Pith. sign in

REVIEW 4 major objections 6 minor 115 references

A Physics-Informed Data-Driven Discovery for Constitutive Modeling of Compressible, Nonlinear, History-Dependent Soft Materials under Multiaxial Cyclic Loading

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A hybrid GPR-LSTM surrogate trained on tensor-invariant response functions reproduces a nonlinear visco-hyperelastic model under multiaxial cyclic loading, including extrapolated stretches, strain rates, and tension/compression.

desk verdict A coherent hybrid GPR/LSTM constitutive surrogate with real architectural merit, but the 'discovery' claim is overreaching: all evidence is synthetic Holzapfel data and the non-equilibrium basis has a representation gap the axisymmetric tests cannot detect. read the letter →

arxiv 2507.12683 v1 pith:D3R4QVEG submitted 2025-07-16 cond-mat.soft

Alireza Ostadrahimi , Amir Teimouri , Kshitiz Upadhyay , Guoqiang Li This is my paper
classification cond-mat.soft MSC 74D1074B2068T07 PACS 83.60.Bc83.80.Va07.05.Mh
keywords physics-informedmachinelearningvisco-hyperelasticityGaussianprocessregressionLSTMintegritybasismultiaxialcyclicloadingthermodynamicconsistencyconstitutivemodeling
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 tries to establish that a physics-informed machine learning surrogate can replace an internal-state-variable visco-hyperelastic constitutive model without losing physical fidelity. It claims that by training a Gaussian Process Regression (GPR) model on the equilibrium stress response functions and a Long Short-Term Memory (LSTM) recurrent network on the non-equilibrium response functions, the surrogate accurately captures nonlinear, rate-dependent, history-dependent stress under multiaxial cyclic loading. The trained model is shown to predict stress for deformation states outside the training range, including different stretch levels, strain rates, tension and compression, while satisfying objectivity, material symmetry, angular momentum balance, and non-negative dissipation. A sympathetic reader would care because this offers a path to data-driven constitutive modeling that generalizes beyond observed loading conditions without enforcing an explicit closed-form strain energy function.

What carries the argument

The key machinery is the integrity basis representation of stress, Eqns. 10-11, where stress components are written as weighted linear combinations of the isotropic generators I, C, and $C^{{-1}}$, with the weights being response functions of the invariants I1, I2 (and J). The equilibrium response functions are learned by GPR with a Matérn 3/2 kernel, while the history-dependent response functions are learned by a three-layer LSTM RNN with a physics-informed loss that adds a penalty when the computed dissipation (from the Clausius-Duhem inequality) is negative. This representation enforces objectivity and material symmetry by construction and converts constitutive modeling into supervised regression of scalar invariant-dependent coefficients.

What would settle it

Train the same GPR-LSTM surrogate on data generated by a visco-hyperelastic model whose dissipative stress requires additional generators, for example a model with a transversely isotropic or orthotropic viscous branch, or a two-network theory with a non-isotropic internal tensor; if the surrogate cannot reproduce the unseen multiaxial stress components despite good training fits, the completeness of the two-generator basis for the non-equilibrium stress is falsified.

Watch

Extended reading notes

Core claim

The central claim is that the total stress of a generalized internal-state-variable visco-hyperelastic material can be decomposed into equilibrium volumetric, equilibrium isochoric, and non-equilibrium isochoric parts, and that each part can be learned separately as a function of tensor invariants using a hybrid GPR-LSTM architecture. The GPR learns the volumetric response function delta(J) and the isochoric response functions chi1(I1,I2) and chi2(I1,I2), while the LSTM learns the time-dependent response functions xi1(t,I1,I2) and xi2(t,I1,I2) that capture the summed Maxwell branches. Because stresses are expressed as linear combinations of a fixed integrity basis (identity, C, $C^{{-1}}$) multiplied by these learned response functions, objectivity, isotropy, and symmetry are built in, and the second law is enforced as a dissipation penalty in the LSTM loss. Using the Holzapfel differential viscoelastic model to generate training data, the paper claims the surrogate generalizes beyond training domains in stretch, strain rate, and tension/compression, with dissipation staying non-negative and predictions robust to about 6% synthetic noise.

Load-bearing premise

The framework assumes that the complete non-equilibrium stress of any history-dependent isotropic viscoelastic material can be represented using only the two isotropic generators C and $C^{{-1}}$, so that two invariant-based response functions suffice for all dissipative branches.

Editorial extensions

If this is right

  • If the surrogate indeed matches the Holzapfel model on unseen multiaxial paths, then data-driven constitutive models can be calibrated from a modest set of multiaxial cyclic experiments and still extrapolate to untrained stretch levels and strain rates.
  • The decomposition into volumetric, isochoric equilibrium, and isochoric non-equilibrium surrogates means each component can be trained separately, making the learning problem lower-dimensional and more data-efficient than raw stress-strain sequence fitting.
  • Because the learned response functions are scalar functions of invariants, the resulting model can be checked against known theoretical forms, connected to classical constitutive theory, and reused in finite element codes via the stress expression of Eq. 16.
  • The non-negative dissipation constraint, enforced as a soft penalty during training, provides a route to thermodynamically consistent RNN-based constitutive models without explicitly enforcing a full evolution equation.
  • Robustness to about 6% synthetic noise suggests the framework could tolerate realistic experimental noise, which matters if the decomposition procedure (asymptote extraction and basis-coefficient inversion via least squares) amplifies measurement error.

Reading between the lines

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

  • The paper's success on extrapolation in stretch and time suggests, but does not prove, that the learned invariant-response functions have locally smooth extrapolation behavior; a testable extension would be to query the surrogate on random multiaxial paths that are not one-cyclic or proportional.
  • The framework is demonstrated only on data generated by the Holzapfel model with isotropic symmetry and two isotropic generators; a natural extension is to anisotropic soft tissues or fiber-reinforced elastomers, where the integrity basis must be enlarged and the claim of complete representation would need to be revisited.
  • The stress-decomposition methodology, which fits asymptotic equilibrium stress and then subtracts it, would fail if a material has a very slow relaxation component that never equilibrates within the test window; a practical extension would require a systematic protocol for choosing test durations such that the asymptotic fit is reliable.
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

4 major / 6 minor

Summary. The paper proposes a hybrid physics-informed machine learning framework for finite-strain, isotropic, compressible visco-hyperelasticity. The equilibrium volumetric and isochoric stresses are represented through Gaussian Process Regression on invariant-based response functions, while the non-equilibrium, history-dependent stress is represented through an LSTM-based RNN whose outputs are projected onto a two-generator integrity basis. Thermodynamic constraints are imposed through a stress-free reference state in the GPR and a dissipation penalty in the RNN loss. Training data are generated from the Holzapfel differential viscoelastic model with Neo-Hookean volumetric and Mooney-Rivlin isochoric energies, for two synthetic datasets (short-term and long-term relaxation). The model is tested on multiaxial cyclic protocols with varied strain rates, stretch levels in tension and compression, extrapolation beyond training ranges, dissipation non-negativity, and synthetic noise. The authors conclude that the framework accurately captures nonlinear, rate-dependent, history-dependent behavior and generalizes beyond its training domain.

