Pith. sign in

REVIEW 4 major objections 7 minor 43 references

Physics Encoded Blocks in Residual Neural Network Architectures for Digital Twin Models

T0 review · 4 major / 7 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read This paper proposes PERNN, an architecture that embeds known physics equations as fixed differentiable blocks inside a neural network and shows that the hybrid outperforms data-only and physics-regularized models on steering and CO2 flux…

desk verdict Clean modular architecture and two real demonstrations, but the abstract overclaims and the NEE physics block is fitted from the same data it constrains. read the letter →

arxiv 2411.11497 v2 pith:LQTFGYN4 submitted 2024-11-18 cs.LG cs.RO

Muhammad Saad Zia , Ashiq Anjum , Lu Liu , Anthony Conway , Anasol Pena Rios This is my paper · ORCID
classification cs.LGcs.RO
keywords physics-encodedneuralnetworksresiduallearningdigitaltwinsdifferentiablephysicsbehaviorcloningNetEcosystemExchangegap-fillingdataefficiency
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 tries to establish that a hybrid architecture called PERNN, made of fixed differentiable 'physics blocks' that implement known equations, small trainable 'learning blocks' that discover unmeasured intermediate variables, and 'residual blocks' that add corrections and keep gradients flowing, can model real systems when both physics knowledge and data are incomplete. The authors argue this targets the common middle ground between pure physics-based and pure data-driven modelling, delivering the interpretability and data efficiency of physics without needing a complete mechanistic description. They demonstrate it in two digital-twin settings: steering an autonomous vehicle in a simulated racing environment, and gap-filling Net Ecosystem Exchange, the net CO2 exchange between an ecosystem and the atmosphere, from eight years of flux-tower data. In both settings they report that PERNN outperforms conventional feed-forward networks, loss-regularized physics-informed networks, and, for CO2 gap-filling, Random Forest and XGBoost, while using less training data and fewer parameters. If the claim holds, digital twins for many partially understood systems could be built faster and with far less data.

What carries the argument

The load-bearing mechanism is the 'knowledge block' decomposition with a residual bypass. A physics block is a non-trainable computational graph representing known equations; a learning block is a small fully connected network that predicts the unknown intermediate variables the physics block needs; a residual block takes an intermediate feature vector from the learning block and predicts a correction $r$, so the final prediction is $\hat{a} = P(x, l) + r$. The residual path is what makes the whole architecture trainable: without it, gradients must flow through static physics operators, and the paper's PENN ablation shows that convergence fails. Two-phase training also does real work: the learning block is warm-started with heuristic labels for the intermediate variables, then the full model is trained end-to-end. In the NEE application, skip-connection layers inside the blocks are likewise reported as necessary for the best results.

What would settle it

Retrain PERNN on the same flux data with E0 and r_b_night estimated only from data strictly before each test gap; if the gap-filling advantage over Random Forest disappears, the original gains came from test-period information rather than from the physics block.

Watch

Extended reading notes

Core claim

The central claim is that a neural network can contain the known equations of a system as a fixed, differentiable computational graph and still be trained end-to-end on observational data. PERNN splits the model into three blocks: a learning block maps observed inputs to intermediate variables the physics needs but that are not measured; a physics block applies the known operators with no trainable weights; and a residual block adds a learned correction to the physics output. In the steering task the physics block is the pure pursuit geometry that turns lookahead distance and heading difference into a steering angle, while in the NEE task it is an Arrhenius-based ODE for ecosystem respiration. The paper argues that the residual correction is necessary for convergence because fixed physics operators otherwise disrupt gradient flow, and that the learning blocks make the model interpretable by exposing human-understandable intermediate variables. The reported result is that PERNN matches or beats much larger conventional networks with far fewer parameters and less data, and improves NEE gap-filling over Random Forest and XGBoost on both error and distributional metrics.

Load-bearing premise

The approach assumes the physics block is a faithful-enough skeleton of the real process; for NEE that means the Arrhenius respiration equation and its parameters E0 and r_b_night, estimated from the same flux-tower data by a partitioning algorithm, are unbiased and do not leak information from the test period.

