Pith. sign in

REVIEW 3 major objections 6 minor 22 references

History-informed Lagrangian Neural Networks

T0 review · 3 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read The paper claims that conditioning a Lagrangian network's mass, potential, and damping on a latent context inferred from position history beats fixed physics-guided and black-box models at long-horizon position-only forecasting.

desk verdict A genuinely new context-conditioned Lagrangian architecture with solid pendulum results, but the paper's 'position-only' framing is undercut by full-state supervision in the training loss. read the letter →

arxiv 2608.13215 v1 pith:TACA5354 submitted 2026-08-13 cs.LG

Tianshuo Zhang , Xianglei Xing , Wenzhe Zhai , Jia Gao , He Cao This is my paper · ORCID
classification cs.LG
keywords physicalforecastingLagrangianneuralnetworkspartialobservationssystemidentificationstructureddynamicsrecurrentencoderdifferentiableRK4rolloutenergyconsistency
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 sets out to solve long-horizon forecasting of mechanical systems when only positions have been observed. It claims that a short position history contains enough information to reconstruct the missing velocity and to identify trajectory-specific physical parameters, and that a Lagrangian neural network whose mass, potential, and damping are conditioned on a latent context drawn from that history can outperform models with fixed global dynamics. On conservative, damped, and variable-parameter pendulums, the proposed History-informed Lagrangian Neural Network (HiLNN) reduces average rollout MSE from 0.279 to 0.103 on the standard pendulum and from 0.894 to 0.061 on the variable-parameter pendulum, with lower final-step and energy errors. A sympathetic reader would take the paper to establish that structured, history-adaptive dynamics is a viable alternative to black-box sequence models for partially observed mechanical systems.

What carries the argument

The load-bearing object is the latent context $z_t$, a 32-dimensional vector produced by a gated recurrent unit (GRU) encoder over the position history $H^q_t = \{q_{t-L+1}, \dots, q_t\}$. It plays two roles: it reconstructs the unobserved initial velocity via $\hat{\dot q}_t = \mathrm{MLP}_v([q_t, z_t])$, and it conditions the structured dynamics—mass matrix $M_\theta(q,z_t)$, potential $V_\phi(q,z_t)$, and optional damping $D_\psi(q,z_t)$—inside a context-conditioned Lagrangian $L_\theta = T - V$ whose accelerations come from the Euler–Lagrange equation with possibly non-conservative forces. The context stays fixed while a differentiable RK4 integrator rolls the state forward, so the encoder, velocity head, and dynamics module are trained end-to-end under rollout loss, initial-velocity loss, and energy-consistency loss.

What would settle it

Run the published training objective on a mechanical benchmark where only position data are available—drop the velocity-supervision and energy terms, or replace them with finite-difference and conservation residuals—and measure 32-step rollout MSE against the fixed LNN baseline; if HiLNN no longer beats LNN or its rollout error grows steeply, the claim that position history alone suffices for long-horizon structured forecasting is refuted.

Watch

Extended reading notes

Core claim

The central discovery is that a latent context extracted from past positions can stand in for the unobserved state and system identity inside a structured Lagrangian model. A recurrent encoder reads a length-8 position history and outputs a context vector; that vector is used to predict the missing initial velocity, to build a positive mass matrix and potential energy, and, for dissipative systems, to set a non-negative damping coefficient. The resulting acceleration comes from the Euler–Lagrange equation, and a differentiable RK4 rollout propagates the state 32 steps ahead with the context held fixed. Trained with multi-step trajectory supervision plus energy-consistency regularization, HiLNN reports the lowest average MSE, final-step MSE, and energy MSE across conservative, damped, and variable-parameter pendulum benchmarks, reducing the variable-parameter final-step error from 2.295 (LNN-multistep) to 0.193.

Load-bearing premise

The load-bearing premise is that ground-truth velocities and energies are available for every rollout step during training, because the proposed loss functions require those labels even though the task is described as position-only.

Editorial extensions

If this is right

  • Position-only sensing becomes enough to run structured, physics-constrained forecasting: any system whose coordinates are recorded but whose velocities and parameters are unknown can in principle use the same history-conditioned Lagrangian setup.
  • Conditioning mass, potential, and damping on a per-trajectory context removes the fixed-global-model limitation of standard LNNs, so one learned model can serve heterogeneous systems with different physical parameters.
  • Differentiable RK4 rollout with full backpropagation is part of the recipe: the ablation shows detached Euler training fails, while full-BPTT RK4 training stabilizes long-horizon errors.
  • Energy regularization buys physical consistency without sacrificing accuracy at the right weight, and over-regularizing hurts both, so there is a usable accuracy–energy trade-off.
  • The reported reductions (e.g., variable-parameter average MSE from 0.894 to 0.061 and energy MSE from 19.516 to 1.759) imply the method is not just a short-horizon fitter but controls long-horizon error accumulation.