Significance. If the central claims hold, this is a useful step toward interpretable, physics-constrained surrogates for history-dependent soft materials. The strength of the paper is its architectural design: learning response functions on an integrity basis rather than raw stress-strain pairs, combining GPR (equilibrium) with LSTM (memory), and explicitly checking stress-free states, objectivity, symmetry, and dissipation. The multiaxial cyclic test suite, including extrapolation and noise, is broader than in many comparable studies. However, the evidence presented is largely self-consistency validation: all training and test data derive from the same Holzapfel generator, the dissipation constraint itself imports the Holzapfel-specific form of the internal-variable evolution, and the non-equilibrium stress representation in Eq. (11) is assumed complete without a representation theorem for internal-variable functionals. The generalizability and 'data-driven discovery' claims therefore need either additional evidence or substantial reframing.

major comments (4)
  1. [Section 2.2, Eq. (11)] The load-bearing assumption that the sum of non-equilibrium branches can be expressed using the same two-generator basis {I, C^{-1}} as the equilibrium stress is not justified by a representation theorem. For an internal-variable viscoelastic material, the non-equilibrium stress is a functional of C and the internal variables/history; isotropic functions of two tensors require additional generators involving the internal variables (e.g., H, CH+HC, C^2H+HC^2). The LSTM hidden state can encode history, but the output is still projected onto Dev(I) and Dev(C^{-1}), so non-equilibrium stresses that are not coaxial with the current C cannot be represented. All numerical tests use lambda_2 = lambda_3 with fixed principal axes, for which any deviatoric stress with the same two-eigenvalue structure lies in the span of these generators; hence the test suite cannot detect this gap. The central generality claim ('general', 'multiaxial') is therefore unsupported. Please either provide a representation theorem for the assumed form, or restrict the claims to axisymmetric coaxial histories and add a non-proportional/rotating-principal-axis test.
  2. [Sections 3.2, 4.1, and Case 4] The phrase 'data-driven discovery' is too strong for the present validation. Training data, test data, and the physics penalty all come from the same Holzapfel-Mooney-Rivlin generator: the response functions are extracted from that model, and the dissipation constraint in Eqs. (29)-(30) is imported from the same generating model. Consequently, the extrapolation tests in Cases 1-3 are self-consistency checks of the surrogate's ability to interpolate/extrapolate the generator, not evidence that the framework discovers a general constitutive law. The dissipation verification in Case 4 is likewise a self-consistency check: a model trained on outputs of a dissipative generator and penalized with that generator's own dissipation inequality will naturally produce non-negative dissipation. To support the discovery claim, the framework should be tested on data generated by a different constitutive class or on experimental data, or the claims should be explicitly limited to 'surrogate modeling of a given generator family.'
  3. [Section 4.3, Eq. (24)] The paper defines a percent relative error metric in Eq. (24) but never reports quantitative error values for any of the five cases. The conclusions rely on visual inspection of figures ('model predictions match the training data closely', 'accurate predictions'), which is not a reproducible quantitative standard. Please report mean and maximum errors for training and testing in each case, including the extrapolation and noise cases, and state which cases use normalized vs. absolute errors. This is needed to support the accuracy and generalization claims.
  4. [Section 4.1.2 and Figures 15] The exact relaxation time spectra and branch weights for the short-term and long-term datasets are never specified. The text says the ST dataset has four Maxwell branches with 'a relaxation time spectrum and uniform weighting factors' and the LT dataset has a single 'long relaxation time', but no numerical values are given. Without these values the synthetic datasets cannot be reproduced, and it is impossible to assess whether the tested time ranges (0.1 s to 200 s) actually sample the intended relaxation regimes. Please provide the full generation parameters, including relaxation times, weights, mu_alpha, and the loading/unloading time histories.
minor comments (6)
  1. [Section 4.2] The GPR kernel description is inconsistent: Section 3.4 states a Matérn 3/2 kernel is used, while Section 4.2 says the kernel is a constant kernel combined with a radial basis function kernel. Please reconcile these statements and specify which kernel was actually used.
  2. [Equation numbering] The equation numbering is disordered: Eq. (25) is introduced in Section 4.1.2, then Eqs. (26)-(27), then Eq. (28) in Section 3.5.2, and Eq. (24) is defined later in Section 4.3. This makes cross-referencing difficult; please renumber sequentially.
  3. [Section 3.5.2] In Eqs. (29)-(31) the notation is unclear: Gamma_alpha is both an internal deviatoric history variable and appears with a finite-difference update, while Q_alpha is called both a deviatoric internal variable and a non-equilibrium stress. Please define the relation between Gamma_alpha and Q_alpha explicitly and state which quantity the RNN actually outputs.
  4. [Figure 14 caption] The caption contains a typo: 'Isochoric viscoelastic tress-stretch behavior' should read 'stress-stretch behavior'.
  5. [Section 4.3.1] The GPR evaluation in Case 1 only shows qualitative curves. Since the GPR is a core component of the framework, please report its training and test errors separately from the RNN errors, particularly for the extrapolation cases.
  6. [General] No code or data availability statement is provided. Given the synthetic-data setup, releasing the generation scripts and trained model configurations would substantially strengthen reproducibility.

Circularity Check

2 steps flagged · score 6.0 of 10

The dissipation check in Case 4 is imposed, not discovered, because Eqs. 28-31 from the same Holzapfel model that generated the data enter the RNN loss; separately, the general non-equilibrium integrity-basis representation is imported from the authors' prior work without a representation theorem for internal variables.

  1. self definitional [Section 3.5.2 and Section 4.3.4 (Case 4)]
    "Using the equilibrium isochoric stress prediction from the trained GPR model in Eq. 31 and substituting the resulting equation into Eq. 28 yields the viscous dissipation at a given timestep and deformation; the non-negativity of this viscous dissipation is incorporated into the loss function of our RNN model. ... The dissipation rates remain strictly non-negative throughout the process, satisfying the thermodynamic requirement."

    The RNN loss is defined with a penalty for negative dissipation, and Eq. 31 is the finite-difference form of the Holzapfel evolution equation (Eq. 30) that also generated the non-equilibrium training data. Therefore the Case 4 result is not an independent verification: non-negative dissipation is enforced by construction during training, and the subsequent report of adherence to the second law reads the constraint back out of the model. The same Eqs. 28-31 used for evaluation are the ones embedded in the loss.

  2. ansatz smuggled in via citation [Section 2.2, Eq. 11]
    "Recently, Upadhyay et al. demonstrated [106] that the individual stress components in the generalized external state variable-based visco-hyperelastic constitutive framework can be written as a weighted linear combination of the components of certain irreducible integrity bases. ... Inspired by the integrity basis representation in the equilibrium stress Eq. 10b and the representation of each non-equilibrium branch evolution through Eq. 9, the sum of all non-equilibrium branches can also be expressed using the same integrity basis:"

    The cited support for the integrity-basis expansion is the authors' own prior work ([106]) for the external-state-variable framework; the internal-variable, history-dependent branch in Eq. 11 is added by assertion rather than by a representation theorem for functions of C plus internal variables or history. Because the RNN output is projected only onto Dev(I) and Dev(C^{-1}), any non-equilibrium stress not coaxial with the current C is excluded by construction. All numerical tests use lambda2 = lambda3 diagonal states, so the tests cannot reveal this restriction; the claimed general multiaxial validity is thus inherited from a self-citation rather than established.

full rationale

The paper is a synthetic-data benchmark: an LSTM/GPR surrogate is trained on outputs of the Holzapfel visco-hyperelastic model and tested on held-out states of that same model. That is not circular by itself, since the networks do not encode the Holzapfel stress formula directly and the extrapolation tests are non-trivial. The circularity is confined to two load-bearing places. First, the Case 4 thermodynamic-consistency check is imposed through the RNN loss using Eqs. 28-31, which come from the same Holzapfel model that generated the data; reporting the resulting non-negative dissipation as a verification is reading the constraint back. Second, the general non-equilibrium representation in Eq. 11 is imported from the authors' prior external-state-variable work by analogy, without a representation theorem covering internal variables, so the framework's general multiaxial claim rests on an ansatz rather than a derivation. The stress-extrapolation results themselves are not forced by construction, which keeps the score at partial circularity rather than full equivalence.

