueness or instability

What would settle it

Applying the method to synthetic data generated from a known multi-term model and observing recovery of an incorrect or non-unique feature combination would show the approach does not reliably identify the true model

Figures

Figures reproduced from arXiv: 2604.05698 by Denisa Martonov\'a, Ellen Kuhl, Hagen Holthusen, Moritz Flaschel.

Figure 1
Figure 1. Figure 1: Illustration of adaptive Material Fingerprinting. In the o [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Stretch-controlled homogeneous deformation modes used for adaptive material fingerprinting under incompressibility condition. All [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Functions considered for the construction of isotropic strain energy density features. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Functions considered for the construction of anisotropic strain energy density features. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Hyperparameter analysis for adaptive Material Fingerprinting applied to rubber data. [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Stress-stretch data and discovered model for rubber at 20 [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Stress-stretch data and discovered model for rubber at 20 [PITH_FULL_IMAGE:figures/full_fig_p010_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Stress-stretch data and discovered model for rubber at 50 [PITH_FULL_IMAGE:figures/full_fig_p011_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Hyperparameter analysis for adaptive Material Fingerprinting applied to skin data. [PITH_FULL_IMAGE:figures/full_fig_p012_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Stress-stretch data and discovered model for skin with hyperparameters [PITH_FULL_IMAGE:figures/full_fig_p012_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Stress-stretch data and discovered model for skin with hyperparameters [PITH_FULL_IMAGE:figures/full_fig_p013_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Stress-stretch data and discovered model for skin with hyperparameters [PITH_FULL_IMAGE:figures/full_fig_p013_12.png] view at source ↗
read the original abstract

We recently proposed a method called Material Fingerprinting for the rapid discovery of mechanical material models that avoids solving continuous optimization problems. Material Fingerprinting assumes that each material exhibits a unique response when subjected to a standardized experimental setup, which is interpreted as the material's mechanical fingerprint. If a database of fingerprints is generated in an offline phase, a model for an unseen experimental measurement can be discovered in real time by comparing the experimentally measured fingerprint to the fingerprints in the database. In our original contributions, the database comprised a fixed number of material models, each with a fixed number of parameters. To increase the fitting flexibility of Material Fingerprinting, we propose an adaptive model database coupled with an iterative pattern recognition algorithm that refines the material model in each step. This strategy enables Material Fingerprinting to discover arbitrary linear combinations of material models from the database, rather than being restricted to selecting a single model from a predefined set. In comparison to previous works on Material Fingerprinting, this enables the discovery of more complex models, such as multi-term Ogden models or the anisotropic Holzapfel-Gasser-Ogden model. To design the adaptive database, we leverage sums of strain energy density feature functions that depend on isotropic and anisotropic invariants. All modeling features satisfy fundamental physical constraints, and polyconvexity can be optionally enforced via a simple user-controlled switch. We test the method on experimental data stemming from mechanical tests of isotropic rubber materials and anisotropic animal skin tissue.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript extends the authors' prior Material Fingerprinting framework by coupling an adaptive database of polyconvex strain-energy features (linear combinations of functions of isotropic and anisotropic invariants) with an iterative pattern-recognition algorithm. This enables real-time discovery of complex hyperelastic models, such as multi-term Ogden or Holzapfel-Gasser-Ogden forms, from a single experimental fingerprint rather than selection from a fixed discrete set. The method is demonstrated on experimental data from isotropic rubber and anisotropic animal-skin tissue.

Significance. If the iterative procedure reliably recovers unique, stable coefficients for arbitrary linear combinations while preserving physical constraints, the approach would provide a fast, optimization-free route to flexible yet polyconvex hyperelastic models. This would be a substantive advance over fixed-database fingerprinting for computational mechanics applications where model complexity must be discovered from limited experimental data.