Editorial extensions

If this is right

  • In the steering task, PERNN reaches a test mean absolute error comparable to a much larger fully connected network while using about 260 times fewer parameters and about five times less training data, and it produces smoother steering on unseen tracks.
  • Removing the residual block (the PENN variant) makes the physics-encoded model struggle to converge, so the residual correction is a required part of the architecture rather than an optional refinement.
  • For NEE gap-filling, PERNN improves on Random Forest by about 39 percent in Wasserstein distance and about 11 percent in KL divergence while also lowering mean absolute error, and it matches the shape of NEE distributions better than the other tested methods.
  • The intermediate variables discovered by the learning blocks (lookahead distance, heading difference, $E_0$, $r_{b,\text{night}}$, and $dT/dt$) are human-readable, so the model's decisions can be inspected.
  • The same three-block pattern applies to any differentiable partial physics model, which is the basis of the paper's claim that the approach is a generic framework for rapid digital-twin prototyping with scarce data.

Reading between the lines

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

  • Because the NEE experiment assumes GPP is zero, its results cover only nighttime conditions; applying the same physics-block idea to daytime NEE would require modelling or learning photosynthesis, a natural extension the paper leaves untested.
  • A stricter comparison would fix the amount of training data and vary it for every architecture, since the paper compares PERNN trained on six tracks against conventional networks trained on six, fourteen, and twenty-five tracks.
  • If the residual block can compensate for a deliberately wrong physics equation, then the method is best understood as a general inductive-bias architecture rather than as a claim specifically about pure pursuit or Arrhenius respiration.
  • The same decomposition could be applied to other partially known ODE and PDE models, where unobserved intermediate variables are usually the main obstacle to mechanistic modelling.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 7 minor

Summary. The paper introduces Physics Encoded Residual Neural Network (PERNN), a hybrid architecture in which a differentiable physics block implements known equations, learning blocks predict intermediate unobserved variables, and residual blocks add learned corrections to the physics output. The framework is applied to two digital-twin tasks: steering in the TORCS simulation using a pure-pursuit geometric model, and gap-filling of net ecosystem exchange (NEE) flux-tower data using an Arrhenius-type ODE for ecosystem respiration. The paper claims that PERNN outperforms conventional neural networks and state-of-the-art physics-informed methods, with better generalizability, lower data requirements, and reduced model complexity. The reported experiments compare PERNN against FCNN, PINN, PENN, Random Forest, and XGBoost, and include analyses of the discovered intermediate variables. The NEE evaluation is restricted to night-time data, and the steering results show that the largest FCNN baseline outperforms PERNN on the main driving metrics, which is in tension with the abstract's claims.

Significance. The modular idea of embedding fixed differentiable physics operators directly in the computation graph, with residual bypasses to stabilize gradient flow, is attractive and potentially transferable to other digital-twin settings. The paper ships publicly available code and presents two quite different application domains, with an interpretability analysis of the learned intermediate variables. If the claimed gains survive the corrections discussed below, the framework would be a useful contribution to physics-informed machine learning. However, the current evidence is weakened by claims that contradict the reported tables, a physics block whose parameters are estimated from the target data themselves, and an evaluation that only covers night-time NEE. The significance of the contribution therefore cannot be fully assessed until these load-bearing points are addressed.