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

The framework relies heavily on synthetic generation and imported physics: the material parameters and the dissipation form are taken from the Holzapfel model, and the neural network hyperparameters are chosen by hand. The central claim therefore rests on the representational assumption in Eq. 11 and on the representativeness of the synthetic generator.

free parameters (7)
  • Bulk modulus K = 10 (chosen)
    Used in the Neo-Hookean volumetric energy to generate synthetic training data; not learned from experiments.
  • Mooney-Rivlin parameters C10, C01 = C10=10, C01=5
    Used to generate equilibrium isochoric stress; hand-chosen generator parameters.
  • Viscous parameter mu_alpha = 1 (set)
    In Eq. 29, the non-negative parameter mu_alpha is set to 1, simplifying the Holzapfel dissipation form.
  • Relaxation time spectrum and branch weights = not specified numerically
    ST dataset has 4 Maxwell branches with a relaxation time spectrum; LT dataset has 1 branch with a long relaxation time; exact values are not reported, which also hurts reproducibility.
  • GPR kernel hyperparameters = optimized by maximum log-likelihood
    Length scale and noise level are fitted to the synthetic training set; they are not derived from physics.
  • LSTM architecture hyperparameters = layers 128,64,32; learning rate 1e-3; epochs 500; batch 32/64
    Chosen by hand, not optimized against a benchmark.
  • Synthetic noise level = approximately 6%
    Chosen for the noise sensitivity test; not tied to a measured experimental uncertainty.
assumptions (7)
  • standard math The Coleman-Noll procedure and the Clausius-Duhem inequality are the correct thermodynamic framework for isothermal viscoelasticity.
    Used in Section 2.1 to derive the stress decomposition and the dissipation inequality.
  • standard math The multiplicative split of the deformation gradient into volumetric and isochoric parts is admissible.
    Used throughout Section 2.
  • domain assumption Material isotropy; the stress is an isotropic function of C.
    The integrity basis Eq. 12 uses only isotropic invariants; no anisotropy is addressed.
  • ad hoc to paper The non-equilibrium stress can be represented by the same two-generator integrity basis as the equilibrium stress (Eq. 11).
    No representation theorem for functions of multiple tensors is invoked; this is the weakest mathematical premise in the paper.
  • domain assumption The Holzapfel differential viscoelastic model (Eq. 25) is a valid generator of training data representative of real soft materials.
    All training and test data are synthetic from this model family; the assumption is untested on experiments.
  • domain assumption Stress relaxation at constant stretch reaches an asymptote that can be identified by curve fitting (Section 3.2.1).
    For real experiments, this asymptotic extraction is approximate and may bias the learned response functions.
  • ad hoc to paper The dissipation constraint can be imposed as a soft penalty in the RNN loss using the Holzapfel-specific form of Eqs. 29-30.
    This imports a specific evolution equation into the supposedly data-driven model, constraining the learned response.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Physics-Informed Data-Driven Discovery for Constitutive Modeling of Compressible, Nonlinear, History-Dependent Soft Materials under Multiaxial Cyclic Loading." pith.science (2026). https://pith.science/paper/D3R4QVEG

@misc{pith2026250712683,
  author       = {Pith},
  title        = {Pith review of: A Physics-Informed Data-Driven Discovery for Constitutive Modeling of Compressible, Nonlinear, History-Dependent Soft Materials under Multiaxial Cyclic Loading},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/D3R4QVEG}},
  note         = {Machine review of arXiv:2507.12683}
}
read the original abstract

We propose a general hybrid physics-informed machine learning framework for modeling nonlinear, history-dependent viscoelastic behavior under multiaxial cyclic loading. The approach is built on a generalized internal state variable-based visco-hyperelastic constitutive formulation, where stress is decomposed into volumetric, isochoric hyperelastic, and isochoric viscoelastic components. Gaussian Process Regression (GPR) models the equilibrium response, while Recurrent Neural Networks (RNNs) with Long Short-Term Memory (LSTM) units capture time-dependent viscoelastic effects. Physical constraints, including objectivity, material symmetry, and thermodynamic consistency, are enforced to ensure physically valid predictions. After developing the general form of the surrogate model based on tensor integrity bases and response functions, we employed the nonlinear Holzapfel differential viscoelastic model to generate training data. Two datasets, one for short-term and another for long-term relaxation, are constructed to span a wide range of material memory characteristics. The model is trained and tested under diverse multiaxial loading conditions, including different stretch levels applied independently in the longitudinal and transverse directions, varying strain rates, and both tension and compression states, even beyond the training domain. Energy dissipation is explicitly analyzed at different strain rates for both datasets to verify thermodynamic consistency through the second law. The results show that the proposed framework accurately captures complex, nonlinear, and rate-dependent material responses. Moreover, it demonstrates strong robustness to synthetic noise, enabling generalizable and physically consistent predictions under realistic and variable loading scenarios.

Figures

Figures reproduced from arXiv: 2507.12683 by the authors.

Figure 1
Figure 1. Modeling framework combining viscoelastic behavior and multiaxial stress conditions. [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

115 extracted references · 50 canonical work pages

  1. [1]

    A new constitutive framework for arterial wall mechanics and a comparative study of material models,

    G. A. Holzapfel, T. C. Gasser, and R. W. Ogden, “A new constitutive framework for arterial wall mechanics and a comparative study of material models,” J Elast , vol. 61, no. 1 –3, 2000, doi: 10.1023/A:1010835316564. 36

    doi:10.1023/a:1010835316564

  2. [2]

    Hyperelastic constitutive modeling of hydrogels based on primary deformation modes and validation under 3D stress states,

    K. Upadhyay, G. Subhash, and D. Spearot, “Hyperelastic constitutive modeling of hydrogels based on primary deformation modes and validation under 3D stress states,” Int J Eng Sci , vol. 154, Sep. 2020, doi: 10.1016/j.ijengsci.2020.103314

    arXiv 2020

  3. [3]

    Visco-hyperelastic constitutive modeling of strain rate sensitive soft materials,

    K. Upadhyay, G. Subhash, and D. Spearot, “Visco-hyperelastic constitutive modeling of strain rate sensitive soft materials,” J Mech Phys Solids, vol. 135, Feb. 2020, doi: 10.1016/j.jmps.2019.103777

    arXiv 2020

  4. [4]

    A physics-informed multi-agents model to predict thermo-oxidative/hydrolytic aging of elastomers,

    A. Ghaderi, V . Morovati, Y . Chen, and R. Dargazany, “A physics-informed multi-agents model to predict thermo-oxidative/hydrolytic aging of elastomers,” Int J Mech Sci , vol. 223, Jun. 2022, doi: 10.1016/j.ijmecsci.2022.107236

    arXiv 2022

  5. [5]

    On a fully three-dimensional finite-strain viscoelastic damage model: Formulation and computational aspects,

    J. C. Simo, “On a fully three-dimensional finite-strain viscoelastic damage model: Formulation and computational aspects,” Comput Methods Appl Mech Eng, vol. 60, no. 2, 1987, doi: 10.1016/0045- 7825(87)90107-1

    doi:10.1016/0045- 1987

  6. [6]

    A THEORY OF FINITE VISCOELASTICITY AND NUMERICAL ASPECTS,

    S. A. Reese and S. Y . Govindjee, “A THEORY OF FINITE VISCOELASTICITY AND NUMERICAL ASPECTS,” 1998

    1998

  7. [7]

    From machine learning to deep learning: progress in machine intelligence for rational drug discovery,

    L. Zhang, J. Tan, D. Han, and H. Zhu, “From machine learning to deep learning: progress in machine intelligence for rational drug discovery,” Nov. 01, 2017, Elsevier Ltd . doi: 10.1016/j.drudis.2017.08.010

    work page doi:10.1016/j.drudis.2017.08.010 2017

  8. [8]

    Skamniotis, D

    C. Skamniotis, D. Long, M. Wenman, and D. S. Balint, “On the effect of nuclear fission cladding stresses on Zirconium hydride orientation and dislocation strain energy fields via discrete dislocation plasticity and crystal plasticity finite element modelling,” J Mech Phys Solids, vol. 195, p. 105924, Feb. 2025, doi: 10.1016/j.jmps.2024.105924

    arXiv 2025

Show all 115 references
  1. [9]

