Pith. sign in

REVIEW 4 major objections 6 minor 40 references

Recurrent U-Net-Based Graph Neural Network (RUGNN) for Accurate Deformation Predictions in Sheet Material Forming

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 RUGNN, a recurrent U-Net graph neural network, predicts sheet-metal deformation fields across stamping timesteps with lower accumulated positional error than three GNN baselines.

desk verdict RUGNN is a plausible and useful combination of known GNN mechanisms, but the headline accuracy ranking needs repeated-seed evidence before it can be trusted. read the letter →

arxiv 2507.11547 v1 pith:PMZTNFTY submitted 2025-07-10 cs.LG

Yingxue Zhao , Qianyi Chen , Haoran Li , Haosu Zhou , Hamid Reza Attar , Tobias Pfaff , Tailin Wu , Nan Li This is my paper
classification cs.LG
keywords graphneuralnetworkssheetmetalformingsurrogatemodelsGatedRecurrentUnitsU-Netcontactmodellingautoregressiverollout
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

This paper proposes RUGNN, a graph neural network surrogate for sheet-metal stamping that predicts how a blank deforms over the full sequence of forming timesteps. The central claim is that combining Gated Recurrent Units inside message-passing blocks with a U-Net-style graph downsample/upsample hierarchy gives lower positional error at every timestep and less autoregressive error drift than plain message passing, recurrent-only, or U-Net-only GNN baselines. This matters because finite-element stamping simulations are slow, so a learned surrogate that stays accurate over many rollout steps could let designers test tool geometries before committing to full simulation. The paper supports the claim with two aluminium case studies, cold dome stamping and hot bulkhead stamping, and with a cheap node-to-surface contact encoding for large-scale tool-contact interactions.

What carries the argument

The load-bearing pieces are: (1) the Recurrent-based GNN Processor Block (RGPB), where each message-passing layer is an edge-MLP update followed by a GRU on the edge hidden state and a node-MLP aggregation, so temporal memory is stored on edges and carried across timesteps; (2) a graph downsample/upsample U-Net built from coarser meshes and KD-Tree inter-level edges, with skip-concatenation during upsampling, so most message passing happens on a very coarse graph; and (3) a node-to-surface contact feature that gives each blank node the inverse of its shortest distance to the tool surface, $1/d_n^t$, and the normal vector of the closest tool element, avoiding the construction of expensive contact edges.

What would settle it

A decisive test would be to train RUGNN and a baseline using explicit node-to-node or face-to-face contact edges on a forming scenario where the blank touches two tool surfaces at once or slides over a sharp tool corner; if RUGNN's positional error in those contact regions is much larger than the baseline's, the node-to-surface encoding is dropping interaction details the paper claims are preserved.

Watch

Extended reading notes

Core claim

RUGNN is an encoder-processor-decoder GNN that predicts the displacement vector $\Delta x_n^{t+1}$ for every node at each forming step. The processor uses three Recurrent GNN Processor Blocks: one at the finest level before downsampling, one at the coarsest level after all downsampling, and one at the finest level after upsampling. Each block updates edge features, passes them through a GRU whose hidden edge state is carried from the previous timestep, and aggregates the result into node updates. This recurrence is what the paper credits for reducing accumulated autoregressive error, while the manually coarsened graph hierarchy, mapped with KD-Tree inter-level edges, is what makes long-range message passing affordable on graphs as large as 6,700 nodes. On the dome cold-stamping and bulkhead hot-stamping case studies, the paper reports that RUGNN achieves the lowest positional mean Euclidean error across all timesteps and reduces error accumulation compared with vanillaGNN, RGNN, and UGNN, with predictions closely matching finite-element ground truth.

Load-bearing premise

The load-bearing premise is that encoding contact as only the inverse shortest distance to the tool surface and the closest-surface normal retains enough interaction detail to predict deformation accurately, a premise the paper supports mainly by distance-value agreement with FE results rather than by direct validation of transmitted forces.