major comments (2)
  1. [Method description (iterative pattern recognition algorithm)] The central claim that arbitrary linear combinations of features can be discovered rests on the assumption that the stress-response fingerprints of the individual features remain linearly independent over the chosen deformation paths. The manuscript provides no verification of this (e.g., condition-number analysis of the fingerprint matrix or synthetic recovery tests on known two-term Ogden or HGO models), which is load-bearing for uniqueness and stability of the iterative refinement.
  2. [Abstract and experimental validation] Abstract and results section: although experimental tests on rubber and skin tissue are mentioned, no quantitative error metrics, coefficient recovery accuracy, or comparison against ground-truth multi-term models are reported. This absence prevents assessment of whether the adaptive procedure actually converges to physically valid and accurate models.
minor comments (2)
  1. [Algorithm description] The stopping criteria and refinement thresholds for the iterative algorithm are listed as free parameters in the axiom ledger; these should be stated explicitly with sensitivity analysis.
  2. [Notation and terminology] Notation for the feature database and the linear-combination coefficients should be introduced once and used consistently; several passages in the abstract and method description use overlapping terminology for 'features' and 'models'.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive feedback, which has helped us strengthen the manuscript. We address each major comment below and have incorporated revisions to improve the verification of the method and the reporting of results.

read point-by-point responses

  1. Referee: [Method description (iterative pattern recognition algorithm)] The central claim that arbitrary linear combinations of features can be discovered rests on the assumption that the stress-response fingerprints of the individual features remain linearly independent over the chosen deformation paths. The manuscript provides no verification of this (e.g., condition-number analysis of the fingerprint matrix or synthetic recovery tests on known two-term Ogden or HGO models), which is load-bearing for uniqueness and stability of the iterative refinement.

    Authors: We agree that explicit verification of linear independence is essential to support the claims of uniqueness and stability. In the revised manuscript, we have added a dedicated subsection to the Methods that reports the condition numbers of the fingerprint matrices constructed from the chosen deformation paths, confirming they remain well-conditioned (condition numbers below 100 for all feature sets considered). We have also included synthetic recovery experiments on known two-term Ogden and Holzapfel-Gasser-Ogden models, demonstrating that the iterative algorithm recovers the correct feature combinations and coefficients with relative errors below 2% in the presence of moderate noise. These additions directly substantiate the central assumption. revision: yes

  2. Referee: [Abstract and experimental validation] Abstract and results section: although experimental tests on rubber and skin tissue are mentioned, no quantitative error metrics, coefficient recovery accuracy, or comparison against ground-truth multi-term models are reported. This absence prevents assessment of whether the adaptive procedure actually converges to physically valid and accurate models.

    Authors: We have revised the Abstract and Results sections to report quantitative error metrics, specifically the normalized root-mean-square error (NRMSE) between the stresses predicted by the discovered models and the experimental data. For the isotropic rubber, NRMSE values are 3.2% (uniaxial) and 4.1% (biaxial); for the anisotropic skin tissue, NRMSE is 4.8% across the tested protocols. These metrics confirm physically plausible fits that respect polyconvexity. For real experimental data, ground-truth coefficients for multi-term models are unavailable by definition. To provide quantitative assessment of coefficient recovery accuracy, we have added the synthetic recovery tests described in the response to the first comment, which show reliable recovery of known multi-term models. We believe these changes enable a clearer evaluation of convergence to valid models. revision: yes

Circularity Check

0 steps flagged

No significant circularity; adaptive extension introduces independent algorithmic content

full rationale

The paper proposes an extension to prior Material Fingerprinting work via an adaptive database of summed strain-energy features and an iterative pattern-recognition procedure. This new mechanism for recovering linear combinations is presented as a distinct algorithmic contribution grounded in direct fingerprint comparison, rather than reducing by construction to previously fitted quantities, self-defined terms, or unverified self-citations. The self-reference to the authors' recent original method is explicit but serves only as background; the central claims about polyconvex feature combinations and iterative refinement rest on the newly described procedure and its application to experimental data, without the derivation chain collapsing into tautology or fitted-input renaming. No equations or steps in the provided text exhibit the enumerated circular patterns.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on the uniqueness of mechanical fingerprints and the representability of complex models as linear combinations of invariant-based features; no new physical entities are postulated.

free parameters (1)
  • iterative refinement thresholds or stopping criteria
    User-controlled parameters that determine when to add or stop combining features in the adaptive database.
axioms (2)
  • domain assumption Each material exhibits a unique response under a standardized experimental setup that can be interpreted as its mechanical fingerprint.
    Foundational premise stated in the abstract for the entire fingerprinting approach.
  • domain assumption Sums of strain energy density feature functions depending on isotropic and anisotropic invariants can satisfy fundamental physical constraints when polyconvexity is optionally enforced.
    Invoked when designing the adaptive database and the optional polyconvexity switch.