    Parametric extended physics-informed neural networks for solid mechanics with complex mixed boundary conditions,

    G. Cao and X. Wang, “Parametric extended physics-informed neural networks for solid mechanics with complex mixed boundary conditions,” J Mech Phys Solids , vol. 194, Jan. 2025, doi: 10.1016/j.jmps.2024.105944

    2025

  2. [10]

    Physics-informed recovery of nonlinear residual stress fields in an inverse continuum framework,

    J. A. Sanz-Herrera and A. Goriely, “Physics-informed recovery of nonlinear residual stress fields in an inverse continuum framework,” J Mech Phys Solids, vol. 200, Jul. 2025, doi: 10.1016/j.jmps.2025.106079

    2025

  3. [11]

    Enforcing physics onto PINNs for more accurate inhomogeneous material identification,

    B. van der Heijden, X. Li, G. Lubineau, and E. Florentin, “Enforcing physics onto PINNs for more accurate inhomogeneous material identification,” Comput Methods Appl Mech Eng, vol. 441, p. 117993, Jun. 2025, doi: 10.1016/j.cma.2025.117993

    2025

  4. [12]

    Statistical-Physics-Informed Neural Networks (Stat-PINNs): A Machine Learning Strategy for Coarse-graining Dissipative Dynamics,

    S. Huang, Z. He, N. Dirr, J. Zimmer, and C. Reina, “Statistical-Physics-Informed Neural Networks (Stat-PINNs): A Machine Learning Strategy for Coarse-graining Dissipative Dynamics,” Dec. 2023, doi: 10.1016/j.jmps.2024.105908

    2023

  5. [13]

    Non -linear viscoelastic laws for soft biological tissues,

    D. P. Pioletti and L. R. Rakotomanana, “Non -linear viscoelastic laws for soft biological tissues,” 2000

    2000

  6. [14]

    Data-driven homogenisation of viscoelastic porous elastomers: Feedforward versus knowledge -based neural networks,

    M. O. Bozkurt and V . L. Tagarielli, “Data-driven homogenisation of viscoelastic porous elastomers: Feedforward versus knowledge -based neural networks,” Int J Mech Sci , vol. 286, Jan. 2025, doi: 10.1016/j.ijmecsci.2024.109824

    2025

  7. [15]

    Elasticity of soft tissues in simple elongation’,

    Y . C. B Fung and Y . C. B Elasticity, “Elasticity of soft tissues in simple elongation’,” 1967. 37

    1967

  8. [16]

    On large strain viscoelasticity: Continuum formulation and finite element applications to elastomeric structures,

    G. A. Holzapfel, “On large strain viscoelasticity: Continuum formulation and finite element applications to elastomeric structures,” Int J Numer Methods Eng, vol. 39, no. 22, 1996, doi: 10.1002/(SICI)1097-0207(19961130)39:22<3903::AID-NME34>3.0.CO;2-C

    1996 doi

  9. [17]

    Nonlinear Solid Mechanics II,

    A. G. Holzapfel, “Nonlinear Solid Mechanics II,” International Coal Preparation Congress 2010, Conference Proceedings, vol. 2, 2010

    2010

  10. [18]

    ON THE LARGE DEFORMATION BEHA VIOUR OF REINFORCED RUBBER AT DIFFERENT TEMPERATURES,

    A. Lion, “ON THE LARGE DEFORMATION BEHA VIOUR OF REINFORCED RUBBER AT DIFFERENT TEMPERATURES,” 1997

    1997

  11. [19]

    Physics -driven neural networks for nonlinear micromechanics,

    Z. Xiong, P. Yang, and P. Zhao, “Physics -driven neural networks for nonlinear micromechanics,” Int J Mech Sci, vol. 273, Jul. 2024, doi: 10.1016/j.ijmecsci.2024.109214

    2024

  12. [20]

    Modeling finite-strain plasticity using physics - informed neural network and assessment of the network performance,

    S. Niu, E. Zhang, Y . Bazilevs, and V . Srivastava, “Modeling finite-strain plasticity using physics - informed neural network and assessment of the network performance,” J Mech Phys Solids, vol. 172, Mar. 2023, doi: 10.1016/j.jmps.2022.105177

    2023

  13. [21]

    A physics-informed 3D surrogate model for elastic fields in polycrystals,

    L. Monteiro Fernandes et al., “A physics-informed 3D surrogate model for elastic fields in polycrystals,” Comput Methods Appl Mech Eng, vol. 441, Jun. 2025, doi: 10.1016/j.cma.2025.117944

    2025

  14. [22]

    Data-driven elastoplastic constitutive modelling with physics -informed RNNs using the Virtual Fields Method for indirect training,

    R. Lourenço, P. Georgieva, and A. Andrade -Campos, “Data-driven elastoplastic constitutive modelling with physics -informed RNNs using the Virtual Fields Method for indirect training,” Comput Methods Appl Mech Eng, vol. 440, May 2025, doi: 10.1016/j.cma.2025.117935

    2025

  15. [23]

    The deep finite element method: A deep learning framework integrating the physics -informed neural networks with the finite element method,

    W. Xiong, X. Long, S. P. A. Bordas, and C. Jiang, “The deep finite element method: A deep learning framework integrating the physics -informed neural networks with the finite element method,” Comput Methods Appl Mech Eng, vol. 436, Mar. 2025, doi: 10.1016/j.cma.2024.117681

    2025

  16. [24]

    Extended physics -informed extreme learning machine for linear elastic fracture mechanics,

    B. Zhu, H. Li, and Q. Zhang, “Extended physics -informed extreme learning machine for linear elastic fracture mechanics,” Comput Methods Appl Mech Eng, vol. 435, Feb. 2025, doi: 10.1016/j.cma.2024.117655

    2025

  17. [25]

    A physics-informed neural network-based method for dispersion calculations,

    Z. Cheng, T. Yu, G. Jia, and Z. Shi, “A physics-informed neural network-based method for dispersion calculations,” Int J Mech Sci , vol. 291 –292, p. 110111, Apr. 2025, doi: 10.1016/j.ijmecsci.2025.110111

    2025

  18. [26]

    Viscohyperelastic Strain Energy Function,

    A. V ogel, L. Rakotomanana, and D. P. Pioletti, “Viscohyperelastic Strain Energy Function,” in Biomechanics of Living Organs: Hyperelastic Constitutive Laws for Finite Element Modeling , Elsevier Inc., 2017, pp. 59–78. doi: 10.1016/B978-0-12-804009-6.00003-1

    2017 doi

  19. [27]

    Inverse Physics-Informed Neural Networks for transport models in porous materials,

    M. Berardi, F. Difonzo, and M. Icardi, “Inverse Physics-Informed Neural Networks for transport models in porous materials,” Jul. 2024, doi: 10.1016/j.cma.2024.117628

    2024

  20. [30]

    Viscoelastic constitutive law in large deformations: application to human knee ligaments and tendons,

    D. P. Pioletti, L. R. Rakotomanana, J.-F. Benvenuti, and P. -F. Leyvraz, “Viscoelastic constitutive law in large deformations: application to human knee ligaments and tendons,” 1998. 38

    1998

  21. [31]

    Analytical and experimental study of shape memory alloy reinforcement on the performance of butt -fusion welded joints in high-density polyethylene pipe,

    E. Choi, A. Ostadrahimi, and H. Youn, “Analytical and experimental study of shape memory alloy reinforcement on the performance of butt -fusion welded joints in high-density polyethylene pipe,” J Intell Mater Syst Struct, vol. 31, no. 17, 2020, doi: 10.1177/1045389X20942583

    2020 doi

  22. [32]

    A Review of Recent Advances in Surrogate Models for Uncertainty Quantification of High-Dimensional Engineering Applications,

