Pith. sign in

REVIEW 5 major objections 5 minor 43 references

A Graph Neural Network approach to zero-shot Digital Twins

T0 review · 5 major / 5 minor · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read This paper claims that a single geometry-agnostic, thermodynamics-informed graph neural network, paired with live video feedback, can simulate unseen solids and fluids in real time and infer hidden stress and velocity fields without retrain

desk verdict A well-engineered integration of the authors' own Local-TIGNN with vision and assimilation, but the real-world zero-shot claim is weaker than advertised because open-loop physics fails and the hidden fields are never directly validated. read the letter →

arxiv 2607.20535 v1 pith:RCTA5IJO submitted 2026-07-10 cs.LG cs.AIcs.CV

Alicia Tierz , Icíar Alfaro , David González , Elías Cueto This is my paper
classification cs.LGcs.AIcs.CV MSC 68T0774S0576M2880A05
keywords DigitaltwinGraphneuralnetworkZero-shotgeneralizationPhysics-informedmachinelearningThermodynamics-informedDataassimilationMetriplecticformalismReal-timesimulation
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 predictive digital twin does not need to be rebuilt for each new object: the same learned physical engine, which encodes energy conservation and entropy production as local graph interactions, can be dropped onto geometries it has never seen and still produce physically accurate simulations. To make that work on real video, the authors couple the engine to a perception module that reconstructs the visible boundary and a closed-loop assimilation step that continuously nudges the simulation back toward what the camera sees. They argue this makes hidden fields—stress in a bending beam, velocity and internal energy in sloshing fluid—physically anchored, not just visually aligned. If true, the practical payoff is large: structural and fluid monitoring could run at interactive rates, roughly 10–25 ms per frame, on arbitrary scenes without case-specific retraining. The two demonstrations, a viscoelastic cantilever and viscous sloshing, are meant to show the framework spans disparate physics with one unified solver.

What carries the argument

The Local-TIGNN (Thermodynamics-Informed Graph Neural Network): a message-passing solver built on the GENERIC metriplectic formalism, in which each node obeys a nodal port-metriplectic evolution equation that keeps the Poisson operator skew-symmetric and the friction operator positive semi-definite, enforcing energy conservation and non-negative entropy production locally. The auxiliary initialization network Ψini maps observed geometry to latent thermodynamic states to avoid cold-start transients, and the closed-loop assimilation uses column-wise vertical rescaling for fluids and keypoint nudging for solids to anchor the autoregressive rollout to the video feed.

What would settle it

Run the closed-loop twin on a real beam instrumented with embedded strain gauges or on a real tank with particle image velocimetry, then compare the inferred internal stress or velocity fields to the sensors while the visible boundary is being corrected; if the hidden fields diverge substantially despite perfect boundary alignment, the claim that boundary nudging physically anchors internal states is falsified.

Watch

Extended reading notes

Core claim

The central claim is that thermodynamics can be encoded in a graph neural network as purely local, port-based interaction laws—each node is an open thermodynamic system exchanging energy and entropy fluxes with neighbors—so that the learned solver is inherently geometry-agnostic. Because no global Poisson or dissipation matrix is assembled, the same trained network can be applied to any mesh or point cloud the perception system produces, including objects whose shape was never in the training set. Combined with an auxiliary network that initializes latent fields from sparse geometry, and a continuous visual correction loop, the system claims zero-shot deployment on unseen geometries with phy

Load-bearing premise

The load-bearing premise is that continuously correcting the visible boundary, the beam's tracked grid or the fluid's free surface, forces the network's unseen internal fields—stress, velocity, and energy—to be physically accurate, even though no experiment in the paper measures those internal fields directly.

Editorial extensions

If this is right

  • A single trained physics engine can be deployed on arbitrary unseen geometries at inference time, removing the retraining bottleneck that currently limits predictive digital twins.
  • Hidden mechanical fields—stress tensors, velocity, internal energy—can be reconstructed from boundary vision alone, enabling augmented-reality overlays that make invisible quantities visible in real time.
  • Continuous visual assimilation prevents autoregressive drift even when offline accuracy is imperfect, so the twin stays synchronized with the physical asset over long horizons.
  • The same core solver works across structurally and fluid-dynamically distinct regimes, large-deformation viscoelastic solids and nonlinear free-surface sloshing, suggesting a path toward general-purpose learned physics engines.
  • End-to-end latencies of about 9–25 ms per frame show the pipeline can run above standard real-time visualization thresholds, leaving budget for control or rendering.