Reading between the lines

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

  • Because the latent context is trained only through dynamics, it may function as a learned parameter-identification vector; a direct test would fix the true mass, length, and damping of test trajectories and check whether the inferred context varies smoothly along those parameter axes.
  • The architecture transfers naturally to higher-dimensional or under-actuated systems (double pendulum, cart-pole, soft bodies); if the recurrent encoder can infer multiple hidden parameters from position history there, the paper's mechanism would be a general tool rather than a pendulum result.
  • The training objective still requires ground-truth velocity and energy labels; a strictly position-only variant could replace the velocity-supervision term with finite-difference or shooting-based velocity targets and the energy term with a conservation residual, and it would be informative to see how far accuracy falls.
  • A competing simpler design—finite-difference velocity estimates plus a context vector from a short window fed into an LNN—might achieve similar gains; comparing against such a baseline would isolate how much of the benefit comes from the learned recurrent encoder rather than from merely having more input features.
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

3 major / 6 minor

Summary. The paper proposes History-informed Lagrangian Neural Networks (HiLNN) for long-horizon forecasting of mechanical systems from position histories. A recurrent encoder (GRU) maps a length-L position history into a latent context z_t; this context is used both to infer the missing initial velocity and to condition the mass matrix, potential energy, and optional damping term of a structured Lagrangian model. The model is trained end-to-end with differentiable RK4 rollout, multi-step trajectory supervision, and energy-consistency regularization. Experiments on a fixed conservative pendulum, a fixed damped pendulum, and a variable-parameter pendulum report that HiLNN outperforms LNN, HNN, Neural ODE, and MLP baselines on trajectory MSE, final-step error, and energy error. The paper claims this demonstrates superior long-horizon prediction from position-only observations under conservative, dissipative, and heterogeneous dynamics.

Significance. If the empirical results hold, HiLNN is a useful step toward physics-guided forecasting under partial observability: it combines a structured Lagrangian prior with trajectory-adapted latent conditioning, and the ablation study supports the value of full rollout backpropagation and high-order integration. The paper ships a public code repository, which is a concrete asset for reproducibility. The main benchmark claim is an honest comparison of trajectory prediction accuracy, and the latent context is inferred from observed past data while the target is future evolution, so there is no obvious circularity in the central evaluation. However, the significance is currently limited by three issues: the training objective uses full-state supervision despite the 'position-only' framing, all quantitative claims rest on single runs without error bars, and the baseline protocol for handling unobserved velocities is unspecified. These issues need to be resolved before the paper's central claim can be accepted at face value.