major comments (4)
  1. [§5.1.3, Table 3] Table 3 directly contradicts the abstract's claim that the method 'outperforms conventional neural network approaches': the FCNN-large baseline has lower test MAE (0.01 rad vs 0.0214 rad) and a greater average distance (3803 m vs 3209 m) than PERNN. Section 5.1.3 itself acknowledges that PERNN only shows 'comparable MAE scores' to FCNN-large, but §6 restates the unqualified outperformance claim. The central empirical claim must be revised to 'PERNN matches or exceeds the small and medium FCNNs and does so with fewer parameters and less training data', or additional evidence must be supplied to support the original claim.
  2. [§4.2.4, Eqs. (13)–(14)] The NEE physics block is not an independent first-principles constraint in the reported experiment. The parameters E0 and r_b_night are not measured but are estimated by REddyProc from the same flux-tower record, and the Phase I and Phase II losses (Eqs. 13-14) supervise the learning block directly on those estimates. The manuscript does not state whether REddyProc was fit only on the 2012-2017 training years or whether its 15-day moving windows can include 2018-2019 test observations; the latter would mean the test metrics reflect data leakage. Even without leakage, the 'physics' term is a data-derived curve fit of the same distribution being gap-filled, so the experiment needs either a leakage-free protocol specification or a redesigned benchmark that treats the physics block as an external constraint.
  3. [Table 4 caption, §5.2.3] The NEE evaluation is explicitly limited to night-time data (Table 4 caption) and the model assumes GPP≈0 for Rg<20 W/m² (§4.2.2), so the experiments only support gap-filling of night-time ecosystem respiration, not general NEE gap-filling. The abstract and introduction claim general gap-filling and forecasting of NEE from flux tower data. The claims and the experimental scope must be aligned: either evaluate daytime data as well, or restrict the claims to night-time respiration gap-filling.
  4. [§5.2.3, Fig. 18] The comparison does not isolate the contribution of the physics block from other architectural choices. The PERNN model differs from FCNN and PINN not only by the presence of the physics block, but also by two-phase training, the autoregressive Euler update in Eq. (8), the auxiliary losses on intermediate variables, and the specific skip-connected learning and residual blocks. A matched-capacity baseline without the physics term, or an ablation that replaces the physics block with a trainable operator of the same intermediate variables, is needed before the Table 4 gains can be attributed to physics encoding rather than to the residual capacity or the training procedure.
minor comments (7)
  1. [§4.1.1] The manuscript refers to 'Equation (4.1.1)' but no equation with that number is defined; the reference-point definition z=(l, θ_target) is not numbered. In addition, the pure-pursuit derivation is cited to reference [6], which is the Hamilton-Jacobi/Lax-Oleinik paper, not a pure-pursuit reference.
  2. [Table 2] Table 2 lists Tsoil2 at a depth of 0.05 m, identical to Tsoil1; this appears to be a typo for a different depth and should be corrected.
  3. [§4.2.4] The text says the physics block computes dNEE/dt 'based on Equation 10', but the displayed formula is Equation (9); the equation numbering in this section should be checked throughout.
  4. [Algorithm 1] Algorithm 1, line 27, refers to 'learning (L), learning (R) and learning (P) blocks'; it should say learning, residual, and physics blocks.
  5. [Table 3, §5.1.3] The text states that PERNN has approximately 260 times fewer parameters than FCNN-large, but the reported counts give a factor of approximately 1,368 (7,676,170 / 5,611); either the parameter count or the factor needs correction.
  6. [Figure 19] The fourth column of Figure 19 is labeled a quarter but the dates span 2019-01-01 to 2019-12-30, which is a full year; the label or the experimental window should be corrected.
  7. [Tables 3 and 4] No multiple-seed or confidence-interval information is reported for the steering or NEE results; since some differences are small (e.g., PERNN MAE 0.866 vs RF 0.901 in Table 4), please report variability over repeated runs.

Circularity Check

0 steps flagged · score 0.0 of 10

No by-construction circularity: NEE parameters are calibrated on training data and all benchmarks are external; remaining issues are leakage risk and claim overreach, not circularity.

full rationale

The derivation chain is self-contained in the sense required for circularity analysis. In the steering case, the pure-pursuit physics block (Eq. 2) is an independent kinematic formula, the learning block outputs lookahead/heading that are trained first on a hand-defined heuristic and then end-to-end on expert steering, and the residual block is a trainable correction; no prediction is defined as the fitted parameter itself. In the NEE case, the Arrhenius/Lloyd-Taylor respiration equation (Eq. 4) and its chain-rule derivative (Eq. 5) are standard external models, and the paper states that E0 and r_b_night are estimated by REddyProc from the flux record; although this means the 'physics' is calibrated rather than first-principles, the learning block is trained on those estimates and the full model is evaluated on held-out years 2018-2019, so the NEE forecast does not reduce by construction to the REddyProc fit. The paper does not specify whether REddyProc ran on the full 2012-2019 record with 15-day windows that could reach into the test period; that is a potential leakage/validity threat, not a demonstrated circularity. The paper also concedes the NEE physics model is an over-simplification, and Table 3 shows FCNN-large outperforms PERNN on test MAE and distance, contradicting the abstract's unconditional 'outperforms' wording; these are correctness and framing issues. No self-citations are load-bearing, and no uniqueness theorem is imported from the authors. Verdict: no significant circularity.