Reading between the lines

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

  • If the geometry-agnostic claim generalizes, the same engine could plausibly be transferred to other dissipative continua, such as soft tissues, granular media, or multiphase flows, provided the perception module can supply boundary geometry; this is an extension the paper does not test.
  • The assimilation loop may be responsible for much of the apparent physical accuracy, since hidden fields are never directly measured; an ablation that replaces the physics engine with a geometry-only interpolator while keeping the visual correction would isolate how much thermodynamic structure contributes.
  • The fixed-depth monocular projection implies the framework should degrade under out-of-plane motion; a natural test is to introduce depth variation and compare twin fidelity, which would also motivate the RGB-D or stereoscopic extension the authors propose as future work.
  • If the internal-field inference is trusted, the framework becomes a candidate sensor itself: the twin's stress or velocity estimates could drive health monitoring or control decisions without adding physical sensors, but that trust depends on validation the current experiments do not yet provide.
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

5 major / 5 minor

Summary. The paper proposes a unified zero-shot Digital Twin framework that couples a Thermodynamics-Informed Graph Neural Network (Local-TIGNN) physics engine with a real-time monocular vision pipeline, an auxiliary initialization network (Ψini) for latent-state inference, and a closed-loop data-assimilation mechanism that nudges the simulation toward observed boundaries. The framework is evaluated on two disparate physical regimes: large deformations of a viscoelastic cantilever beam and nonlinear sloshing of a viscous fluid. The authors report sub-percent positional RRMSE on synthetic solid tests (0.30%), 2.49% positional RRMSE on synthetic fluid tests, real-time latencies of 9.1 ms (solid) and 25.2 ms (fluid), and claim that the system projects physically accurate latent stress/velocity/energy fields onto real, never-seen objects via Augmented Reality.

Significance. If fully substantiated, the framework would be a significant integration of structure-preserving learned physics with real-time perception and assimilation, with clear value for cognitive digital twins and industrial monitoring. The paper's strengths include: the synthetic training data are independently grounded in Abaqus FEM and SPH simulations; the closed-loop boundary tracking is demonstrated on real experiments; and the measured latencies support the real-time claim. The main weakness is that the central claim of physically accurate latent fields on real objects is not directly validated: the closed-loop assimilation corrects only observable boundaries, and the internal stress/velocity/energy fields are never compared against direct measurements on the physical system. The real-world evidence therefore supports the boundary-tracking result but not the hidden-field claim.