major comments (3)
  1. [§3.1, §3.6, Eqs. (20)–(23)] The paper frames the task as 'position-only observations' (Eq. 1), but the training objective is not position-only. L_roll in Eq. (20) includes a velocity MSE term λ_qdot ||qdot_hat − qdot||² at every rollout step; L_v0 in Eq. (21) requires the exact initial velocity; and L_E in Eq. (22) requires ground-truth mechanical energy at every step. Thus velocities and energies are used as supervision throughout training, and only test-time inference is position-only. The ablation in Table 8 removes only λ_v0, not the trajectory velocity or energy terms, so the reported results do not show what happens when velocities are truly unavailable. Please restate the claim as 'position-input with full-state supervision during training' and either add an experiment with λ_qdot = 0 and without L_v0/L_E, or explicitly justify why such supervision is available in the intended applications.
  2. [Tables 3–7] All quantitative results are reported as single point estimates with no error bars, no seed counts, and no significance tests. Deep-learning training is stochastic, and some headline comparisons are close (for example, Table 4 Step 1 shows HiLNN's one-step MSE of 0.061 is much worse than LNN's 2.17×10⁻³, while the claimed advantage appears at longer horizons). Without multiple random seeds and a measure of variance, the reader cannot assess whether the reported improvements are robust or due to a particular run. Please report mean ± standard deviation over at least five seeds and provide a statistical test for the main MSE and Energy MSE comparisons.
  3. [§4.1 Evaluation protocol] The evaluation section states that 'all methods follow the same position-only forecasting protocol,' but LNN, HNN, and Neural ODE require the full state (q, qdot) as input. The paper does not specify how these baselines obtain the initial velocity at test time: finite differences, ground-truth velocity, or a learned estimator. This choice directly affects the comparison, because the baselines' accuracy depends on the quality of the velocity input. Please specify the exact baseline input protocol and, if ground-truth or finite-difference velocities are supplied, state this explicitly and discuss how it affects the fairness of the comparison against HiLNN, which must infer velocity from history.
minor comments (6)
  1. [Eq. (22)] The energy loss contains stray vertical bars in the notation '||E(hat_q, hat_qdot) − E(q, qdot)||²_2'; please use a single consistent norm notation.
  2. [§4.1 and Table 3] The baselines 'MLP-one-step' and 'LNN-multistep' are named in the tables but not described in the setup section. Please add one or two sentences specifying their architecture, input, and training loss.
  3. [Table 8] The 'Init. velocity' rows label the default weight as 'λ_v0 = 0.1' but do not mark it as the default configuration, and the row 'λ_v0 = 0.1' duplicates the default setting rather than acting as a distinct ablation. Please relabel to make the default clearly identifiable and clarify that the comparison is between λ_v0 = 0.1 and λ_v0 = 0.
  4. [§4.4] The paragraph after Table 7 contains two nearly identical concluding sentences ('These results suggest that...' and 'These results indicate that...'); please condense to avoid redundancy.
  5. [Table 4] HiLNN's Step-1 MSE (0.061) is substantially higher than LNN's (2.17×10⁻³). The text should acknowledge this explicitly and explain why the multi-step roll-out error is the appropriate criterion for the paper's claim.
  6. [§4.1 and Conclusion] The paper claims 'superior long-term prediction accuracy' and 'state-of-the-art baselines,' but the experiments are limited to three pendulum systems. Please scope the conclusion to these benchmarks or add experiments on higher-dimensional or more complex mechanical systems.

Circularity Check

0 steps flagged · score 1.0 of 10

No circular derivation found: HiLNN's prediction target is an external simulated trajectory, the latent context is inferred rather than fitted to the forecast, and the self-citations are non-load-bearing.

full rationale

The central claim (Section 4.2, Table 3) is an empirical comparison of 32-step open-loop rollouts against simulated pendulum trajectories. The prediction target is the future trajectory, which is not used to construct the model's input: the history encoder consumes only H_q^t = {q_{t-L+1}, ..., q_t} (Eq. 3), and the latent context z_t is fixed during rollout. No parameter is fitted to the test quantity and then reported as a prediction. The latent context is a learned representation supervised on training trajectories and evaluated on held-out test windows from the same simulator; this is standard amortized system identification, not circular reasoning. The energy regularization (Eq. 22) uses ground-truth mechanical energy as supervision, so Energy MSE is partly optimized during training, but it is evaluated on unseen test rollouts and is not the quantity being predicted. The self-citations ([9] and [21]) appear only in a broad motivation list and a related-work sentence ('[14,21]'), and neither is load-bearing for the empirical conclusions. The one substantive weakness is that Eqs. (20)-(22) require ground-truth velocities and energies at every training step and for the initial state, so the method is position-input, full-state-supervised; this narrows the 'position-only observations' claim (Section 3.1) but is an applicability limitation, not a circularity.

Assumptions & free parameters 4 free parameters · 4 assumptions · 1 invented entities

The ledger includes hand-set hyperparameters that influence results, domain assumptions that restrict the class of systems, and the latent context as an internal construct. Neural network weights themselves are treated as fitted model parameters and are not enumerated; the listed items are the macroscopic choices on which the paper's contribution depends.