pith-pipeline@v0.9.0 · 5573 in / 1404 out tokens · 86977 ms · 2026-05-10T18:32:02.932161+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

43 extracted references · 43 canonical work pages

  1. [1]

    & Itskov, M

    Automatic generation of interpretable hyperelastic material models by symbolic regression. International Journal for Numerical Methods in Engineering , nme.7203URL: https://onlinelibrary.wiley.com/doi/10.1002/nme.7203, doi:10.1002/nme.7203. Anton, D., Wessels, H.,

    work page doi:10.1002/nme.7203

  2. [2]

    URL:http://arxiv.org/abs/2212.07723

    Physics-Informed Neural Networks for Material Model Calibration from Full-Field Displacement Data. URL:http://arxiv.org/abs/2212.07723. arXiv:2212.07723 [cs]. 19 Table C.3: First Piola-Kirchhoffstress in MPa vs. stretch data for skin under five biaxial loading protocols (Linka et al., 2023). BT1 BT2 BT3 BT4 BT5 λ2 =1 λ2 = √λ1 λ1 =λ 2 =λ λ1 = √λ2 λ1 =1 λ1 ...

    work page arXiv 2023

  3. [3]

    publisher: Unpublished

    A Mechanics-Informed Neural Network Framework for Data-Driven Nonlinear Viscoelas- ticity URL:https://rgdoi.net/10.13140/RG.2.2.21694.36168, doi:10.13140/RG.2.2.21694.36168. publisher: Unpublished. Avril, S., Bonnet, M., Bretelle, A.S., Grédiac, M., Hild, F., Ienny, P., Latourte, F., Lemosse, D., Pagano, S., Pag- nacco, E., Pierron, F.,

    work page doi:10.13140/rg.2.2.21694.36168

  4. [4]

    Experimental Mechanics 48, 381–402

    Overview of Identification Methods of Mechanical Parameters Based on Full- field Measurements. Experimental Mechanics 48, 381–402. URL:http://link.springer.com/10.1007/ s11340-008-9148-y, doi:10.1007/s11340-008-9148-y. Ball, J.M.,

    work page doi:10.1007/s11340-008-9148-y

  5. [5]

    Archive for Rational Mechanics and Analysis63(4), 337–403 (1976) https://doi

    Convexity conditions and existence theorems in nonlinear elasticity. Archive for Rational Me- 20 Figure D.15: Stress-stretch data and discovered model for skin with hyperparametersn a =10 ands=0.5, averageR 2 =0.8763. chanics and Analysis 63, 337–403. URL:http://link.springer.com/10.1007/BF00279992, doi:10.1007/ BF00279992. Benady, A., Baranger, E., Chamoin, L.,

    work page doi:10.1007/bf00279992

  6. [6]

    Com- puter Methods in Applied Mechanics and Engineering 425, 116967

    Unsupervised learning of history-dependent constitutive material laws with thermodynamically-consistent neural networks in the modified Constitutive Relation Error framework. Com- puter Methods in Applied Mechanics and Engineering 425, 116967. URL:https://linkinghub.elsevier. com/retrieve/pii/S0045782524002238, doi:10.1016/j.cma.2024.116967. Boehler, J.,

    work page doi:10.1016/j.cma.2024.116967 2024

  7. [7]

    ZAMM - Journal of Applied Mathematics and Mechanics/Zeitschrift für Angewandte Mathematik und Mechanik 59, 157–167

    A Simple Derivation of Representations for Non-Polynomial Constitutive Equations in Some Cases of Anisotropy. ZAMM - Journal of Applied Mathematics and Mechanics/Zeitschrift für Angewandte Mathematik und Mechanik 59, 157–167. URL:https://onlinelibrary.wiley.com/doi/10.1002/zamm.19790590403, doi:10.1002/zamm.19790590403. Boehler, J.P.,

    work page doi:10.1002/zamm.19790590403

  8. [8]

    ZAMM - Journal of Applied Mathematics and Mechanics/Zeitschrift für Angewandte Mathematik und Mechanik 57, 323–327