major comments (5)
  1. [3.1, Eqs. (3)–(4)] The central claim of 'physically accurate' latent fields on real, unseen objects is not established. The closed-loop correction injects only observed boundary geometry—tracked nodal positions in the solid case (§5.1.5) and the free-surface profile in the fluid case (§5.2.5)—while the inferred internal stress, velocity, and energy fields are never compared against direct measurements (e.g., strain gauges, DIC, PIV). The statement in §5.1.5 that geometric correction 'guarantees that the inferred latent stress fields (σ) remain physically anchored' is asserted, not validated. The only quantitative internal-field accuracies are the synthetic offline metrics (Table 1: stress RRMSE 11.34%; Table 2: velocity RRMSE 26.34%). Therefore the headline claim that the system projects physically accurate latent mechanical variables on real objects is unsupported exactly where it matters most. This is an
  2. [3.1, Eqs. (3)–(4)] The abstract and §3.1 claim that the Local-TIGNN 'enforces energy conservation and non-negative entropy production locally through graph message passing' and 'guarantees thermodynamic consistency by construction.' However, the degeneracy conditions (4) are only incorporated as a soft constraint in the loss function, not enforced as a hard architectural constraint. Skew-symmetry of L and positive semi-definiteness of M are structural, but without the degeneracy conditions, the GENERIC evolution does not strictly conserve energy or produce entropy monotonically. Please clarify the strength of the guarantee and provide empirical verification of conservation/dissipation properties on the test trajectories, or revise the abstract accordingly.
  3. [5.2.4] The 'zero-shot deployment ... without case-specific retraining' claim is partially contradicted by the warm-start transfer learning protocol: the fluid network is initialized with weights optimized on a water sloshing baseline and then 'subsequently fine-tuning on the specific bi-distilled glycerin dataset.' This is case-specific training for the fluid material regime. If the zero-shot claim refers only to geometry changes within a fixed material regime, that scope should be stated explicitly in the abstract and Section 4.
  4. [3.2.1, Eq. (5)] The monocular reconstruction assumes a known constant depth d and planar motion. This assumption is acknowledged in the conclusion, but it places a strong restriction on the claimed 'novel, unseen geometries' and 'fully unconstrained three-dimensional dynamics' framing. Out-of-plane motion directly corrupts the observed boundary that anchors the entire closed loop. The planarity restriction should be listed as a formal limitation in the abstract or contributions, and the zero-shot claim should be scoped to quasi-planar scenes.
  5. [3.2.2, Eqs. (6)–(7)] The column-wise vertical rescaling is claimed to preserve 'local volume consistency,' but no proof or quantitative assessment is provided. Rescaling the vertical coordinate of every particle by a column-dependent factor γ(x) changes the volume element unless the column width and the Jacobian are accounted for; for curved or merging/splitting free surfaces, this operation can artificially compress or expand the fluid. Please provide a volume-error analysis on the test cases, or replace the assertion with a measured volume-conservation metric.
minor comments (5)
  1. [General] The introduction refers to 'Section II' and 'Section III' using Roman numerals, while the actual section headings are numbered 1, 2, 3, etc. Please harmonize the cross-references.
  2. [§5.1.4] The reported stress RMSE for σ11 is given as 40.28 without units. Specify the units (presumably Pa) for clarity and consistency with Table 1.
  3. [§5.2.4] The explanation of the inflated velocity RRMSE (26.34%) as a 'numerical artifact' of near-zero denominators is plausible but should be supported by reporting an additional error metric, such as RMSE normalized by the maximum velocity magnitude or a velocity threshold-based metric.
  4. [§3.1, Eq. (3)] The notation for the local matrices L_i, M_i, L_ij, M_ij is introduced only briefly; a reader unfamiliar with port-metriplectic formulations would benefit from a more explicit definition of how these matrices are constructed from node/edge features and how the structural constraints are parameterized in the neural network.
  5. [§4.1] The term 'geometry-agnostic' is used as a synonym for 'trained on local interaction rules.' Since the model is trained on a specific material regime and then transferred, the term 'geometry-agnostic' may be misleading; consider 'mesh-agnostic' or 'topology-agnostic' to avoid overclaiming.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular reduction: the physics engine is trained on independent FEM/SPH ground truth, and data assimilation corrects only observable boundaries without defining the hidden fields.

full rationale

The derivation chain is not circular. The Local-TIGNN is trained on Abaqus FEM/SPH outputs (Sections 5.1.2, 5.2.2), so the learned dynamics are grounded in independent high-fidelity solvers rather than in the quantities the paper later claims to predict. The auxiliary initialization network Ψini is likewise trained on synthetic ground truth to map boundary geometry to stress/velocity/energy fields; it is not defined in terms of the real observations. The closed-loop assimilation (Eqs. 6–7 for fluids; keypoint nudging for solids) explicitly corrects only observable boundary geometry and is never presented as a prediction of that geometry. Consequently, the hidden fields (stress, velocity, energy) are not statistically forced by the fit; the paper's assertion that they are 'physically anchored' on real objects is an unvalidated inference, not a circular one. The many self-citations ([5], [19], [25], [40], etc.) are to published prior work with stated formulations and independent experiments; the GENERIC equations and training data are also stated in the paper, so the core result does not reduce to a self-citation chain. The acknowledged planar-motion/depth limitation (Conclusion) is a correctness/robustness concern, not circularity. No equation in the paper is equivalent to its inputs by construction.