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

The central claim rests on two physics models borrowed from prior literature (pure pursuit and Lloyd-Taylor respiration), both with parameters that are either hand-set or estimated from the same data. The learning block's latent variables are effectively fitted values. No entirely new entities are postulated.

free parameters (5)
  • E0 (temperature sensitivity) = estimated by REddyProc, range [50,400]
    Estimated from the same flux data via REddyProc; used as ground truth for the learning block and required by the physics block.
  • r_b_night (base respiration) = estimated by REddyProc, > 0
    Also estimated from the flux data; used as ground truth in training and in the Reco term of the physics block.
  • dT/dt (rate of change of air temperature) = computed via finite differences from Tair
    The learning block is trained to predict this data-derived quantity, which is not a known physical law.
  • Lookahead distance l (steering) = learned by network
    A latent variable predicted by the learning block in the steering case; no independent measurement, fitted to match the expert through the pure pursuit formula.
  • Heading difference theta_target (steering) = learned by network
    A latent variable predicted by the learning block; fitted to match the expert steering through the physics block.
assumptions (5)
  • domain assumption Pure pursuit kinematic model assumptions (rigid body on a circle, kinematic bicycle model)
    Section 4.1.1 derives delta = arctan(2L sin(theta) / l) using these assumptions.
  • domain assumption Arrhenius-type respiration model (Lloyd-Taylor)
    Equation 4, cited from references [35,36]; assumed to describe ecosystem respiration.
  • domain assumption GPP is approximately 0 for Rg < 20 W/m2
    Section 4.2.2 states this, restricting the NEE ODE to night-time conditions.
  • domain assumption The physics models are differentiable
    Section 3 assumes differentiability for end-to-end gradient training.
  • ad hoc to paper Heuristic warm-start labeling for steering (argmax of range finder sensors)
    Section 4.1.3 defines the warm-start labels for lookahead distance and heading difference; a hand-crafted rule specific to this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Physics Encoded Blocks in Residual Neural Network Architectures for Digital Twin Models." pith.science (2026). https://pith.science/paper/LQTFGYN4

@misc{pith2026241111497,
  author       = {Pith},
  title        = {Pith review of: Physics Encoded Blocks in Residual Neural Network Architectures for Digital Twin Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LQTFGYN4}},
  note         = {Machine review of arXiv:2411.11497}
}
read the original abstract

Physics Informed Machine Learning has emerged as a popular approach for modeling and simulation in digital twins, enabling the generation of accurate models of processes and behaviors in real-world systems. However, existing methods either rely on simple loss regularizations that offer limited physics integration or employ highly specialized architectures that are difficult to generalize across diverse physical systems. This paper presents a generic approach based on a novel physics-encoded residual neural network (PERNN) architecture that seamlessly combines data-driven and physics-based analytical models to overcome these limitations. Our method integrates differentiable physics blocks-implementing mathematical operators from physics-based models with feed-forward learning blocks, while intermediate residual blocks ensure stable gradient flow during training. Consequently, the model naturally adheres to the underlying physical principles even when prior physics knowledge is incomplete, thereby improving generalizability with low data requirements and reduced model complexity. We investigate our approach in two application domains. The first is a steering model for autonomous vehicles in a simulation environment, and the second is a digital twin for climate modeling using an ordinary differential equation (ODE)-based model of Net Ecosystem Exchange (NEE) to enable gap-filling in flux tower data. In both cases, our method outperforms conventional neural network approaches as well as state-of-the-art Physics Informed Machine Learning methods.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 37 canonical work pages

  1. [1]

    Nature Computational Science 1, 166–168 (2021) https://doi.org/10.1038/s43588-021-00040-z

    Willcox, K.E., Ghattas, O., Heimbach, P.: The imperative of physics-based mod- eling and inverse theory in computational science. Nature Computational Science 1, 166–168 (2021) https://doi.org/10.1038/s43588-021-00040-z

    doi:10.1038/s43588-021-00040-z 2021

  2. [2]

    Journal of Manufacturing Systems63, 381–391 (2022) https://doi.org/10.1016/j.jmsy.2022

    Wang, J., Li, Y., Gao, R., Zhang, F.: Hybrid physics-based and data-driven mod- els for smart manufacturing: Modelling, simulation, and explainability. Journal of Manufacturing Systems63, 381–391 (2022) https://doi.org/10.1016/j.jmsy.2022. 04.004

    work page doi:10.1016/j.jmsy.2022 2022

  3. [3]

    IEEE Access8, 21980–22012 (2020) https://doi

    Rasheed, A., San, O., Kvamsdal, T.: Digital twin: Values, challenges and enablers from a modeling perspective. IEEE Access8, 21980–22012 (2020) https://doi. org/10.1109/ACCESS.2020.2970143

    work page arXiv 2020

  4. [4]