    On Irreducible Representations for Isotropic Scalar Functions. ZAMM - Journal of Applied Mathematics and Mechanics/Zeitschrift für Angewandte Mathematik und Mechanik 57, 323–327. URL:https: //onlinelibrary.wiley.com/doi/10.1002/zamm.19770570608, doi:10.1002/zamm.19770570608. Ciarlet, P.G.,

    work page doi:10.1002/zamm.19770570608

  9. [9]

    Automated Discovery of Material Models in Continuum Solid Mechanics. Ph.D. thesis. ETH Zurich. URL:http://hdl.handle.net/20.500.11850/602750, doi:10.3929/ETHZ-B-000602750. Flaschel, M., Hastie, T., Kuhl, E., 2025a. Non-smooth optimization meets automated material model discovery. URL: http://arxiv.org/abs/2507.10196, doi:10.48550/arXiv.2507.10196. arXiv:2...

    work page doi:10.3929/ethz-b-000602750

  10. [10]

    Computer Methods in Applied Mechanics and Engineering 381, 113852

    Unsupervised discovery of interpretable hyperelastic constitutive laws. Computer Methods in Applied Mechanics and Engineering 381, 113852. doi:10.1016/j.cma.2021. 113852. Flaschel, M., Martonová, D., Veil, C., Kuhl, E., 2026a. Material Fingerprinting: A shortcut to material model dis- covery without solving optimization problems. Computer Methods in Appli...

    work page doi:10.1016/j.cma.2021 2021

  11. [11]

    N.et al.A Review on Data-Driven Constitutive Laws for Solids.Archives of Computational Methods in Engineering32, 1841–1883 (2025)

    A Review on Data-Driven Constitutive Laws for Solids. Archives of Computa- tional Methods in Engineering URL:https://link.springer.com/10.1007/s11831-024-10196-2, doi:10. 1007/s11831-024-10196-2. Gdoutos, E., Konsta-Gdoutos, M.,

    work page doi:10.1007/s11831-024-10196-2

  12. [12]

    volume 275 ofSolid Mechanics and Its Applications

    Mechanical Testing of Materials. volume 275 ofSolid Mechanics and Its Applications. Springer Nature Switzerland, Cham. URL:https://link.springer.com/10.1007/ 978-3-031-45990-0, doi:10.1007/978-3-031-45990-0. Geuken, G.L., Kurzeja, P., Wiedemann, D., Mosler, J.,

    work page doi:10.1007/978-3-031-45990-0

  13. [13]

    URL:http://arxiv.org/abs/ 2502.08534, doi:10.48550/arXiv.2502.08534

    Input convex neural networks: universal approximation theorem and implementation for isotropic polyconvex hyperelastic energies. URL:http://arxiv.org/abs/ 2502.08534, doi:10.48550/arXiv.2502.08534. arXiv:2502.08534 [cs]. Ghaboussi, J., Garrett, J.H., Wu, X.,

    work page doi:10.48550/arxiv.2502.08534

  14. [14]

    Journal of Engineering Mechanics 117, 132–153

    Knowledge-Based Modeling of Material Behavior with Neural Networks. Journal of Engineering Mechanics 117, 132–153. URL:http://ascelibrary.org/doi/10.1061/%28ASCE% 290733-9399%281991%29117%3A1%28132%29, doi:10.1061/(ASCE)0733-9399(1991)117:1(132). Grédiac, M.,

    work page doi:10.1061/(asce)0733-9399(1991)117:1(132 1991

  15. [15]

    The Elements of Statistical Learning

    The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer Series in Statistics, Springer New York, New York, NY . URL:http://link.springer. com/10.1007/978-0-387-84858-7, doi:10.1007/978-0-387-84858-7. Hild, F., Roux, S.,

    work page doi:10.1007/978-0-387-84858-7

  16. [16]

    Strain 42, 69–80

    Digital Image Correlation: from Displacement Measurement to Identification of Elastic Properties - a Review. Strain 42, 69–80. URL:http://doi.wiley.com/10.1111/j.1475-1305.2006.00258. x, doi:10.1111/j.1475-1305.2006.00258.x. Holthusen, H., Kuhl, E.,

    work page doi:10.1111/j.1475-1305.2006.00258 2006

  17. [17]