Assumptions & free parameters 5 free parameters · 6 assumptions · 1 invented entities

The paper introduces no new physical entities or forces. Its load-bearing free parameters are the beam's effective material constants (calibrated, not measured) and the assumed constant camera depth d. Its axioms are the adequacy of GENERIC at the resolved scales, the validity of the local port-metriplectic decomposition, the well-posedness of the boundary-to-hidden-state inverse map, and the untested claim that boundary nudging yields correct hidden fields. This ledger is the honest measure of what is assumed versus demonstrated.

free parameters (5)
  • Beam density ρ = 18.44 kg/m³
    Assigned in the Abaqus KSV-viscoelastic calibration (§5.1.1); unphysically low for a solid polymer, indicating it is an effective parameter tuned so the FEM matches video dynamics.
  • Young's modulus E = 4.5×10⁴ Pa
    Obtained from an Euler-Bernoulli estimate refined with the Abaqus KSV model (§5.1.1); fitted to the real beam's response.
  • Prony shear relaxation parameters = ḡ₁=0.5, τ₁=0.05 s
    Chosen in the mechanical characterization (§5.1.1) to reproduce observed viscoelastic relaxation; not measured independently.
  • Camera depth d = assumed constant
    Eq. (5) reconstructs 3D from a monocular camera using a known, constant depth d; never measured, and load-bearing for every 3D graph geometry in both experiments.
  • Training hyperparameters = λd=50, noise 8×10⁻³, 2×100 units, 7 MP steps, 768×768 px, 300 epochs
    Reported in §5.2.3-5.2.4; standard tuning choices, minor compared with the material and depth parameters.
assumptions (6)
  • domain assumption GENERIC/metriplectic formalism (Eqs. 1-2) is an adequate governing framework for both viscoelastic solids and viscous free-surface fluids.
    Invoked in §3.1 as the foundation of the engine; the paper assumes its sufficiency at the resolved scales rather than arguing it.
  • domain assumption The nodal port-metriplectic decomposition (Eq. 3) with bulk degeneracy conditions (Eq. 4) preserves the global GENERIC structure.
    The decomposition is asserted in §3.1; the degeneracy conditions — the part that actually secures energy conservation — enter only as a soft loss term, so the preservation is approximate.
  • domain assumption Boundary observations suffice to recover the full internal thermodynamic state (well-posed inverse problem).
    Ψini (§3.3) maps sparse initial boundary geometry to stress/velocity/energy fields; no identifiability or uniqueness argument is given.
  • domain assumption Monocular reconstruction with known constant depth d and planar motion faithfully represents the real scene.
    Eq. (5) and §3.2.2; the conclusion explicitly admits this bounds the system to near-planar motions.
  • ad hoc to paper Column-wise rescaling (Eqs. 6-7) preserves local volume consistency of the fluid.
    Stated in §3.2.2 without derivation; a uniform vertical stretch γ(x) of each column changes per-column volume, and no conservation argument is supplied.
  • ad hoc to paper Nudging the visible boundary makes the hidden fields physically accurate.
    The core leap in §5.1.5/§5.2.5: geometric correction is claimed to keep latent stress/velocity fields 'physically anchored'; never tested against measured internal fields in the physical demos.
invented entities (1)
  • Auxiliary initialization network Ψini
    purpose: Maps sparse initial boundary geometry to unobserved state fields (stress tensor; velocity and internal energy) to avoid numerical start-up transients.
    A learned module introduced in §3.3; its outputs are validated only on synthetic test sets from the same Abaqus/SPH generation pipeline and never against direct measurement in the physical experiments.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Graph Neural Network approach to zero-shot Digital Twins." pith.science (2026). https://pith.science/paper/RCTA5IJO

@misc{pith2026260720535,
  author       = {Pith},
  title        = {Pith review of: A Graph Neural Network approach to zero-shot Digital Twins},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RCTA5IJO}},
  note         = {Machine review of arXiv:2607.20535}
}
read the original abstract