    Z. Azarhoosh and M. Ilchi Ghazaan, “A Review of Recent Advances in Surrogate Models for Uncertainty Quantification of High-Dimensional Engineering Applications,” Jan. 01, 2025, Elsevier B.V . doi: 10.1016/j.cma.2024.117508

    2025

  23. [33]

    History-Matching of Imbibition Flow in Multiscale Fractured Porous Media Using Physics-Informed Neural Networks (PINNs),

    J. Abbasi et al., “History-Matching of Imbibition Flow in Multiscale Fractured Porous Media Using Physics-Informed Neural Networks (PINNs),” Oct. 2024, doi: 10.1016/j.cma.2025.117784

    2024

  24. [34]

    Development and validation of subject-specific 3D human head models based on a nonlinear visco-hyperelastic constitutive framework,

    K. Upadhyay et al., “Development and validation of subject-specific 3D human head models based on a nonlinear visco-hyperelastic constitutive framework,” J R Soc Interface, vol. 19, no. 195, Oct. 2022, doi: 10.1098/rsif.2022.0561

    2022

  25. [35]

    A transversely isotropic viscohyperelastic material Application to the modeling of biological soft connective tissues,

    G. Limbert and J. Middleton, “A transversely isotropic viscohyperelastic material Application to the modeling of biological soft connective tissues,” Int J Solids Struct , vol. 41, no. 15, pp. 4237–4260, Jul. 2004, doi: 10.1016/j.ijsolstr.2004.02.057

    2004 doi

  26. [36]

    A Nonlinear Thermo -Visco- Green-Elastic Constitutive Model for Mullins Damage of Shape Memory Polymers under Giant Elongations,

    A. Ostadrahimi, A. Enferadi, M. Baghani, S. Sarrafan, and G. Li, “A Nonlinear Thermo -Visco- Green-Elastic Constitutive Model for Mullins Damage of Shape Memory Polymers under Giant Elongations,” J Mech Phys Solids, vol. 196, Mar. 2025, doi: 10.1016/j.jmps.2025.106029

    2025

  27. [37]

    Deep learning predicts path -dependent plasticity,

    M. A. B. M. Designed Research; M, “Deep learning predicts path -dependent plasticity,” vol. 116, no. 52, 2019, doi: 10.1073/pnas.1911815116/-/DCSupplemental.y

    2019 doi

  28. [38]

    THESE DE DOCTORAT DE L’ÉCOLE CENTRALE DE NANTES Auriane Platzer Rapporteurs avant soutenance : Composition du Jury

    E. N. Doctorale, J. Yvonnet, D. Ryckelynck Professeur, M. Paris Tech Examinateurs, S. Reese Professeure, and R. Aachen Université Piotr Breitkopf, “THESE DE DOCTORAT DE L’ÉCOLE CENTRALE DE NANTES Auriane Platzer Rapporteurs avant soutenance : Composition du Jury.”

  29. [39]

    An analytical study on SMA beam-column actuators for anti -buckling phenomenon,

    A. Ostadrahimi, F. Taheri-Behrooz, E. Choi, and G. Li, “An analytical study on SMA beam-column actuators for anti -buckling phenomenon,” Int J Solids Struct , vol. 295, p. 112792, Jun. 2024, doi: 10.1016/j.ijsolstr.2024.112792

    2024

  30. [40]

    ‘A constitutive model for the Mullins effect with permanent set in a particle -reinforced rubber’ by A. Dorfmann and R.W. Ogden,

    A. Zhong, “‘A constitutive model for the Mullins effect with permanent set in a particle -reinforced rubber’ by A. Dorfmann and R.W. Ogden,” 2005. doi: 10.1016/j.ijsolstr.2004.12.002

    2005 doi

  31. [41]

    Finite Bending of Fiber -Reinforced Visco-Hyperelastic Material: Analytical Approach and FEM,

    J. Pashazadeh, A. Ostadrahimi, M. Baghani, and E. Choi, “Finite Bending of Fiber -Reinforced Visco-Hyperelastic Material: Analytical Approach and FEM,” Materials, vol. 17, no. 1, Jan. 2024, doi: 10.3390/ma17010005

    2024 doi

  32. [42]

    A viscoelastic constitutive model for compressible polymers based on logarithmic strain and its finite element implementation,

    R. Naghdabadi, M. Baghani, and J. Arghavani, “A viscoelastic constitutive model for compressible polymers based on logarithmic strain and its finite element implementation,” Finite Elements in Analysis and Design, vol. 62, pp. 18–27, Dec. 2012, doi: 10.1016/j.finel.2012.05.001

    2012 doi

  33. [43]

    An analytical study on the bending of prismatic SMA beams,

    A. Ostadrahimi, J. Arghavani, and S. Poorasadion, “An analytical study on the bending of prismatic SMA beams,” Smart Mater Struct, vol. 24, no. 12, 2015, doi: 10.1088/0964-1726/24/12/125035

    2015 doi

  34. [44]

    Mechanics -informed, model -free symbolic regression framework for solving fracture problems,

    R. Yi, D. Georgiou, X. Liu, and C. E. Athanasiou, “Mechanics -informed, model -free symbolic regression framework for solving fracture problems,” J Mech Phys Solids, vol. 194, Jan. 2025, doi: 10.1016/j.jmps.2024.105916. 39

    2025

  35. [45]

    Data -driven continuum damage mechanics with built -in physics,

    V . Tac, E. Kuhl, and A. B. Tepole, “Data -driven continuum damage mechanics with built -in physics,” Extreme Mech Lett, vol. 71, Sep. 2024, doi: 10.1016/j.eml.2024.102220

    2024

  36. [46]

    A neural network -based enrichment of reproducing kernel approximation for modeling brittle fracture,

    J. Baek and J. S. Chen, “A neural network -based enrichment of reproducing kernel approximation for modeling brittle fracture,” Comput Methods Appl Mech Eng, vol. 419, Feb. 2024, doi: 10.1016/j.cma.2023.116590

    2024

  37. [47]

    Tensor Basis Gaussian Process Models of Hyperelastic Materials,

    A. Frankel, R. Jones, and L. Swiler, “Tensor Basis Gaussian Process Models of Hyperelastic Materials,” Dec. 2019, [Online]. Available: http://arxiv.org/abs/1912.10872

    2019 arXiv

  38. [48]

    Local approximate Gaussian process regression for data- driven constitutive models: development and comparison with neural networks,

    J. N. Fuhg, M. Marino, and N. Bouklas, “Local approximate Gaussian process regression for data- driven constitutive models: development and comparison with neural networks,” Comput Methods Appl Mech Eng, vol. 388, Jan. 2022, doi: 10.1016/j.cma.2021.114217

    2022

  39. [49]

    Data-driven hyperelasticity, Part I: A canonical isotropic formulation for rubberlike materials,

    H. Dal, F. A. Denli, A. K. Açan, and M. Kaliske, “Data-driven hyperelasticity, Part I: A canonical isotropic formulation for rubberlike materials,” J Mech Phys Solids , vol. 179, Oct. 2023, doi: 10.1016/j.jmps.2023.105381

    2023

  40. [50]

    On physics-informed data-driven isotropic and anisotropic constitutive models through probabilistic machine learning and space -filling sampling,

    J. N. Fuhg and N. Bouklas, “On physics-informed data-driven isotropic and anisotropic constitutive models through probabilistic machine learning and space -filling sampling,” Comput Methods Appl Mech Eng, vol. 394, May 2022, doi: 10.1016/j.cma.2022.114915

    2022

  41. [51]

    Intriguing properties of neural networks,

    C. Szegedy et al., “Intriguing properties of neural networks,” Dec. 2013, [Online]. Available: http://arxiv.org/abs/1312.6199

    2013 arXiv

  42. [52]

    Machine learning of evolving physics - based material models for multiscale solid mechanics,

    I. B. C. M. Rocha, P. Kerfriden, and F. P. van der Meer, “Machine learning of evolving physics - based material models for multiscale solid mechanics,” Mechanics of Materials, vol. 184, Sep. 2023, doi: 10.1016/j.mechmat.2023.104707