Editorial extensions

If this is right

  • On both the dome cold-stamping and bulkhead hot-stamping datasets, RUGNN produces the lowest positional mean Euclidean error at every timestep and slows the error accumulation seen in autoregressive rollout.
  • Recurrence and graph U-Net downsampling are complementary: the U-Net handles long-range spatial propagation while the GRU hidden edge states suppress drift over time, as shown by the bulkhead errors concentrating in the top region when recurrence is absent.
  • The proposed contact features are load-bearing: removing them raises the dome-case validation MSE from 0.0791 to 0.0974, so the cheap node-to-surface encoding carries much of the contact signal.
  • With a node reduction ratio near 3 and most message-passing layers placed at the coarsest level, RUGNN keeps GPU memory close to the U-Net-only model while improving accuracy, making 6,700-node stamping graphs feasible on one GPU.
  • The model generalizes to unseen tool geometries drawn from the same design space, since every test sample uses a tool geometry not seen in training.

Reading between the lines

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

  • If the trend holds, the same combination of recurrent edge states and multiscale graph processing could transfer to other large-contact deformation problems such as forging, crash simulation, or composite forming, though the paper only demonstrates stamping of two aluminium shapes.
  • The node-to-surface contact encoding discards information about which tool surface is contacted and about sliding; a natural test is whether adding contact pressure or a signed-distance gradient improves accuracy on cases where the blank contacts two tool surfaces at once.
  • Because teacher forcing trained better than autoregressive training in the paper's comparison, a scheduled noising or curriculum that gradually shifts from ground-truth to predicted inputs might further reduce rollout drift; this was not explored.
  • The hierarchy relies on fixed manually coarsened meshes for each case study, so learning the coarsening or adapting it automatically to new tool shapes would be the natural next step toward a general tool-design exploration tool.
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 RUGNN, a graph neural network surrogate for sheet metal stamping that combines GRU-based recurrent processor blocks with a U-Net-style graph downsample/upsample hierarchy, together with a node-to-surface contact feature representation. The model is evaluated on FE-generated datasets for a dome-shaped cold stamping case (50 training samples) and a larger bulkhead-shaped hot stamping case (100 training samples), and is compared against three ablations (vanillaGNN, RGNN, UGNN). The authors report that RUGNN achieves the lowest positional mean Euclidean error across all timesteps and reduces autoregressive error accumulation, and they analyze hyperparameters, training strategies, and input features.

Significance. If the reported accuracy and efficiency gains are robust, RUGNN is a useful contribution to learned simulation for forming processes: it directly addresses two known weaknesses of mesh-based GNN surrogates, namely autoregressive error accumulation and long-range information propagation on large graphs. The paper also proposes a lightweight node-to-surface contact encoding that avoids the cost of dense contact-edge graphs. The ablation design is reasonable, the ground truth is held-out FE simulation rather than the model's own predictions, and the authors provide feature and hyperparameter analyses. These are genuine strengths. The central empirical claim, however, is currently supported only by single-run point estimates without uncertainty quantification, and the paper does not provide code, data, or quantitative accuracy numbers, so the headline ordering is not yet independently verifiable.