Traditional Predictive Digital Twins often remain geometrically rigid, requiring extensive retraining or fine-tuning whenever the underlying physical domain or boundary conditions change. To overcome this limitation, we present a novel framework for \textit{Zero-Shot Digital Twins} that seamlessly couples real-time visual perception with a geometry-agnostic, physics-informed reasoning engine. At the core of our architecture is the Thermodynamics-Informed Graph Neural Network architecture, a Geometric Deep Learning solver grounded in a metriplectic thermodynamic formalism that enforces energy conservation and non-negative entropy production locally through graph message passing. The framework integrates an auxiliary Graph Neural Network to infer unobservable fields (such as stress tensors or velocity and energy distributions) directly from sparse initial visual boundaries, mitigating numerical start-up transients. To bridge the sim-to-real gap, we implement a continuous closed-loop data assimilation mechanism; the pipeline tracks macroscopic deformations and free-surface fluid boundaries in real-time using deep segmentation networks combined with sparse optical flow, dynamically correcting the autoregressive simulation rollout and eliminating numerical drift. To test the validity of our approach, we demonstrate the extreme generalization capabilities of our approach across two disparate physical regimes: the large deformations of a viscoelastic beam and the non-linear sloshing of a viscous fluid. In both scenarios, the unified framework instantiates physically accurate simulations on novel, unseen geometries without case-specific retraining, operating well within real-time latency budgets (approximately 25 ms per frame) and enabling the direct projection of latent mechanical variables via Augmented Reality.

Figures

Figures reproduced from arXiv: 2607.20535 by the authors.

Figure 1
Figure 1. Comparison of Digital Twin Pipelines. (a) In the solid mechanics scenario, semantic segmentation (U-Net) [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Validation of the mechanical characterization. The Finite Element Method (FEM) solution computed in Abaqus (colored mesh) is superimposed onto the experimental video frame. The spatial alignment between the numerical prediction and the physical deformation corroborates the validity of the calibrated elastic parameters and the Euler-Bernoulli beam assumption. 9 [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Compensation of Sim-to-Real Drift. (A) Open-Loop Prediction: Despite high offline accuracy, the uncorrected [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Representative operation of the Cognitive Digital Twin on the hyperelastic beam. The top panel shows the [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]
Figure 5
Figure 5. Figure 5: Dynamic synchronization and latent vertical velocity field ( [PITH_FULL_IMAGE:figures/full_fig_p015_5.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

43 extracted references · 1 linked inside Pith

  1. [1]

    Tolle, Stewart Tansley, and Tony Hey

    K.M. Tolle, Stewart Tansley, and Tony Hey. The fourth paradigm: Data-intensive scientific discovery [point of view].Proceedings of the IEEE, 99:1334–1337, 08 2011

    2011

  2. [2]

    Academic press, 2019

    Fei Tao, Meng Zhang, and Andrew Yeh Chris Nee.Digital twin driven smart manufacturing. Academic press, 2019

    2019

  3. [3]

    Bronstein, Joan Bruna, Taco Cohen, and Petar Veliˇckovi´c

    Michael M. Bronstein, Joan Bruna, Taco Cohen, and Petar Veliˇckovi´c. Geometric deep learning: Grids, groups, graphs, geodesics, and gauges, 2021

    2021

  4. [4]

    Battaglia, Jessica B

    Peter W. Battaglia, Jessica B. Hamrick, Victor Bapst, Alvaro Sanchez-Gonzalez, Vinicius Zambaldi, Mateusz Malinowski, Andrea Tacchetti, David Raposo, Adam Santoro, Ryan Faulkner, Caglar Gulcehre, Francis Song, Andrew Ballard, Justin Gilmer, George Dahl, Ashish Vaswani, Kelsey Allen, Charles Nash, Victoria Langston, Chris Dyer, Nicolas Heess, Daan Wierstra...