    2023

  43. [53]

    Prediction of damage evolution in CMCs considering the real microstructures through a deep-learning scheme,

    R. Zhu, G. Niu, P. Wang, C. He, Z. Qu, and D. Fang, “Prediction of damage evolution in CMCs considering the real microstructures through a deep-learning scheme,” Comput Methods Appl Mech Eng, vol. 439, May 2025, doi: 10.1016/j.cma.2025.117923

    2025

  44. [54]

    t -PiNet: A thermodynamics -informed hierarchical learning for discovering constitutive relations of geomaterials,

    P. Zhang, K. Karapiperis, and O. Weeger, “t -PiNet: A thermodynamics -informed hierarchical learning for discovering constitutive relations of geomaterials,” J Mech Phys Solids , vol. 197, Apr. 2025, doi: 10.1016/j.jmps.2025.106049

    2025

  45. [55]

    A new family of Constitutive Artificial Neural Networks towards automated model discovery,

    K. Linka and E. Kuhl, “A new family of Constitutive Artificial Neural Networks towards automated model discovery,” Comput Methods Appl Mech Eng, vol. 403, Jan. 2023, doi: 10.1016/j.cma.2022.115731

    2023

  46. [56]

    Physics-informed data-driven discovery of constitutive models with application to strain-rate-sensitive soft materials,

    K. Upadhyay, J. N. Fuhg, N. Bouklas, and K. T. Ramesh, “Physics-informed data-driven discovery of constitutive models with application to strain-rate-sensitive soft materials,” Comput Mech, 2024, doi: 10.1007/s00466-024-02497-x

    2024 doi

  47. [57]

    Thermodynamics-based Artificial Neural Networks for constitutive modeling,

    F. Masi, I. Stefanou, P. Vannucci, and V . Maffi-Berthier, “Thermodynamics-based Artificial Neural Networks for constitutive modeling,” J Mech Phys Solids , vol. 147, Feb. 2021, doi: 10.1016/j.jmps.2020.104277

    2021

  48. [58]

    Evolution TANN and the identification of internal variables and evolution equations in solid mechanics,

    F. Masi and I. Stefanou, “Evolution TANN and the identification of internal variables and evolution equations in solid mechanics,” J Mech Phys Solids , vol. 174, May 2023, doi: 10.1016/j.jmps.2023.105245. 40

    2023

  49. [60]

    A review on data-driven constitutive laws for solids,

    J. N. Fuhg et al., “A review on data-driven constitutive laws for solids,” May 2024, [Online]. Available: http://arxiv.org/abs/2405.03658

    2024 arXiv

  50. [61]

    Parametrized polyconvex hyperelasticity with physics-augmented neural networks,

    D. K. Klein, F. J. Roth, I. Valizadeh, and O. Weeger, “Parametrized polyconvex hyperelasticity with physics-augmented neural networks,” Data-Centric Engineering , vol. 4, no. 2, Nov. 2023, doi: 10.1017/dce.2023.21

    2023 doi

  51. [62]

    Finite electro-elasticity with physics- augmented neural networks,

    D. K. Klein, R. Ortigosa, J. Martínez-Frutos, and O. Weeger, “Finite electro-elasticity with physics- augmented neural networks,” Comput Methods Appl Mech Eng, vol. 400, Oct. 2022, doi: 10.1016/j.cma.2022.115501

    2022

  52. [63]

    Input Convex Neural Networks,

    B. Amos, L. Xu, and J. Zico Kolter, “Input Convex Neural Networks,” 2017

    2017

  53. [64]

    A mechanics-informed deep learning framework for data-driven nonlinear viscoelasticity,

    F. As’ad and C. Farhat, “A mechanics-informed deep learning framework for data-driven nonlinear viscoelasticity,” Comput Methods Appl Mech Eng, vol. 417, Dec. 2023, doi: 10.1016/j.cma.2023.116463

    2023

  54. [65]

    A mechanics-informed artificial neural network approach in data- driven constitutive modeling,

    F. As’ad, P. Avery, and C. Farhat, “A mechanics-informed artificial neural network approach in data- driven constitutive modeling,” Int J Numer Methods Eng , vol. 123, no. 12, pp. 2738– 2759, Jun. 2022, doi: 10.1002/nme.6957

    2022 doi

  55. [66]

    Hypersparse Neural Network Analysis of Large-Scale Internet Traffic,

    J. Kepner, K. Cho, K. Claffy, V . Gadepally, P. Michaleas, and L. Milechin, “Hypersparse Neural Network Analysis of Large-Scale Internet Traffic,” Apr. 2019, doi: 10.1109/HPEC.2019.8916263

    2019

  56. [67]

    Data-Driven Network Neuroscience: On Data Collection and Benchmark

    J. Xu et al., “Data-Driven Network Neuroscience: On Data Collection and Benchmark”, doi: 10.17608/k6.auckland.21397377

    doi

  57. [68]

    Patient Risk Assessment and Warning Symptom Detection Using Deep Attention- Based Neural Networks,

    I. Girardi et al., “Patient Risk Assessment and Warning Symptom Detection Using Deep Attention- Based Neural Networks,” Sep. 2018, [Online]. Available: http://arxiv.org/abs/1809.10804

    2018 arXiv

  58. [69]

    Predicting Progression Events in Multiple Myeloma from Routine Blood Work,

    M. Ferle et al., “Predicting Progression Events in Multiple Myeloma from Routine Blood Work,” May 2024, [Online]. Available: http://arxiv.org/abs/2405.18051

    2024 arXiv

  59. [70]

    A machine learning approach to predict in vivo skin growth,

    M. Nagle, H. Conroy Broderick, A. Buganza Tepole, M. Fop, and A. Ní Annaidh, “A machine learning approach to predict in vivo skin growth,” Sci Rep , vol. 14, no. 1, Dec. 2024, doi: 10.1038/s41598-024-67056-z

    2024 doi

  60. [71]

    Discovering plasticity models without stress data,

    M. Flaschel, S. Kumar, and L. De Lorenzis, “Discovering plasticity models without stress data,” NPJ Comput Mater, vol. 8, no. 1, Dec. 2022, doi: 10.1038/s41524-022-00752-4

    2022 doi

  61. [72]

    Bayesian -EUCLID: Discovering hyperelastic material laws with uncertainties,

    A. Joshi et al., “Bayesian -EUCLID: Discovering hyperelastic material laws with uncertainties,” Comput Methods Appl Mech Eng, vol. 398, Aug. 2022, doi: 10.1016/j.cma.2022.115225

    2022

  62. [73]

    Automated identification of linear viscoelastic constitutive laws with EUCLID,

    E. Marino, M. Flaschel, S. Kumar, and L. De Lorenzis, “Automated identification of linear viscoelastic constitutive laws with EUCLID,” Mechanics of Materials , vol. 181, Jun. 2023, doi: 10.1016/j.mechmat.2023.104643. 41

    2023

  63. [74]

    Learning viscoelasticity models from indirect data using deep neural networks,

    K. Xu, A. M. Tartakovsky, J. Burghardt, and E. Darve, “Learning viscoelasticity models from indirect data using deep neural networks,” Comput Methods Appl Mech Eng, vol. 387, Dec. 2021, doi: 10.1016/j.cma.2021.114124

    2021

  64. [76]

    Hybrid Monte Carlo for Failure Probability Estimation with Gaussian Process Surrogates,

    A. S. Booth and S. A. Renganathan, “Hybrid Monte Carlo for Failure Probability Estimation with Gaussian Process Surrogates,” 2024

    2024

  65. [77]

    Viscoelastic constitutive artificial neural networks (vCANNs) – A framework for data -driven anisotropic nonlinear finite viscoelasticity,

    K. P. Abdolazizi, K. Linka, and C. J. Cyron, “Viscoelastic constitutive artificial neural networks (vCANNs) – A framework for data -driven anisotropic nonlinear finite viscoelasticity,” J Comput Phys, vol. 499, Feb. 2024, doi: 10.1016/j.jcp.2023.112704

    2024

  66. [78]

    Constitutive artificial neural networks: A fast and general approach to predictive data-driven constitutive modeling by deep learning,

    K. Linka, M. Hillgärtner, K. P. Abdolazizi, R. C. Aydin, M. Itskov, and C. J. Cyron, “Constitutive artificial neural networks: A fast and general approach to predictive data-driven constitutive modeling by deep learning,” J Comput Phys, vol. 429, Mar. 2021, doi: 10.1016/j.jcp....