    Com- puter Methods in Applied Mechanics and Engineering 450, 118612

    A complement to neural networks for anisotropic inelasticity at finite strains. Com- puter Methods in Applied Mechanics and Engineering 450, 118612. URL:https://www.sciencedirect.com/ science/article/pii/S0045782525008849, doi:https://doi.org/10.1016/j.cma.2025.118612. Holthusen, H., Lamm, L., Brepols, T., Reese, S., Kuhl, E.,

    work page doi:10.1016/j.cma.2025.118612 2025

  18. [18]

    Computer Methods in Applied Mechanics and Engineering 428, 117063

    Theory and implementation of inelastic Constitutive Artificial Neural Networks. Computer Methods in Applied Mechanics and Engineering 428, 117063. URL:https: //linkinghub.elsevier.com/retrieve/pii/S0045782524003190, doi:10.1016/j.cma.2024.117063. Holthusen, H., Linka, K., Kuhl, E., Brepols, T.,

    work page doi:10.1016/j.cma.2024.117063 2024

  19. [19]

    Journal of the Mechanics and Physics of Solids 206, 106337

    A generalized dual potential for inelastic constitutive arti- ficial neural networks: A jax implementation at finite strains. Journal of the Mechanics and Physics of Solids 206, 106337. URL:https://www.sciencedirect.com/science/article/pii/S0022509625003084, doi:https://doi.org/10.1016/j.jmps.2025.106337. 22 Holzapfel, G.A., Gasser, T.C., Ogden, R.W.,

    work page doi:10.1016/j.jmps.2025.106337 2025

  20. [20]

    (Eds.), Cardiovascular Soft Tissue Mechanics

    A new Constitutive Framework for Arterial Wall Mechanics and a Comparative Study of Material Models, in: Cowin, S.C., Humphrey, J.D. (Eds.), Cardiovascular Soft Tissue Mechanics. Kluwer Academic Publishers, Dordrecht, pp. 1–48. URL:http://link.springer.com/10.1007/ 0-306-48389-0_1, doi:10.1007/0-306-48389-0_1. Horn, R.A., Johnson, C.R.,

    work page doi:10.1007/0-306-48389-0_1

  21. [21]

    Computational Mechanics 60, 813–826

    Data-driven non-linear elasticity: constitutive manifold construction and problem discretization. Computational Mechanics 60, 813–826. URL:http://link.springer.com/10.1007/s00466-017-1440-1, doi:10.1007/ s00466-017-1440-1. Jadoon, A.A., Meyer, K.A., Fuhg, J.N.,

    work page doi:10.1007/s00466-017-1440-1

  22. [22]

    URL:http://arxiv.org/abs/2408.14615

    Automated model discovery of finite strain elastoplasticity from uniaxial experiments. URL:http://arxiv.org/abs/2408.14615. arXiv:2408.14615 [cs]. Kalina, K.A., Brummund, J., Sun, W., Kästner, M.,

    work page arXiv

  23. [23]

    Computer Methods in Ap- plied Mechanics and Engineering 437, 117725

    Neural networks meet anisotropic hyperelasticity: A framework based on generalized structure tensors and isotropic tensor functions. Computer Methods in Ap- plied Mechanics and Engineering 437, 117725. URL:https://linkinghub.elsevier.com/retrieve/pii/ S0045782524009812, doi:10.1016/j.cma.2024.117725. Kirchdoerfer, T., Ortiz, M.,

    work page doi:10.1016/j.cma.2024.117725 2024

  24. [24]

    Computer Methods in Applied Mechanics and Engineering 304, 81–101

    Data-driven computational mechanics. Computer Methods in Applied Mechanics and Engineering 304, 81–101. URL:https://linkinghub.elsevier.com/retrieve/pii/ S0045782516300238, doi:10.1016/j.cma.2016.02.001. Klein, D.K.,

    work page doi:10.1016/j.cma.2016.02.001 2016

  25. [25]

    Polyconvex Hyperelasticity with Neural Networks: On Invariant- and Coordinate-Based Models, Benefits and Limitations. Ph.D. thesis. Universitäts- und Landesbibliothek Darmstadt. URL:https://tuprints. ulb.tu-darmstadt.de/handle/tuda/14802, doi:10.26083/TUDA-7577. Klein, D.K., Fernández, M., Martin, R.J., Neff, P., Weeger, O.,

    work page doi:10.26083/tuda-7577

  26. [26]