    2018

  5. [5]

    Graph neural networks informed locally by thermodynamics.Engineering Applications of Artificial Intelligence, 144:110108, 2025

    Alicia Tierz, Icíar Alfaro, David González, Francisco Chinesta, and Elías Cueto. Graph neural networks informed locally by thermodynamics.Engineering Applications of Artificial Intelligence, 144:110108, 2025

    2025

  6. [6]

    Dynamics and thermodynamics of complex fluids

    Miroslav Grmela and Hans Christian Öttinger. Dynamics and thermodynamics of complex fluids. i. development of a general formalism.Phys. Rev. E, 56:6620–6632, Dec 1997

    1997

  7. [7]

    Learning to simulate complex physics with graph networks

    Alvaro Sanchez-Gonzalez, Jonathan Godwin, Tobias Pfaff, Rex Ying, Jure Leskovec, and Peter Battaglia. Learning to simulate complex physics with graph networks. InInternational conference on machine learning, pages 8459–

  8. [8]

    Battaglia

    Tobias Pfaff, Meire Fortunato, Alvaro Sanchez-Gonzalez, and Peter W. Battaglia. Learning mesh-based simulation with graph networks, 2021

    2021

Show all 43 references
  1. [9]

    Lu Jinzhi, Yang Zhaorui, Zheng Xiaochen, Wang Jian, and Kiritsis Dimitris. Exploring the concept of cogni- tive digital twin from model-based systems engineering perspective.The International Journal of Advanced Manufacturing Technology, 121(9):5835–5854, 2022

    2022

  2. [10]

    Integrating scientific knowledge with machine learning for engineering and environmental systems, 2022

    Jared Willard, Xiaowei Jia, Shaoming Xu, Michael Steinbach, and Vipin Kumar. Integrating scientific knowledge with machine learning for engineering and environmental systems, 2022

    2022

  3. [11]

    Physics- informed machine learning.Nature Reviews Physics, 3(6):422–440, 2021

    George Em Karniadakis, Ioannis G Kevrekidis, Lu Lu, Paris Perdikaris, Sifan Wang, and Liu Yang. Physics- informed machine learning.Nature Reviews Physics, 3(6):422–440, 2021

    2021

  4. [12]

    Raissi, P

    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.Journal of Computational Physics, 378:686–707, 2019

    2019

  5. [13]

    Hamiltonian neural networks.Advances in neural information processing systems, 32, 2019

    Samuel Greydanus, Misko Dzamba, and Jason Yosinski. Hamiltonian neural networks.Advances in neural information processing systems, 32, 2019

    2019

  6. [14]

    Discovering symbolic models from deep learning with inductive biases.Advances in neural information processing systems, 33:17429–17442, 2020

    Miles Cranmer, Alvaro Sanchez Gonzalez, Peter Battaglia, Rui Xu, Kyle Cranmer, David Spergel, and Shirley Ho. Discovering symbolic models from deep learning with inductive biases.Advances in neural information processing systems, 33:17429–17442, 2020

    2020

  7. [15]

    Morrison

    Philip J. Morrison. A paradigm for joined hamiltonian and dissipative systems.Physica D: Nonlinear Phenomena, 18(1):410–419, 1986

    1986

  8. [16]

    Structure-preserving neural networks.Journal of Computational Physics, 426:109950, 2021

    Quercus Hernández, Alberto Badías, David González, Francisco Chinesta, and Elías Cueto. Structure-preserving neural networks.Journal of Computational Physics, 426:109950, 2021

    2021

  9. [17]

    Thermodynamics-informed graph neural networks.arXiv preprint arXiv:2203.01874, 2022

    Quercus Hernandez, Alberto Badias, Francisco Chinesta, and Elias Cueto. Thermodynamics-informed graph neural networks.arXiv preprint arXiv:2203.01874, 2022

    2022 arXiv

  10. [18]

    Scaling description of generalization with number of parameters in deep learning.Journal of Statistical Mechanics: Theory and Experiment, 2020(2):023401, feb 2020

    Mario Geiger, Arthur Jacot, Stefano Spigler, Franck Gabriel, Levent Sagun, Stéphane d’Ascoli, Giulio Biroli, Clément Hongler, and Matthieu Wyart. Scaling description of generalization with number of parameters in deep learning.Journal of Statistical Mechanics: Theory and Exper...

    2020

  11. [19]

    Port-metriplectic neural networks: thermodynamics-informed machine learning of complex physical systems.Computational Mechanics, 72(3):553– 561, 2023

    Quercus Hernández, Alberto Badías, Francisco Chinesta, and Elías Cueto. Port-metriplectic neural networks: thermodynamics-informed machine learning of complex physical systems.Computational Mechanics, 72(3):553– 561, 2023. 17 Graph neural networks informed locally by thermodyn...