    Structural and Multidisciplinary Optimization65, 354 (2022) https://doi.org/10.1007/ s00158-022-03425-4

    Thelen, A., Zhang, X., Fink, O.,et al.: A comprehensive review of digi- tal twin — part 1: modeling and twinning enabling technologies. Structural and Multidisciplinary Optimization65, 354 (2022) https://doi.org/10.1007/ s00158-022-03425-4

    work page 2022

  5. [5]

    Nature Machine Intelligence3, 218–229 (2021)

    Lu, L., Jin, P., Pang, G., Zhang, Z., Karniadakis, G.E.: Learning nonlinear oper- ators via deeponet based on the universal approximation theorem of operators. Nature Machine Intelligence3, 218–229 (2021)

    work page 2021

  6. [6]

    Journal of Computational Physics425, 109907 (2021)

    Darbon, J., Meng, T.: On some neural network architectures that can represent 44 viscosity solutions of certain high dimensional hamilton-jacobi partial differential equations. Journal of Computational Physics425, 109907 (2021)

    work page 2021

  7. [7]

    In: Proceedings of the IEEE International Conference on Robotics and Automation (ICRA) (2019)

    Li, Y., Wu, J., Zhu, J.-Y., Tenenbaum, J.B., Torralba, A., Tedrake, R.: Propaga- tion networks for model-based control under partial observation. In: Proceedings of the IEEE International Conference on Robotics and Automation (ICRA) (2019)

    work page 2019

  8. [8]

    In: Proceedings of the IEEE International Conference on Robotics and Automation (ICRA) (2018)

    Toussaint, M., Allen, K.R., Smith, K.A., Tenenbaum, J.B.: Differentiable physics and stable modes for tool-use and manipulation planning. In: Proceedings of the IEEE International Conference on Robotics and Automation (ICRA) (2018)

    work page 2018

Show all 43 references
  1. [9]

    In: Proceedings of the International Conference on Machine Learning (ICML) (2018)

    Sanchez-Gonzalez, A., Heess, N., Springenberg, J.T., Merel, J., Riedmiller, M., Hadsell, R., Battaglia, P.: Graph networks as learnable physics engines for infer- ence and control. In: Proceedings of the International Conference on Machine Learning (ICML) (2018)

    2018

  2. [10]

    In: Advances in Neural Information Processing Systems (NeurIPS) (2022)

    Seo, J., Smith, B.D., Greydanus, S., Saha, S., Paduraru, C., Wan, C., Bansal, A., Jain, G., Peng, A., Narvekar, S., Pezeshki, G., Brevdo, E., Moore, C.X., Patel, S., Wang, H., Merel, J., Tunyasuvunakool, K., Heess, N., Hadsell, R., Varma, S.: Learning to optimize complex real-...

    2022

  3. [11]

    In: NeurIPS Workshop on Machine Learning for Physics and the Physical Sciences (ML4PS) (2019)

    Stewart, R., Ermon, S.: How to train your differentiable physical simulator. In: NeurIPS Workshop on Machine Learning for Physics and the Physical Sciences (ML4PS) (2019)

    2019

  4. [12]

    In: Advances in Neural Information Processing Systems (NeurIPS) (2019)

    Greydanus, S., Dzamba, M., Yosinski, J.: Hamiltonian neural networks. In: Advances in Neural Information Processing Systems (NeurIPS) (2019)

    2019

  5. [13]

    Journal of Computational Physics378, 686–707 (2019)

    Raissi, M., Perdikaris, P., Karniadakis, G.E.: Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics378, 686–707 (2019)

    2019

  6. [14]

    In: International Conference on Learning Represen- tations (ICLR) (2020)

    Cranmer, M., Greydanus, S., Hoyer, S., Battaglia, P., Spergel, D., Ho, S.: Lagrangian neural networks. In: International Conference on Learning Represen- tations (ICLR) (2020)

    2020

  7. [15]

    In: Advances in Neural Information Processing Systems (NeurIPS) (2018)

    Mrowca, D., Zhuang, C., Wang, E., Haber, N., Fei-Fei, L.F., Tenenbaum, J., Yamins, D.: Flexible neural representation for physics prediction. In: Advances in Neural Information Processing Systems (NeurIPS) (2018)

    2018

  8. [16]

    In: IEEE/RSJ International 45 Conference on Intelligent Robots and Systems (IROS) (2018)

    Ajay, A., Wu, J., Fazeli, N., Bauza, M., Kaelbling, L.P., Tenenbaum, J.B., Rodriguez, A.: Augmenting physical simulators with stochastic neural net- works: Case study of planar pushing and bouncing. In: IEEE/RSJ International 45 Conference on Intelligent Robots and Systems (IR...