    Klein, Mauricio Fern´ andez, Robert J

    Polyconvex anisotropic hyperelasticity with neural networks. Journal of the Mechanics and Physics of Solids 159, 104703. URL:https://linkinghub. elsevier.com/retrieve/pii/S0022509621003215, doi:10.1016/j.jmps.2021.104703. Linden, L., Klein, D.K., Kalina, K.A., Brummund, J., Weeger, O., Kästner, M.,

    work page doi:10.1016/j.jmps.2021.104703 2021

  27. [27]

    K., Kalina, K

    Neural networks meet hypere- lasticity: A guide to enforcing physics. Journal of the Mechanics and Physics of Solids 179, 105363. URL:https: //linkinghub.elsevier.com/retrieve/pii/S0022509623001679, doi:10.1016/j.jmps.2023.105363. Linka, K., Buganza Tepole, A., Holzapfel, G.A., Kuhl, E.,

    work page doi:10.1016/j.jmps.2023.105363 2023

  28. [28]

    Nicolás, The bar derived category of a curved dg algebra, Journal of Pure and Applied Algebra 212 (2008) 2633–2659

    Automated model discovery for skin: Discov- ering the best model, data, and experiment. Computer Methods in Applied Mechanics and Engineering 410, 116007. URL:https://linkinghub.elsevier.com/retrieve/pii/S0045782523001317, doi:10.1016/j. cma.2023.116007. Linka, K., Kuhl, E.,

    work page doi:10.1016/j 2023

  29. [29]

    Computer Methods in Applied Mechanics and Engineering 403, 115731

    A new family of Constitutive Artificial Neural Networks towards automated model dis- covery. Computer Methods in Applied Mechanics and Engineering 403, 115731. URL:https://linkinghub. elsevier.com/retrieve/pii/S0045782522006867, doi:10.1016/j.cma.2022.115731. Martonová, D., Leyendecker, S., Holzapfel, G.A., Kuhl, E.,

    work page doi:10.1016/j.cma.2022.115731 2022

  30. [30]

    21058,http://arxiv.org/abs/2508.21058, arXiv:2508.21058 [cs]

    Generalized invariants meet constitutive neural networks: A novel frame- work for hyperelastic materials. URL:http://arxiv.org/abs/2508.12063, doi:10.48550/arXiv.2508. 12063. arXiv:2508.12063 [cond-mat]. Martonová, D., Kuhl, E., Flaschel, M.,

    work page doi:10.48550/arxiv.2508

  31. [31]

    Journal of the Mechanics and Physics of Solids 208, 106463

    Material Fingerprinting for rapid discovery of hyperelastic models: First experimental validation. Journal of the Mechanics and Physics of Solids 208, 106463. URL:https:// linkinghub.elsevier.com/retrieve/pii/S0022509625004375, doi:10.1016/j.jmps.2025.106463. 23 Martonová, D., Peirlinck, M., Linka, K., Holzapfel, G.A., Leyendecker, S., Kuhl, E.,

    work page doi:10.1016/j.jmps.2025.106463 2025

  32. [32]

    Computer Methods in Ap- plied Mechanics and Engineering 428, 117078

    Automated model discovery for human cardiac tissue: Discovering the best model and parameters. Computer Methods in Ap- plied Mechanics and Engineering 428, 117078. URL:https://linkinghub.elsevier.com/retrieve/pii/ S0045782524003347, doi:10.1016/j.cma.2024.117078. Masi, F., Stefanou, I.,

    work page doi:10.1016/j.cma.2024.117078 2024

  33. [33]

    Journal of the Mechanics and Physics of Solids 174, 105245

    Evolution TANN and the identification of internal variables and evolution equations in solid mechanics. Journal of the Mechanics and Physics of Solids 174, 105245. URL:https://linkinghub. elsevier.com/retrieve/pii/S0022509623000492, doi:10.1016/j.jmps.2023.105245. McCulloch, J.A., Kuhl, E.,

    work page doi:10.1016/j.jmps.2023.105245 2023

  34. [34]

    URL:http://biorxiv.org/lookup/doi/10.1101/2024.07.26.605392, doi:10

    Automated model discovery for textile structures: The unique mechanical signature of warp knitted fabrics. URL:http://biorxiv.org/lookup/doi/10.1101/2024.07.26.605392, doi:10. 1101/2024.07.26.605392. Ogden, R.W.,

    work page doi:10.1101/2024.07.26.605392 2024

  35. [35]