major comments (4)
  1. [§7.1, §7.2, Figs. 10(b), 12(a)] The central claim that RUGNN 'achieves the lowest positional MEE across all timesteps' rests entirely on point estimates with no uncertainty quantification. The paper does not state how many random initializations were used, how final checkpoints were selected, or whether differences were consistent across seeds. Because deep GNN training has nontrivial run-to-run variance, and because the configuration analysis in Table 2 reports validation-loss differences as small as 0.0791 versus 0.0879, the reported ranking over RGNN and UGNN could be due to seed noise rather than architecture. The authors should report repeated-seed mean and variance, error bars or confidence intervals on the MEE curves, and ideally paired significance tests or effect sizes for the final-timestep comparison.
  2. [§7.3, Table 2] The hyperparameters and feature choices that define the final RUGNN model, including message-passing layer counts, latent dimensions, node reduction ratio, batch size, and use of contact features and hidden-state initialization, were selected on validation MSE. There is no separate evaluation of the selected configuration on the held-out test set that accounts for this model-selection step, and no repeated-seed evaluation of the chosen configuration. This makes it difficult to separate genuine architectural gains from selection on the validation set, especially given the small training sets (50 and 100 samples) and small validation/test sets. The revised paper should report test-set results for the selected configuration with uncertainty, or describe a nested validation procedure.
  3. [§4.2, Appendix A, Table 2] The claim that the node-to-surface contact representation 'ensures no loss of critical interaction details' is stronger than the evidence supports. Appendix A only compares node-to-surface distance magnitudes at one timestep of one sample and for seven selected nodes; it does not validate that the inverse-distance and closest-normal-vector encoding preserves contact force transmission or deformation accuracy. Table 2 provides only one validation-loss comparison with and without contact features. The authors should add a quantitative test-set comparison of deformation fields with and without contact features, and ideally an error breakdown by distance to the tool surface, to substantiate the contact-modeling contribution.
  4. [Data availability and reproducibility] The empirical comparison cannot currently be independently reproduced because the code and data are only 'available upon reasonable request' and the paper does not provide training configurations such as optimizer schedules, random seeds, checkpoint selection criteria, or absolute efficiency measurements. For a methods paper whose central claim is an empirical ordering, the authors should release code and data (or a detailed artifact description) and report the exact training protocol, including the number of runs and the rule for choosing the final model.
minor comments (6)
  1. [Equation (6)] The loss notation is inconsistent: the text defines ℒMSE but the equation and surrounding prose use ℒMS; please correct the label.
  2. [§3.2 and §7.2] Section 3.2 describes the bulkhead material as a 'time-independent elastic-viscoplastic material model,' while Section 7.2 and the introduction describe the material as temperature- and strain-rate dependent and 'temporal dependant.' This is contradictory and should be corrected, since the time dependence is part of the motivation for the recurrent architecture.
  3. [Appendix B, reference [34]] The gPartition method is cited to reference [34], but reference [34] is the ST-UNet paper by Yu et al.; please verify and correct this citation to the actual source of the graph partition algorithm.
  4. [Table 2] Table 2 is difficult to read: the row and column semantics are unclear, and it is not immediately obvious which column corresponds to which factor. Please restructure the table so each factor is a separate column with clear header labels.
  5. [§7.1, §7.2] The text reports only qualitative statements such as 'closely matching' and 'significantly reduces'; no absolute MEE values, GPU memory values, or training times are given. Please include the actual numbers in the text or in a table so readers can assess the magnitude of the reported improvements.
  6. [Throughout] There are several typos and grammatical errors, including 'extendes' (Section 2.1), 'the punch move' (Sections 3.1 and 3.2), 'obejects' (Section 3.2), and 'the positional mean positional Euclidean errors' (Section 7.1). A careful proofreading pass is needed.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central accuracy ranking is measured against held-out FE ground truth, not derived from the model's inputs or from self-citations.

full rationale

The paper's central claim is an empirical ordering: RUGNN achieves the lowest positional MEE across timesteps relative to vanillaGNN, RGNN, and UGNN (Sections 7.1 and 7.2, Figs. 10-13). This ordering is computed from Equation (7), which compares predicted node positions with FE ground-truth positions on disjoint test sets (50/20/20 and 100/45 splits in Section 3). The displacement loss in Equation (6) also targets FE ground truth, so the fitted parameters do not by construction define the evaluation metric. The contact representation in Section 4.2 is an input feature (inverse node-to-surface distance and closest-surface normal); it is not the training target, and its effectiveness is tested by an ablation (Table 2) and by an independent geometric check in Appendix A. The recurrent processor block is attributed to prior work [23] with overlapping authorship, but this is architectural inheritance rather than load-bearing derivation: the paper additionally isolates the contribution of recurrence through the RGNN-versus-vanillaGNN and RUGNN-versus-UGNN comparisons, and the cited prior work is a separate published study with its own datasets. Similarly, the data-generation pipeline follows [5], but that is methodological reuse, not a self-citation used to justify predictions. Hyperparameter selection on the validation set (Fig. 14 and Table 2) is standard model tuning and does not make the test-set comparison circular. I find no step in which a prediction is equivalent by construction to its inputs or in which a self-citation is the sole support for the central claim.