    2021

  67. [79]

    Discovering uncertainty: Bayesian constitutive artificial neural networks,

    K. Linka, G. A. Holzapfel, and E. Kuhl, “Discovering uncertainty: Bayesian constitutive artificial neural networks,” Comput Methods Appl Mech Eng, vol. 433, p. 117517, Jan. 2025, doi: 10.1016/j.cma.2024.117517

    2025

  68. [81]

    ViscoelasticNet: A physics informed neural network framework for stress discovery and model selection,

    S. Thakur, M. Raissi, and A. M. Ardekani, “ViscoelasticNet: A physics informed neural network framework for stress discovery and model selection,” J Nonnewton Fluid Mech, vol. 330, Aug. 2024, doi: 10.1016/j.jnnfm.2024.105265

    2024

  69. [82]

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

    M. Raissi, P. Perdikaris, and G. E. Karniadakis, “Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations,” J Comput Phys, vol. 378, pp. 686–707, Feb. 2019, doi: 10.1016/j.jcp.2018.10.045

    2019 doi

  70. [83]

    Model -free Data-Driven viscoelasticity in the frequency domain,

    H. Salahshoor and M. Ortiz, “Model -free Data-Driven viscoelasticity in the frequency domain,” Comput Methods Appl Mech Eng, vol. 403, Jan. 2023, doi: 10.1016/j.cma.2022.115657

    2023

  71. [84]

    Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning,

    Y . Gal and Z. Ghahramani, “Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning,” Jun. 2015, [Online]. Available: http://arxiv.org/abs/1506.02142

    2015 arXiv

  72. [85]

    Data-driven physics-informed constitutive metamodeling of complex fluids: A multifidelity neural network (MFNN) framework,

    M. Mahmoudabadbozchelou, M. Caggioni, S. Shahsavari, W. H. Hartt, G. Em Karniadakis, and S. Jamali, “Data-driven physics-informed constitutive metamodeling of complex fluids: A multifidelity neural network (MFNN) framework,” J Rheol (N Y N Y) , vol. 65, no. 2, pp. 179– 198, Ma...

    2021 doi

  73. [86]

    Weight Uncertainty in Neural Networks,

    C. Blundell, J. Cornebise, K. Kavukcuoglu, and D. Wierstra, “Weight Uncertainty in Neural Networks,” May 2015, [Online]. Available: http://arxiv.org/abs/1505.05424

    2015 arXiv

  74. [87]

    What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?

    A. Kendall and Y . Gal, “What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?” 42

  75. [88]

    A generalized dual potential for inelastic Constitutive Artificial Neural Networks: A JAX implementation at finite strains,

    H. Holthusen, K. Linka, E. Kuhl, and T. Brepols, “A generalized dual potential for inelastic Constitutive Artificial Neural Networks: A JAX implementation at finite strains,” Feb. 2025, [Online]. Available: http://arxiv.org/abs/2502.17490

    2025

  76. [89]

    Bayesian Physics Informed Neural Networks for real -world nonlinear dynamical systems,

    K. Linka, A. Schäfer, X. Meng, Z. Zou, G. E. Karniadakis, and E. Kuhl, “Bayesian Physics Informed Neural Networks for real -world nonlinear dynamical systems,” Comput Methods Appl Mech Eng , vol. 402, Dec. 2022, doi: 10.1016/j.cma.2022.115346

    2022

  77. [90]

    Empowering approximate Bayesian neural networks with functional priors through anchored ensembling for mechanics surrogate modeling applications,

    J. Ghorbanian, N. Casaprima, and A. Olivier, “Empowering approximate Bayesian neural networks with functional priors through anchored ensembling for mechanics surrogate modeling applications,” Comput Methods Appl Mech Eng, Apr. 2025, doi: 10.1016/j.cma.2024.117645

    2025

  78. [91]

    Uncertainty quantification for noisy inputs -outputs in physics-informed neural networks and neural operators,

    Z. Zou, X. Meng, and G. E. Karniadakis, “Uncertainty quantification for noisy inputs -outputs in physics-informed neural networks and neural operators,” Nov. 2023, doi: 10.1016/j.cma.2024.117479

    2023

  79. [92]

    FE ANN : an efficient data -driven multiscale approach based on physics-constrained neural networks and automated data mining,

    K. A. Kalina, L. Linden, J. Brummund, and M. Kästner, “FE ANN : an efficient data -driven multiscale approach based on physics-constrained neural networks and automated data mining,” Comput Mech, vol. 71, no. 5, pp. 827–851, May 2023, doi: 10.1007/s00466-022-02260-0

    2023 doi

  80. [93]

    Viscoelasticty with physics- augmented neural networks: Model formulation and training methods without prescribed internal variables,

    M. Rosenkranz, K. A. Kalina, J. Brummund, W. Sun, and M. Kästner, “Viscoelasticty with physics- augmented neural networks: Model formulation and training methods without prescribed internal variables,” Jan. 2024, [Online]. Available: http://arxiv.org/abs/2401.14270

    2024 arXiv

  81. [94]

    Neural network - based multiscale modeling of finite strain magneto -elasticity with relaxed convexity criteria,

    K. A. Kalina, P. Gebhart, J. Brummund, L. Linden, W. C. Sun, and M. Kästner, “Neural network - based multiscale modeling of finite strain magneto -elasticity with relaxed convexity criteria,” Comput Methods Appl Mech Eng, vol. 421, Mar. 2024, doi: 10.1016/j.cma.2023.116739

    2024

  82. [95]

    Data-oriented constitutive modeling of plasticity in metals,

    A. Hartmaier, “Data-oriented constitutive modeling of plasticity in metals,” Materials, vol. 13, no. 7, Apr. 2020, doi: 10.3390/ma13071600

    2020 doi

  83. [96]

    Application of artificial neural networks for the prediction of interface mechanics: a study on grain boundary constitutive behavior,

    M. Fernández, S. Rezaei, J. Rezaei Mianroodi, F. Fritzen, and S. Reese, “Application of artificial neural networks for the prediction of interface mechanics: a study on grain boundary constitutive behavior,” Adv Model Simul Eng Sci, vol. 7, no. 1, Dec. 2020, doi: 10.1186/s4032...

    2020 doi

  84. [97]

    Multiscale Modeling Meets Machine Learning: What Can We Learn?,

    G. C. Y . Peng et al. , “Multiscale Modeling Meets Machine Learning: What Can We Learn?,” Archives of Computational Methods in Engineering, vol. 28, no. 3, pp. 1017–1037, May 2021, doi: 10.1007/s11831-020-09405-5

    2021 doi

  85. [98]

    Integrating machine learning and multiscale modeling—perspectives, challenges, and opportunities in the biological, biomedical, and behavioral sciences,

    M. Alber et al., “Integrating machine learning and multiscale modeling—perspectives, challenges, and opportunities in the biological, biomedical, and behavioral sciences,” Dec. 01, 2019, Nature Research. doi: 10.1038/s41746-019-0193-y

    2019 doi

  86. [99]

    Hernández, A

    Q. Hernández, A. Badías, F. Chinesta, and E. Cueto, “Thermodynamics-informed neural networks for physically realistic mixed reality ScienceDirect Thermodynamics-informed neural networks for physically realistic mixed reality ✩ -NC-ND license (http://creativecommons.org/license...