    2023

  12. [20]

    The graph neural network model.IEEE transactions on neural networks / a publication of the IEEE Neural Networks Council, 20:61–80, 01 2009

    Franco Scarselli, Marco Gori, Ah Tsoi, Markus Hagenbuchner, and Gabriele Monfardini. The graph neural network model.IEEE transactions on neural networks / a publication of the IEEE Neural Networks Council, 20:61–80, 01 2009

    2009

  13. [21]

    Schoenholz, Patrick F

    Justin Gilmer, Samuel S. Schoenholz, Patrick F. Riley, Oriol Vinyals, and George E. Dahl. Neural message passing for quantum chemistry. In Doina Precup and Yee Whye Teh, editors,Proceedings of the 34th International Conference on Machine Learning, volume 70 ofProceedings of Ma...

    2017

  14. [22]

    Iparraguirre, Iciar Alfaro, David Gonzalez, Francisco Chinesta, and Elias Cueto

    Alicia Tierz, Mikel M. Iparraguirre, Iciar Alfaro, David Gonzalez, Francisco Chinesta, and Elias Cueto. On the feasibility of foundational models for the simulation of physical phenomena, 2024

    2024

  15. [23]

    Michael Grieves and John Vickers.Digital Twin: Mitigating Unpredictable, Undesirable Emergent Behavior in Complex Systems, pages 85–113. 08 2017

    2017

  16. [24]

    Toward predictive digital twins via component-based reduced-order models and interpretable machine learning

    Michael Kapteyn, Karen Willcox, and David Knezevic. Toward predictive digital twins via component-based reduced-order models and interpretable machine learning. 01 2020

    2020

  17. [25]

    Digital twins that learn and correct themselves.International Journal for Numerical Methods in Engineering, 123(13):3034–3044, 2022

    Beatriz Moya, Alberto Badías, Icíar Alfaro, Francisco Chinesta, and Elías Cueto. Digital twins that learn and correct themselves.International Journal for Numerical Methods in Engineering, 123(13):3034–3044, 2022

    2022

  18. [26]

    Virtual, digital and hybrid twins: A new paradigm in data-based engineering and engineered data: F

    Francisco Chinesta, Elias Cueto, Emmanuelle Abisset-Chavanne, Jean Louis Duval, and Fouad El Khaldi. Virtual, digital and hybrid twins: A new paradigm in data-based engineering and engineered data: F. chinesta et al. Archives of computational methods in engineering, 27(1):105–...

    2020

  19. [27]

    Cognitive digital twin: An approach to improve the maintenance management.CIRP Journal of Manufacturing Science and Technology, 38:613–630, 2022

    Rosario Davide D’Amico, John Ahmet Erkoyuncu, Sri Addepalli, and Steve Penver. Cognitive digital twin: An approach to improve the maintenance management.CIRP Journal of Manufacturing Science and Technology, 38:613–630, 2022

    2022

  20. [28]

    Zero-shot learning - the good, the bad and the ugly

    Yongqin Xian, Bernt Schiele, and Zeynep Akata. Zero-shot learning - the good, the bad and the ugly. 03 2017

    2017

  21. [29]

    Digital twin (dt)-cyclegan: Enabling zero-shot sim-to-real transfer of visual grasping models.IEEE Robotics and Automation Letters, 8(5):2421–2428, 2023

    David Liu, Yuzhong Chen, and Zihao Wu. Digital twin (dt)-cyclegan: Enabling zero-shot sim-to-real transfer of visual grasping models.IEEE Robotics and Automation Letters, 8(5):2421–2428, 2023

    2023

  22. [30]

    Virtual, digital and hybrid twins: a new paradigm in data-based engineering and engineered data

    Francisco Chinesta, Elias Cueto, Emmanuelle Abisset-Chavanne, Jean Louis Duval, and Fouad El Khaldi. Virtual, digital and hybrid twins: a new paradigm in data-based engineering and engineered data. Technical report, 2018

    2018

  23. [31]

    Machine learning with data assimilation and uncertainty quantification for dynamical systems: a review.IEEE/CAA Journal of Automatica Sinica, 10(6):1361–1387, 2023

    Sibo Cheng, César Quilodrán-Casas, Said Ouala, Alban Farchi, Che Liu, Pierre Tandeo, Ronan Fablet, Didier Lucor, Bertrand Iooss, Julien Brajard, et al. Machine learning with data assimilation and uncertainty quantification for dynamical systems: a review.IEEE/CAA Journal of Au...