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

All central claims rest on FE-generated datasets from commercial software and on several architecture choices tuned on the validation set. The model does not introduce new physical entities, but it depends on the adequacy of the contact feature representation and the ability of the small training sets to cover the design space.

free parameters (7)
  • Message passing layers per RGPB = 2 in RGPB1, 10 in RGPB2, 2 in RGPB3 (dome); same structure for bulkhead
    Selected via hyperparameter study in Fig. 14(a) and Appendix C/E.
  • Latent feature dimensions per graph level = Dome: 32, 32, 64, 128; bulkhead: 32, 32, 64, 128, 128
    Architecture choice from model design in Appendix C/E.
  • Node reduction ratio = 3
    Chosen from Fig. 14(b) as a balance between spatial preservation and efficiency.
  • Batch size = 1
    Chosen from Table 2; batch size 1 gave lower rollout validation loss.
  • Number of downsample/upsample layers = 3 for dome, 4 for bulkhead
    Chosen from Fig. 14(c) based on graph size.
  • Hidden state initialization = zero vectors
    Table 2 shows zero initialization slightly outperformed global-feature initialization.
  • Learning rate schedule = exponential decay from exp(-4) to exp(-6)
    Training strategy in Section 6; not derived from first principles.
assumptions (5)
  • domain assumption PAMSTAMP finite element simulation outputs are valid ground truth for real sheet forming behavior.
    Section 3 uses FE simulations as the only supervision; no physical experiments validate the surrogate against measured deformation.
  • domain assumption Node-to-surface distance and closest-surface normal are sufficient contact features.
    Section 4.2 states this representation 'ensures no loss of critical interaction details'; Appendix A checks only distance magnitudes, not force transmission or deformation accuracy.
  • domain assumption Uniform blank shape and specified DoE parameter ranges define the usable domain.
    Sections 3.1-3.2 train and test on LHC-sampled geometries within fixed ranges; generalization to other blank shapes or parameters outside ranges is untested.
  • domain assumption Manual graph coarsening preserves the geometry needed for hierarchical message passing.
    Section 4.3 adopts manual coarsening based on qualitative comparison in Appendix B; no quantitative geometric error is reported.
  • standard math Standard neural network training with Adam and MSE minimizes to a useful local optimum.
    Section 6 relies on standard supervised training; no convergence guarantees are provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Recurrent U-Net-Based Graph Neural Network (RUGNN) for Accurate Deformation Predictions in Sheet Material Forming." pith.science (2026). https://pith.science/paper/PMZTNFTY

@misc{pith2026250711547,
  author       = {Pith},
  title        = {Pith review of: Recurrent U-Net-Based Graph Neural Network (RUGNN) for Accurate Deformation Predictions in Sheet Material Forming},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PMZTNFTY}},
  note         = {Machine review of arXiv:2507.11547}
}
read the original abstract

In recent years, various artificial intelligence-based surrogate models have been proposed to provide rapid manufacturability predictions of material forming processes. However, traditional AI-based surrogate models, typically built with scalar or image-based neural networks, are limited in their ability to capture complex 3D spatial relationships and to operate in a permutation-invariant manner. To overcome these issues, emerging graph-based surrogate models are developed using graph neural networks. This study developed a new graph neural network surrogate model named Recurrent U Net-based Graph Neural Network (RUGNN). The RUGNN model can achieve accurate predictions of sheet material deformation fields across multiple forming timesteps. The RUGNN model incorporates Gated Recurrent Units (GRUs) to model temporal dynamics and a U-Net inspired graph-based downsample/upsample mechanism to handle spatial long-range dependencies. A novel 'node-to-surface' contact representation method was proposed, offering significant improvements in computational efficiency for large-scale contact interactions. The RUGNN model was validated using a cold forming case study and a more complex hot forming case study using aluminium alloys. Results demonstrate that the RUGNN model provides accurate deformation predictions closely matching ground truth FE simulations and outperforming several baseline GNN architectures. Model tuning was also performed to identify suitable hyperparameters, training strategies, and input feature representations. These results demonstrate that RUGNN is a reliable approach to support sheet material forming design by enabling accurate manufacturability predictions.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