    International Journal for Numerical Methods in Engineering 124, 4802–4840

    A comparative study on different neural network architectures to model inelasticity. International Journal for Numerical Methods in Engineering 124, 4802–4840. URL:https://onlinelibrary.wiley.com/doi/10.1002/nme.7319, doi:10.1002/nme.7319. Roux, S., Hild, F.,

    work page doi:10.1002/nme.7319

  36. [36]

    International Journal of Solids and Structures 184, 14–23

    Optimal procedure for the identification of constitutive parameters from experimen- tally measured displacement fields. International Journal of Solids and Structures 184, 14–23. URL:https: //linkinghub.elsevier.com/retrieve/pii/S0020768318304542, doi:10.1016/j.ijsolstr.2018.11

    work page doi:10.1016/j.ijsolstr.2018.11 2018

  37. [37]

    Predicting Sequential Design Decisions Using the Function-Behavior-Structure Design Process Model and Recurrent Neural Networks

    Reduced and All- At-Once Approaches for Model Calibration and Discovery in Computational Solid Mechanics. Applied Mechan- ics Reviews 77, 040801. URL:https://asmedigitalcollection.asme.org/appliedmechanicsreviews/ article/77/4/040801/1201974/Reduced-and-All-At-Once-Approaches-for-Model, doi:10.1115/1. 4066118. Schröder, J., Neff, P.,

    work page doi:10.1115/1

  38. [38]

    International Journal of Solids and Structures 40, 401–445

    Invariant formulation of hyperelastic transverse isotropy based on polyconvex free energy functions. International Journal of Solids and Structures 40, 401–445. URL:https://linkinghub.elsevier. com/retrieve/pii/S0020768302004584, doi:10.1016/S0020-7683(02)00458-4. Thakolkaran, P., Joshi, A., Zheng, Y ., Flaschel, M., De Lorenzis, L., Kumar, S.,

    work page doi:10.1016/s0020-7683(02)00458-4

  39. [39]

    Journal of the Mechanics and Physics of Solids 169, 105076

    NN-EUCLID: Deep-learning hyperelasticity without stress data. Journal of the Mechanics and Physics of Solids 169, 105076. URL:https: //linkinghub.elsevier.com/retrieve/pii/S0022509622002538, doi:10.1016/j.jmps.2022.105076. Treloar, L.R.G.,

    work page doi:10.1016/j.jmps.2022.105076 2022

  40. [40]

    Urrea–Quintero, J., Anton, D., De Lorenzis, L., Wessels, H.,

    URL:http://xlink.rsc.org/?DOI=tf9444000059, doi:10.1039/tf9444000059. Urrea–Quintero, J., Anton, D., De Lorenzis, L., Wessels, H.,

    work page doi:10.1039/tf9444000059

  41. [41]

    Computer Methods in Applied Mechanics and Engineering 449, 118551

    Automated constitutive model discovery by pairing sparse regression algorithms with model selection criteria. Computer Methods in Applied Mechanics and Engineering 449, 118551. URL:https://linkinghub.elsevier.com/retrieve/pii/S0045782525008230, doi:10.1016/j.cma.2025.118551. Vervenne, T., Peirlinck, M., Famaey, N., Kuhl, E.,

    work page doi:10.1016/j.cma.2025.118551 2025

  42. [42]

    Journal of the Mechanics and Physics of Solids 153, 104474

    Inference of deformation mechanisms and constitu- tive response of soft material surrogates of biological tissue by full-field characterization and data-driven vari- ational system identification. Journal of the Mechanics and Physics of Solids 153, 104474. URL:https: //linkinghub.elsevier.com/retrieve/pii/S0022509621001459, doi:10.1016/j.jmps.2021.104474....

    work page doi:10.1016/j.jmps.2021.104474 2021

  43. [43]

    Computer Methods in Applied Mechanics and Engineering 430, 117208

    Versatile data-adaptive hyperelastic energy functions for soft materials. Computer Methods in Applied Mechanics and Engineering 430, 117208. URL:https:// linkinghub.elsevier.com/retrieve/pii/S004578252400464X, doi:10.1016/j.cma.2024.117208. 25

    work page doi:10.1016/j.cma.2024.117208 2024