    2018

  9. [17]

    In: Advances in Neural Information Processing Systems (NeurIPS) (2021)

    Mohan, A., Chai, H., Khojasteh, M.J., Dey, D., Joshi, K.: Cdpn: Coordinated differentiable physics networks for interacting rigid bodies. In: Advances in Neural Information Processing Systems (NeurIPS) (2021)

    2021

  10. [18]

    Physics of Fluids33, 027104 (2021)

    Kashefi, A., Rempe, D., Guibas, L.J.: A point-cloud deep learning framework for prediction of fluid flow fields on irregular geometries. Physics of Fluids33, 027104 (2021)

    2021

  11. [19]

    In: International Conference on Learning Representations (ICLR) (2021)

    Li, Z.,et al.: Fourier neural operator for parametric partial differential equations. In: International Conference on Learning Representations (ICLR) (2021)

    2021

  12. [20]

    IEEE Transactions on Neural Networks 9, 987–1000 (1998)

    Lagaris, I.E., Likas, A., Fotiadis, D.I.: Artificial neural networks for solving ordi- nary and partial differential equations. IEEE Transactions on Neural Networks 9, 987–1000 (1998)

    1998

  13. [21]

    Journal of Computational Physics394, 56–81 (2019)

    Zhu, Y., Zabaras, N., Koutsourelakis, P.S., Perdikaris, P.: Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quan- tification without labeled data. Journal of Computational Physics394, 56–81 (2019)

    2019

  14. [22]

    Journal of Computational Physics 403, 109056 (2020)

    Geneva, N., Zabaras, N.: Modeling the dynamics of pde systems with physics- constrained deep auto-regressive networks. Journal of Computational Physics 403, 109056 (2020)

    2020

  15. [23]

    Journal of Computational Physics 406, 109209 (2020)

    Wu, J.L.,et al.: Enforcing statistical constraints in generative adversarial net- works for modeling chaotic dynamical systems. Journal of Computational Physics 406, 109209 (2020)

    2020

  16. [24]

    Journal of Franklin Institute346, 898–913 (2009)

    Beidokhti, R.S., Malek, A.: Solving initial-boundary value problems for systems of partial differential equations using neural networks and optimization techniques. Journal of Franklin Institute346, 898–913 (2009)

    2009

  17. [25]

    Communications in Computational Physics28, 2139–2157 (2020)

    Wang, B., Zhang, W., Cai, W.: Multi-scale deep neural network (mscalednn) methods for oscillatory stokes flows in complex domains. Communications in Computational Physics28, 2139–2157 (2020)

    2020

  18. [26]

    SIAM Journal on Scientific Computing42, 3285–3312 (2020)

    Cai, W., Li, X., Liu, L.: A phase shift deep neural network for high frequency approximation and wave problems. SIAM Journal on Scientific Computing42, 3285–3312 (2020)

    2020

  19. [27]

    In: Workshop on Integration of Deep Neural Models and Differential Equations at ICLR (2020)

    Miles, C., Sam, G., Stephan, H., Peter, B., David, S., Shirley, H.: Lagrangian neu- ral networks. In: Workshop on Integration of Deep Neural Models and Differential Equations at ICLR (2020)

    2020

  20. [28]

    Frontiers in Neurorobotics13, 42 (2019)

    Degrave, J., Hermans, M., Dambre, J., Wyffels, F.: A differentiable physics engine 46 for deep learning in robotics. Frontiers in Neurorobotics13, 42 (2019)

    2019

  21. [29]

    arXiv preprint arXiv:2001.07457 (2020)

    Holl, P., Koltun, V., Thuerey, N.: Learning to control PDEs with differentiable physics. arXiv preprint arXiv:2001.07457 (2020)

    2020 arXiv

  22. [30]

    In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp

    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770–778 (2016). https://doi.org/10.1109/CVPR.2016.90

    2016 doi

  23. [31]

    Available online (2005)

    Espi´ e, E., Guionneau, C., Wymann, B., Dimitrakakis, C., Coulom, R., Sumner, A.: TORCS, The Open Racing Car Simulator. Available online (2005)

    2005

  24. [32]