40 extracted references · 26 canonical work pages

  1. [1]

    J. L. D. Z. Marciniak, S.J. Hu, Mechanics of Sheet Metal Forming. Great Britain, 2002

    work page 2002

  2. [2]

    A new design guideline development strategy for aluminium alloy corners formed through cold and hot stamping processes,

    H. Attar, N. Li, and A. Foster, "A new design guideline development strategy for aluminium alloy corners formed through cold and hot stamping processes," Materials & Design, p. 109856, 05/01 2021, doi: 10.1016/j.matdes.2021.109856

    work page arXiv 2021

  3. [3]

    Optimization of an aluminum alloy anti -collision side beam hot stamping process using a multi -objective genetic algorithm,

    J. Zhou, B. Wang, J. Lin, and L. Fu, "Optimization of an aluminum alloy anti -collision side beam hot stamping process using a multi -objective genetic algorithm," Archives of Civil and Mechanical Engineering, vol. 13, no. 3, pp. 401 -411, 2013/09/01/ 2013, doi: https://doi.org/10.1016/j.acme.2013.01.008

    work page doi:10.1016/j.acme.2013.01.008 2013

  4. [4]

    A study on the buckling behaviour of aluminium alloy sheet in deep drawing with macro -textured blankholder,

    K. Zheng, D. J. Politis, J. Lin, and T. A. Dean, "A study on the buckling behaviour of aluminium alloy sheet in deep drawing with macro -textured blankholder," International Journal of Mechanical Sciences, vol. 110, pp. 138 -150, 2016/05/01/ 2016, doi: https://doi.org/10.1016/j.ijmecsci.2016.03.011

    work page doi:10.1016/j.ijmecsci.2016.03.011 2016

  5. [5]

    Rapid feasibility assessment of components to be formed through hot stamping: A deep learning approach,

    H. Attar, H. Zhou, A. Foster, and N. Li, "Rapid feasibility assessment of components to be formed through hot stamping: A deep learning approach," Journal of Manufacturing Processes, vol. 68, pp. 1650-1671, 08/01 2021, doi: 10.1016/j.jmapro.2021.06.011

    doi:10.1016/j.jmapro.2021.06.011 2021

  6. [6]

    A Study on Using Image -Based Machine Learning Methods to Develop Surrogate Models of Stamp Forming Simulations,

    H. Zhou, Q. Xu, Z. Nie, and N. Li, "A Study on Using Image -Based Machine Learning Methods to Develop Surrogate Models of Stamp Forming Simulations," Journal of Manufacturing Science and Engineering, vol. 144, pp. 1-41, 06/28 2021, doi: 10.1115/1.4051604

    work page doi:10.1115/1.4051604 2021

  7. [7]

    Springback prediction for sheet metal cold stamping using convolutional neural networks,

    L. Zhu and N. Li, "Springback prediction for sheet metal cold stamping using convolutional neural networks," in 2022 Workshop on Electronics Communication Engineering, 2023, vol. 12720: SPIE, pp. 278-283

    work page 2022

  8. [8]

    Deep residual U-net with input of static structural responses for efficient U* load transfer path analysis,

    S. Zhao, N. Wu, and Q. Wang, "Deep residual U-net with input of static structural responses for efficient U* load transfer path analysis," Advanced Engineering Informatics, vol. 46, p. 101184, 2020/10/01/ 2020, doi: https://doi.org/10.1016/j.aei.2020.101184

    arXiv 2020

Show all 40 references
  1. [9]