    2023

  24. [32]

    Digital twins that learn and correct themselves.International Journal for Numerical Methods in Engineering, 123(13):3034–3044, 2020

    Beatriz Moya, Alberto Badías, Icíar Alfaro, Francisco Chinesta, and Elías Cueto. Digital twins that learn and correct themselves.International Journal for Numerical Methods in Engineering, 123(13):3034–3044, 2020

    2020

  25. [33]

    A taxonomy of mixed reality visual displays.IEICE TRANSACTIONS on Information and Systems, 77(12):1321–1329, 1994

    Paul Milgram and Fumio Kishino. A taxonomy of mixed reality visual displays.IEICE TRANSACTIONS on Information and Systems, 77(12):1321–1329, 1994

    1994

  26. [34]

    A review on industrial augmented reality systems for the industry 4.0 shipyard.Ieee Access, 6:13358–13375, 2018

    Paula Fraga-Lamas, Tiago M Fernandez-Carames, Oscar Blanco-Novoa, and Miguel A Vilar-Montesinos. A review on industrial augmented reality systems for the industry 4.0 shipyard.Ieee Access, 6:13358–13375, 2018

    2018

  27. [35]

    Morph-dslam: Model order reduction for physics-based deformable slam.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(11):7764–7777, 2021

    Alberto Badias, Iciar Alfaro, David Gonzalez, Francisco Chinesta, and Elías Cueto. Morph-dslam: Model order reduction for physics-based deformable slam.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(11):7764–7777, 2021

    2021

  28. [36]

    U-net: Convolutional networks for biomedical image seg- mentation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image seg- mentation. InMedical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part...

    2015

  29. [37]

    Good features to track

    Jianbo Shi et al. Good features to track. In1994 Proceedings of IEEE conference on computer vision and pattern recognition, pages 593–600. IEEE, 1994

    1994

  30. [38]

    An iterative image registration technique with an application to stereo vision

    Bruce D Lucas and Takeo Kanade. An iterative image registration technique with an application to stereo vision. InIJCAI’81: 7th international joint conference on Artificial intelligence, volume 2, pages 674–679, 1981

    1981

  31. [39]

    Fast poisson disk sampling in arbitrary dimensions

    Robert Bridson. Fast poisson disk sampling in arbitrary dimensions. InSIGGRAPH sketches, page 22. ACM, 2007

    2007

  32. [40]

    A thermodynamics-informed active learning approach to perception and reasoning about fluids.Computational Mechanics, 72(3):577–591, 2023

    Beatriz Moya, Alberto Badías, David González, Francisco Chinesta, and Elías Cueto. A thermodynamics-informed active learning approach to perception and reasoning about fluids.Computational Mechanics, 72(3):577–591, 2023

    2023

  33. [41]

    The initialization of numerical models by a dynamic-initialization technique

    James E Hoke and Richard A Anthes. The initialization of numerical models by a dynamic-initialization technique. Monthly Weather Review, 104(12):1551–1556, 1976. 18 Graph neural networks informed locally by thermodynamicsA PREPRINT

    1976

  34. [42]

    SIAM, 2016

    Mark Asch, Marc Bocquet, and Maëlle Nodet.Data assimilation: methods, algorithms, and applications. SIAM, 2016

    2016

  35. [43]

    Learning skillful medium-range global weather forecasting.Science, 382(6677):1416–1421, 2023

    Remi Lam, Alvaro Sanchez-Gonzalez, Matthew Willson, Peter Wirnsberger, Meire Fortunato, Ferran Alet, Suman Ravuri, Timo Ewalds, Zach Eaton-Rosen, Weihua Hu, et al. Learning skillful medium-range global weather forecasting.Science, 382(6677):1416–1421, 2023. 19

    2023

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.