free parameters (4)
  • Energy-regularization weight lambda_E = 0.01
    Selected via ablation in Table 8; lambda_E=0.001 gives lower Final MSE@32 (0.253), so the default is a manual accuracy-energy compromise.
  • Initial-velocity loss weight lambda_v0 = 0.1
    Eq. 24; ablation shows removing it increases initial velocity error from 0.115 to 0.245.
  • History length L = 8
    Hard-coded in Table 2; Section 4.5 announces a history-length ablation that is not reported.
  • Latent context dimension d_z = 32
    Set in Table 2; no sensitivity analysis is provided.
assumptions (4)
  • standard math The learned acceleration is defined by the Euler-Lagrange equation with generalized force Q: d/dt(dL/dqdot) - dL/dq = Q (Eq. 12).
    Basis for the dynamics module in Section 3.4.
  • domain assumption The mass matrix is positive and diagonal (scalar in 1D), with kinetic energy quadratic in velocity and potential independent of velocity.
    Restricts representable systems; common for rigid-body pendulums but not for general mechanical systems (Section 3.4, Eqs. 9-11).
  • domain assumption An 8-sample position history contains enough information to infer initial velocity and trajectory-specific parameters such as mass, length, and friction.
    No identifiability analysis is given; the paper appeals to delay-coordinate reconstruction (Section 3.3).
  • domain assumption Ground-truth velocities and energies are available during training to compute L_roll, L_v0, and L_E.
    Eqs. 20-22 require velocity and energy labels even though the setting is called position-only; this limits the training regime to datasets with full-state supervision.
invented entities (1)
  • Latent context z_t (32-dimensional)
    purpose: Encodes hidden initial velocity cues and trajectory-specific dynamics; conditions mass, potential, and damping in the Lagrangian; fixed during rollout.
    Internal latent variable learned end-to-end; it predicts no external measurable quantity on its own, and its validity is only tested through downstream rollout accuracy.

how reviews work

0 comments
Cite this review

Pith. "Pith review of History-informed Lagrangian Neural Networks." pith.science (2026). https://pith.science/paper/TACA5354

@misc{pith2026260813215,
  author       = {Pith},
  title        = {Pith review of: History-informed Lagrangian Neural Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TACA5354}},
  note         = {Machine review of arXiv:2608.13215}
}
read the original abstract

Forecasting the long-horizon evolution of mechanical systems from position-only observations is a pivotal yet difficult task, as hidden velocities and trajectory-specific physical properties must be inferred simultaneously. Although physics-guided neural networks like Lagrangian Neural Networks (LNNs) guarantee physical plausibility, they generally require complete state inputs and lack adaptability to changing system parameters. To break these limitations, we introduce History-informed Lagrangian Neural Networks (HiLNN). Grounded in the insight that temporal position sequences implicitly encode underlying dynamics, HiLNN employs a recurrent encoder to extract a latent context from history. This context not only reconstructs the unobserved initial velocity but also adaptively modulates the mass matrix, potential energy, and damping coefficients of a structured Lagrangian system. By leveraging a differentiable RK4 rollout scheme, the entire pipeline is optimized end-to-end under multi-step trajectory supervision and energy-consistency regularization. Empirical evaluations across conservative, dissipative, and heterogeneous variable-parameter systems show that HiLNN delivers superior long-term prediction accuracy and maintains precise energy profiles compared to state-of-the-art baselines. The source code is publicly available at https://github.com/yingtian22/History-informed-LNN.

Figures

Figures reproduced from arXiv: 2608.13215 by the authors.

Figure 1
Figure 1. Overview of HiLNN. Physics-guided neural dynamics address this limitation by incorporating me￾chanical priors. Hamiltonian and Lagrangian Neural Networks derive dynamics from learned energy functions or Lagrangians, improving interpretability and rollout stability [7,4,14,6,22]. Recent extensions further handle constrained, con￾trolled, or dissipative systems through explicit structure and energy-dissipation mechani… view at source ↗
Figure 2
Figure 2. Qualitative rollout comparison. Left: ground truth and HiLNN on three datasets (a–c). Right: ground truth and baseline predictions on the fixed pendulum (d–f). Implementation details. HiLNN uses a one-layer GRU to encode position history into a latent context, which is shared by the initial velocity head and the context-conditioned Lagrangian module. The mass and potential terms are modeled by MLPs, with mass positi… view at source ↗
Figure 3
Figure 3. Step-wise rollout error on the three pendulum systems. HiLNN achieves lower state MSE than LNN and LNN-multistep over the 32-step open-loop prediction horizon [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Energy behavior on the damped, standard, and variable-parameter pendulum systems [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Representative trajectory predictions on the three pendulum systems. HiLNN closely follows the ground-truth position q and velocity ˙q over the 32-step prediction horizon. benefit of combining history-informed context with structured Lagrangian dy￾namics. 4.5 Ablation …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

22 extracted references · 15 canonical work pages

  1. [1]

    In: Advances in Neural Information Processing Systems

    Battaglia, P.W., Pascanu, R., Lai, M., Rezende, D.J., Kavukcuoglu, K.: Interaction networks for learning about objects, relations and physics. In: Advances in Neural Information Processing Systems. vol. 29, pp. 4502–4510 (2016)

    work page 2016

  2. [2]