    Agricultural and Forest Meteorology 314, 108777 (2022)

    Zhu, S., Clement, R., McCalmont, J., Davies, C., Hill, T.: Stable gap-filling for longer eddy covariance data gaps: A globally validated machine-learning approach for carbon dioxide, water, and energy fluxes. Agricultural and Forest Meteorology 314, 108777 (2022)

    2022

  25. [33]

    Global Change Biology11, 1424–1439 (2005)

    Reichstein, M., Falge, E., Baldocchi, D.,et al.: On the separation of net ecosys- tem exchange into assimilation and ecosystem respiration: review and improved algorithm. Global Change Biology11, 1424–1439 (2005)

    2005

  26. [34]

    Agricultural and Forest Meteorology147, 209–232 (2007)

    Moffat, A.M., Papale, D., Reichstein, M.,et al.: Comprehensive comparison of gap-filling techniques for eddy covariance net carbon fluxes. Agricultural and Forest Meteorology147, 209–232 (2007)

    2007

  27. [35]

    Global Change Biology16(1), 187–208 (2010)

    Lasslop, G., Reichstein, M., Papale, D., Richardson, A., Arneth, A., Barr, A., Stoy, P., Wohlfahrt, G.: Separation of net ecosystem exchange into assimilation and respiration using a light response curve approach: critical issues and global evaluation. Global Change Biology16(...

    2010

  28. [36]

    Functional Ecology8(3), 315–323 (1994)

    Lloyd, J., Taylor, J.A.: On the temperature dependence of soil respiration. Functional Ecology8(3), 315–323 (1994). Accessed 2024-07-02

    1994

  29. [37]

    Applied Mathematics and Computation196(2), 686–704 (2008) https://doi.org/10.1016/j.amc.2007.07.004

    White, L., Luo, Y.: Modeling and inversion of net ecological exchange data using an ito stochastic differential equation approach. Applied Mathematics and Computation196(2), 686–704 (2008) https://doi.org/10.1016/j.amc.2007.07.004

    2008 doi

  30. [38]

    IEEE Access7, 127356–127367 (2019)

    Tao, F., Qi, Q., Liu, A., Kusiak, A.: Digital twin-driven smart manufacturing: Connotation, reference model, applications and research issues. IEEE Access7, 127356–127367 (2019)

    2019

  31. [39]

    Renewable and Sustainable Energy Reviews135, 110275 (2020)

    Jones, D., Smith, J., Brown, A.: Digital twin technology for energy and climate management. Renewable and Sustainable Energy Reviews135, 110275 (2020)

    2020

  32. [40]

    NERC Environmental Information Data Centre (2020)

    Cumming, A.M.J., Newman, T.R., Benson, S.J., Balzter, H., Evans, C., Jones, D., Kaduk, J., Morrison, R.D., Page, S.E.: Eddy Covariance measurements of carbon dioxide, energy and water flux at an intensively cultivated lowland deep peat soil, 47 East Anglia, UK, 2012 to 2020. N...

    2020 doi

  33. [41]

    Nature Ecology & Evolution3, 407–415 (2019) https://doi.org/10.1038/s41559-019-0809-2

    Keenan, T.F., Migliavacca, M., Papale, D.,et al.: Widespread inhibition of daytime ecosystem respiration. Nature Ecology & Evolution3, 407–415 (2019) https://doi.org/10.1038/s41559-019-0809-2

    2019 doi

  34. [42]

    Biogeosciences15(16), 5015–5030 (2018) https://doi.org/10.5194/bg-15-5015-2018

    Wutzler, T., Lucas-Moffat, A., Migliavacca, M., Knauer, J., Sickel, K., ˇSigut, L., Menzer, O., Reichstein, M.: Basic and extensible post-processing of eddy covariance flux data with reddyproc. Biogeosciences15(16), 5015–5030 (2018) https://doi.org/10.5194/bg-15-5015-2018

    2018 doi

  35. [43]

    IEEE Transactions on Computational Intelligence and AI in Games9(1), 54–66 (2017) https://doi.org/ 10.1109/TCIAIG.2016.2565661 48

    Bonyadi, M.R., Michalewicz, Z., Nallaperuma, S., Neumann, F.: Ahura: A heuristic-based racer for the open racing car simulator. IEEE Transactions on Computational Intelligence and AI in Games9(1), 54–66 (2017) https://doi.org/ 10.1109/TCIAIG.2016.2565661 48

    2017

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.