    FuS -GCN: Efficient B -rep based graph convolutional networks for 3D -CAD model classification and retrieval,

    J. Hou, C. Luo, F. Qin, Y. Shao, and X. Chen, "FuS -GCN: Efficient B -rep based graph convolutional networks for 3D -CAD model classification and retrieval," Advanced Engineering Informatics, vol. 56, p. 102008, 2023/04/01/ 2023, doi: https://doi.org/10.1016/j.aei.2023.102008

    2023

  2. [10]

    Graph neural networks: A review of methods and applications,

    J. Zhou et al., "Graph neural networks: A review of methods and applications," AI Open, vol. 1, pp. 57-81, 2020/01/01/ 2020, doi: https://doi.org/10.1016/j.aiopen.2021.01.001

    2020 doi

  3. [11]

    Learning mesh -based simulation with graph networks,

    T. Pfaff, M. Fortunato, A. Sanchez -Gonzalez, and P. W. Battaglia, "Learning mesh -based simulation with graph networks," arXiv preprint arXiv:2010.03409, 2020

    2010 arXiv

  4. [12]

    Everything is connected: Graph neural networks,

    P. Veli čković, "Everything is connected: Graph neural networks," Current Opinion in Structural Biology, vol. 79, p. 102538, 2023/04/01/ 2023, doi: https://doi.org/10.1016/j.sbi.2023.102538

    2023

  5. [13]

    Learning to simulate complex physics with graph networks,

    A. Sanchez -Gonzalez, J. Godwin, T. Pfaff, R. Ying, J. Leskovec, and P. W. Battaglia, "Learning to simulate complex physics with graph networks," presented at the Proceedings of the 37th International Conference on Machine Learning, 2020. 30

    2020

  6. [14]

    Relational inductive biases, deep learning, and graph networks,

    P. W. Battaglia et al., "Relational inductive biases, deep learning, and graph networks," arXiv preprint arXiv:1806.01261, 2018

    2018 arXiv

  7. [15]

    A review of graph neural network applications in mechanics -related domains,

    Y. Zhao, H. Li, H. Zhou, H. R. Attar, T. Pfaff, and N. Li, "A review of graph neural network applications in mechanics -related domains," Artificial Intelligence Review, vol. 57, no. 11, p. 315, 2024/10/04 2024, doi: 10.1007/s10462-024-10931-y

    2024 doi

  8. [16]

    Graph neural network -assisted evolutionary algorithm for rapid optimization design of shear-wall structures,

    Y. Fei, S. Qin, W. Liao, H. Guan, and X. Lu, "Graph neural network -assisted evolutionary algorithm for rapid optimization design of shear-wall structures," Advanced Engineering Informatics, vol. 65, p. 103129, 2025/05/01/ 2025, doi: https://doi.org/10.1016/j.aei.2025.103129

    2025

  9. [17]

    Design -condition-informed shear wall layout design based on graph neural networks,

    P. Zhao, Y. Fei, Y. Huang, Y. Feng, W. Liao, and X. Lu, "Design -condition-informed shear wall layout design based on graph neural networks," Advanced Engineering Informatics, vol. 58, p. 102190, 2023/10/01/ 2023, doi: https://doi.org/10.1016/j.aei.2023.102190

    2023

  10. [18]

    Learning finite element convergence with the Multi-fidelity Graph Neural Network,

    N. Black and A. R. Najafi, "Learning finite element convergence with the Multi-fidelity Graph Neural Network," Computer Methods in Applied Mechanics and Engineering, vol. 397, 2022

    2022

  11. [19]

    Learning rigid dynamics with face interaction graph networks,

    K. R. Allen et al., "Learning rigid dynamics with face interaction graph networks," arXiv preprint arXiv:2212.03574, 2022

    2022 arXiv

  12. [20]