    2023

  87. [100]

    Data-driven computational mechanics,

    T. Kirchdoerfer and M. Ortiz, “Data-driven computational mechanics,” Comput Methods Appl Mech Eng, vol. 304, pp. 81–101, Jun. 2016, doi: 10.1016/j.cma.2016.02.001. 43

    2016 doi

  88. [101]

    Model-Free Data-Driven inelasticity,

    R. Eggersmann, T. Kirchdoerfer, S. Reese, L. Stainier, and M. Ortiz, “Model-Free Data-Driven inelasticity,” Comput Methods Appl Mech Eng, vol. 350, pp. 81– 99, Jun. 2019, doi: 10.1016/j.cma.2019.02.016

    2019 doi

  89. [102]

    On sparse regression, Lp -regularization, and automated model discovery,

    J. A. McCulloch, S. R. St. Pierre, K. Linka, and E. Kuhl, “On sparse regression, Lp -regularization, and automated model discovery,” Jul. 30, 2024, John Wiley and Sons Ltd. doi: 10.1002/nme.7481

    2024 doi

  90. [103]

    Keeton, Proceedings of the 12th USENIX conference on Operating Systems Design and Implementation

    Kimberly. Keeton, Proceedings of the 12th USENIX conference on Operating Systems Design and Implementation. USENIX Association, 2016

    2016

  91. [104]

    A POD -TANN approach for the multiscale modeling of materials and macroelement derivation in geomechanics,

    G. Piunno, I. Stefanou, and C. Jommi, “A POD -TANN approach for the multiscale modeling of materials and macroelement derivation in geomechanics,” Aug. 2024, [Online]. Available: http://arxiv.org/abs/2408.07165

    2024 arXiv

  92. [105]

    Tunnel lining defects identification using TPE-CatBoost algorithm with GPR data: A model test study,

    K. Li, X. Xie, J. Zhai, B. Zhou, C. Huang, and C. Wang, “Tunnel lining defects identification using TPE-CatBoost algorithm with GPR data: A model test study,” Tunnelling and Underground Space Technology, vol. 157, Mar. 2025, doi: 10.1016/j.tust.2024.106275

    2025

  93. [106]

    Physics-informed Data-driven Discovery of Constitutive Models with Application to Strain -Rate-sensitive Soft Materials,

    K. Upadhyay, J. N. Fuhg, N. Bouklas, and K. T. Ramesh, “Physics-informed Data-driven Discovery of Constitutive Models with Application to Strain -Rate-sensitive Soft Materials,” Apr. 2023, [Online]. Available: http://arxiv.org/abs/2304.13897

    2023 arXiv

  94. [108]

    Shear strength of beam -end bolted connections in cold-formed steel structures through experiments, numerical simulations and hybrid GPR-ECLPSO modeling,

    V . T. Huynh, C. H. Pham, V . B. Pham, and H. T. Thai, “Shear strength of beam -end bolted connections in cold-formed steel structures through experiments, numerical simulations and hybrid GPR-ECLPSO modeling,” Thin -Walled Structures , vol. 211, Jun. 2025, doi: 10.1016/j.tws....

    2025

  95. [109]

    Intelligent prediction of compressive strength of self -compacting concrete incorporating silica fume using hybrid IWOA -GPR model,

    Y . Yu, G. Wang, G. F. Huseien, Z. Zou, Z. Ding, and C. Zhang, “Intelligent prediction of compressive strength of self -compacting concrete incorporating silica fume using hybrid IWOA -GPR model,” Mater Today Commun, vol. 45, Apr. 2025, doi: 10.1016/j.mtcomm.2025.112282

    2025

  96. [111]

    Recurrent neural networks (RNNs) learn the constitutive law of viscoelasticity,

    G. Chen, “Recurrent neural networks (RNNs) learn the constitutive law of viscoelasticity,” Comput Mech, vol. 67, no. 3, pp. 1009–1019, Mar. 2021, doi: 10.1007/s00466-021-01981-y

    2021 doi

  97. [112]

    A recurrent neural network-accelerated multi- scale model for elasto -plastic heterogeneous materials subjected to random cyclic and non - proportional loading paths,

    L. Wu, V . D. Nguyen, N. G. Kilingar, and L. Noels, “A recurrent neural network-accelerated multi- scale model for elasto -plastic heterogeneous materials subjected to random cyclic and non - proportional loading paths,” Comput Methods Appl Mech Eng, vol. 369, Sep. 2020, doi: ...

    2020

  98. [113]

    Finite element solver for data -driven finite strain elasticity,

    A. Platzer, A. Leygue, L. Stainier, and M. Ortiz, “Finite element solver for data -driven finite strain elasticity,” Comput Methods Appl Mech Eng, vol. 379, Jun. 2021, doi: 10.1016/j.cma.2021.113756

    2021

  99. [114]

    A review of the application of machine learning and data mining approaches in continuum materials mechanics,

    F. E. Bock, R. C. Aydin, C. J. Cyron, N. Huber, S. R. Kalidindi, and B. Klusemann, “A review of the application of machine learning and data mining approaches in continuum materials mechanics,” Apr. 15, 2019, Frontiers Media S.A. doi: 10.3389/fmats.2019.00110. 44

    2019

  100. [115]

    Data-driven and physics-based modelling of process behaviour and deposit geometry for friction surfacing,

    F. E. Bock, Z. Kallien, N. Huber, and B. Klusemann, “Data-driven and physics-based modelling of process behaviour and deposit geometry for friction surfacing,” 2023, doi: 10.5281/zenodo.8263026

    2023 doi

  101. [116]

    Modeling of materials with fading memory using neural networks,

    M. Oeser and S. Freitag, “Modeling of materials with fading memory using neural networks,” Int J Numer Methods Eng, vol. 78, no. 7, pp. 843–862, May 2009, doi: 10.1002/nme.2518

    2009 doi

  102. [117]

    A material description based on recurrent neural networks for fuzzy data and its application within the finite element method,

    S. Freitag, W. Graf, and M. Kaliske, “A material description based on recurrent neural networks for fuzzy data and its application within the finite element method,” Comput Struct , vol. 124, pp. 29– 37, 2013, doi: 10.1016/j.compstruc.2012.11.011

    2013 doi

  103. [118]

    Numerical characterisation of uncured elastomers by a neural network based approach,

    C. Zopf and M. Kaliske, “Numerical characterisation of uncured elastomers by a neural network based approach,” Comput Struct , vol. 182, pp. 504– 525, Apr. 2017, doi: 10.1016/j.compstruc.2016.12.012

    2017 doi

  104. [119]

    Neural network model describing the temperature- And rate- dependent stress-strain response of polypropylene,

    B. Jordan, M. B. Gorji, and D. Mohr, “Neural network model describing the temperature- And rate- dependent stress-strain response of polypropylene,” Int J Plast, vol. 135, Dec. 2020, doi: 10.1016/j.ijplas.2020.102811

    2020

  105. [120]

    A new viscoelastic constitutive model for continuous media at finite thermomechanical changes,

    G. A. Holzapfel and J. C. Simo, “A new viscoelastic constitutive model for continuous media at finite thermomechanical changes,” Int J Solids Struct, vol. 33, no. 20–22, 1996, doi: 10.1016/0020- 7683(95)00263-4

    1996 doi

  106. [121]

    Spatial variation. Stochastic models and their application to some problems in forest surveys and other sampling investigations.,

    B. Matern, “Spatial variation. Stochastic models and their application to some problems in forest surveys and other sampling investigations.,” vol. 49, no. 5, p. 144, 1960

    1960

  107. [122]

    Gaussian Processes in Machine Learning

    C. E. Rasmussen, “Gaussian Processes in Machine Learning.” [Online]. Available: http://www.tuebingen.mpg.de/∼carl

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.