    Physics -Encoded Graph Neural Networks for Deformation Prediction under Contact,

    M. Saleh, M. Sommersperger, N. Navab, and F. Tombari, "Physics -Encoded Graph Neural Networks for Deformation Prediction under Contact," arXiv preprint arXiv:2402.03466, 2024

    2024 arXiv

  13. [21]

    Graph networks as learnable physics engines for inference and control,

    A. Sanchez-Gonzalez et al., "Graph networks as learnable physics engines for inference and control," in International Conference on Machine Learning, 2018: PMLR, pp. 4470-4479

    2018

  14. [22]

    Towards Fast Simulation of Environmental Fluid Mechanics with Multi -Scale Graph Neural Networks,

    M. Lino, S. Fotiadis, A. A. Bharath, and C. D. Cantwell, "Towards Fast Simulation of Environmental Fluid Mechanics with Multi -Scale Graph Neural Networks," ArXiv, vol. abs/2205.02637, 2022

    2022 arXiv

  15. [23]

    Predicting dynamic responses of continuous deformable bodies:A graph-based learning approach,

    Q. Chen, J. Cao, W. Lin, S. Zhu, and S. Wang, "Predicting dynamic responses of continuous deformable bodies:A graph-based learning approach," Computer Methods in Applied Mechanics and Engineering, vol. 420, p. 116669, 2024/02/15/ 2024, doi: https://doi.org/10.1016/j.cma.2023.116669

    2024

  16. [24]

    Physics-informed graph neural Galerkin networks: A unified framework for solving PDE-governed forward and inverse problems,

    H. Gao, M. J. Zahr, and J. -X. Wang, "Physics-informed graph neural Galerkin networks: A unified framework for solving PDE-governed forward and inverse problems," Computer Methods in Applied Mechanics and Engineering, vol. 390, p. 114502, 2022/02/15/ 2022, doi: https://doi.org...

    2022

  17. [25]

    Physics -informed graph neural network emulation of soft-tissue mechanics,

    D. Dalton, D. Husmeier, and H. Gao, "Physics -informed graph neural network emulation of soft-tissue mechanics," Computer Methods in Applied Mechanics and Engineering, vol. 417, p. 116351, 2023/12/01/ 2023, doi: https://doi.org/10.1016/j.cma.2023.116351

    2023

  18. [26]

    On the use of graph neural networks and shape - function-based gradient computation in the deep energy method,

    J. He, D. Abueidda, S. Koric, and I. Jasiuk, "On the use of graph neural networks and shape - function-based gradient computation in the deep energy method," International Journal for Numerical Methods in Engineerin, vol. 124, 2023

    2023

  19. [27]

    Metal forming progress since 2000,

    J. Jeswiet et al. , "Metal forming progress since 2000," CIRP Journal of Manufacturing Science and Technology, vol. 1, no. 1, pp. 2 -17, 2008/01/01/ 2008, doi: https://doi.org/10.1016/j.cirpj.2008.06.005. 31

    2000 doi

  20. [28]

    Numerical analysis on the elastic deformation of the tools in sheet metal forming processes,

    D. M. Neto, J. Coë r, M. C. Oliveira, J. L. Alves, P. Y. Manach, and L. F. Menezes, "Numerical analysis on the elastic deformation of the tools in sheet metal forming processes," International Journal of Solids and Structures, vol. 100 -101, pp. 270 -285, 2016/12/01/ 2016, doi...

    2016 doi

  21. [29]

    Experimental investigation of boron steel at hot stamping conditions,

    N. Li et al., "Experimental investigation of boron steel at hot stamping conditions," Journal of Materials Processing Technology, vol. 228, pp. 2 -10, 2016/02/01/ 2016, doi: https://doi.org/10.1016/j.jmatprotec.2015.09.043

    2016 doi

  22. [30]

    Investigation of deformation and failure features in hot stamping of AA6082: Experimentation and modelling,

    M. S. Mohamed, A. D. Foster, J. Lin, D. S. Balint, and T. A. Dean, "Investigation of deformation and failure features in hot stamping of AA6082: Experimentation and modelling," International Journal of Machine Tools and Manufacture, vol. 53, no. 1, pp. 27-38, 2012/02/01/ 2012,...

    2012 doi

  23. [31]

    MAgNET: A graph U -Net architecture for mesh-based simulations,

    S. Deshpande, S. P. A. Bordas, and J. Lengiewicz, "MAgNET: A graph U -Net architecture for mesh-based simulations," Engineering Applications of Artificial Intelligence, vol. 133, p. 108055, 2024/07/01/ 2024, doi: https://doi.org/10.1016/j.engappai.2024.108055

    2024

  24. [32]

    MultiScale MeshGraphNets,

    M. Fortunato, T. Pfaff, P. Wirnsberger, A. Pritzel, and P. W. Battaglia, "MultiScale MeshGraphNets," ArXiv, vol. abs/2210.00612, 2022

    2022 arXiv

  25. [33]

    Learning rigid -body simulators over implicit shapes for large -scale scenes and vision,

    Y. Rubanova, T. Lopez-Guevara, K. R. Allen, W. F. Whitney, K. Stachenfeld, and T. Pfaff, "Learning rigid -body simulators over implicit shapes for large -scale scenes and vision," arXiv preprint arXiv:2405.14045, 2024

    2024 arXiv

  26. [34]

    St -unet: A spatio-temporal u-network for graph-structured time series modeling,

    B. Yu, H. Yin, and Z. Zhu, "St -unet: A spatio-temporal u-network for graph-structured time series modeling," arXiv preprint arXiv:1903.05631, 2019

    1903 arXiv

  27. [35]

    An integrated preforming -performance model for high - fidelity performance analysis of cured woven composite part with non -orthogonal yarn angles,

    B. Liang, S. Gao, and W. Zhang, "An integrated preforming -performance model for high - fidelity performance analysis of cured woven composite part with non -orthogonal yarn angles," Chinese Journal of Aeronautics, vol. 35, no. 6, pp. 367 -378, 2022/06/01/ 2022, doi: https://d...

    2022 doi

  28. [36]

    An investigation of a new 2D CDM model in predicting failure in HFQing of an automotive panel,

    M. Mohamed et al. , "An investigation of a new 2D CDM model in predicting failure in HFQing of an automotive panel," MATEC Web of Conferences, vol. 21, p. 05011, 2015. [Online]. Available: https://doi.org/10.1051/matecconf/20152105011

    2015

  29. [37]

    An investigation of lubrication and heat transfer for a sheet aluminium heat, form-quench (HFQ) process,

    A. Foster, M. Mohamed, J. Lin, and T. Dean, "An investigation of lubrication and heat transfer for a sheet aluminium heat, form-quench (HFQ) process," Steel Research International, vol. 79, pp. 113-120, 02/01 2008

    2008

  30. [38]

    Sliding interfaces with contact-impact in large-scale Lagrangian computations,

    J. O. Hallquist, G. L. Goudreau, and D. J. Benson, "Sliding interfaces with contact-impact in large-scale Lagrangian computations," Computer Methods in Applied Mechanics and Engineering, vol. 51, no. 1, pp. 107-137, 1985/09/01/ 1985, doi: https://doi.org/10.1016/0045-7825(85)90030-1

    1985 doi

  31. [39]

    Neural message passing for quantum chemistry,

    J. Gilmer, S. S. Schoenholz, P. F. Riley, O. Vinyals, and G. E. Dahl, "Neural message passing for quantum chemistry," in International conference on machine learning , 2017: PMLR, pp. 1263 - 1272

    2017

  32. [40]

    Efficient learning of mesh -based physical simulation with bi -stride multi -scale graph neural network,

    Y. Cao, M. Chai, M. Li, and C. Jiang, "Efficient learning of mesh -based physical simulation with bi -stride multi -scale graph neural network," presented at the Proceedings of the 40th International Conference on Machine Learning, Honolulu, Hawaii, USA, 2023

    2023

Pith tools

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