    Transactions on Machine Learning Research (2023)

    Buisson-Fenet, M., Morgenthaler, V., Trimpe, S., Di Meglio, F.: Recognition mod- els to learn dynamics from partial observations with neural odes. Transactions on Machine Learning Research (2023)

    work page 2023

  3. [3]

    In: Advances in Neural Information Processing Systems

    Chen, R.T.Q., Rubanova, Y., Bettencourt, J., Duvenaud, D.K.: Neural ordinary differential equations. In: Advances in Neural Information Processing Systems. vol. 31, pp. 6572–6583 (2018)

    work page 2018

  4. [4]

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

    Cranmer, M., Greydanus, S., Hoyer, S., Battaglia, P., Spergel, D., Ho, S.: La- grangian neural networks. In: ICLR 2020 Workshop on Integration of Deep Neural Models and Differential Equations (2020)

    work page 2020

  5. [5]

    Physical Review E104(3), 034312 (2021)

    Desai, S.A., Mattheakis, M., Sondak, D., Protopapas, P., Roberts, S.J.: Port-hamiltonian neural networks for learning explicit time- dependent dynamical systems. Physical Review E104(3), 034312 (2021). https://doi.org/10.1103/PhysRevE.104.034312

    doi:10.1103/physreve.104.034312 2021

  6. [6]

    In: Advances in Neural Information Processing Systems

    Finzi, M., Wang, K.A., Wilson, A.G.: Simplifying hamiltonian and lagrangian neu- ral networks via explicit constraints. In: Advances in Neural Information Processing Systems. vol. 33, pp. 13880–13889 (2020)

    work page 2020

  7. [7]

    In: Ad- vances in Neural Information Processing Systems

    Greydanus, S., Dzamba, M., Yosinski, J.: Hamiltonian neural networks. In: Ad- vances in Neural Information Processing Systems. vol. 32, pp. 15353–15363 (2019)

    work page 2019

  8. [8]

    arXiv preprint arXiv:1511.05121 (2015)

    Krishnan, R.G., Shalit, U., Sontag, D.: Deep kalman filters. arXiv preprint arXiv:1511.05121 (2015)

    arXiv 2015

Show all 22 references
  1. [9]

    CAAI Transactions on Intelligent Systems 20(6), 1355–1365 (2025)

    Li, M., Xiao, Y., Xing, X.: Trajectory prediction methods based on analytical mechanics and graph neural networks. CAAI Transactions on Intelligent Systems 20(6), 1355–1365 (2025). https://doi.org/10.11992/tis.202501020

    2025 doi

  2. [10]

    In: IEEE International Conference on Acoustics, Speech and Signal Processing (2026)

    Li, Y., Ding, K., Yang, C., Chen, S.Y., Tian, Y.: Distilling time series foundation models for efficient forecasting. In: IEEE International Conference on Acoustics, Speech and Signal Processing (2026)

    2026

  3. [11]

    arXiv preprint arXiv:2511.16715 (2025)

    Li, Y., Ding, K., Yang, C., Wang, H., Wang, H., Duan, H., Liu, J., Tian, Y.: Ddtime: Dataset distillation with spectral alignment and information bottleneck for time-series forecasting. arXiv preprint arXiv:2511.16715 (2025)

    2025

  4. [12]

    IEEE Transactions on Evolutionary Computa- tion (2026)

    Li, Y., Dong, J., Liu, J., Koniusz, P., Zeng, H., Yang, C., Liu, J., Tian, Y., Huang, T., Wu, H.: Evolving multimodal models for physical dynamics: A multi- objective neuroevolution approach. IEEE Transactions on Evolutionary Computa- tion (2026). https://doi.org/10.1109/TEVC....

    2026

  5. [13]

    In: Pro- ceedings of the IEEE/CVF International Conference on Computer Vision

    Li, Y., Yang, C., Zeng, H., Dong, Z., An, Z., Xu, Y., Tian, Y., Wu, H.: Frequency- aligned knowledge distillation for lightweight spatiotemporal forecasting. In: Pro- ceedings of the IEEE/CVF International Conference on Computer Vision. pp. 7262–7272 (2025)

    2025

  6. [14]

    In: International Conference on Learning Representations (2019)

    Lutter, M., Ritter, C., Peters, J.: Deep lagrangian networks: Using physics as model prior for deep learning. In: International Conference on Learning Representations (2019)

    2019

  7. [15]

    In: 2018 IEEE International Conference on Robotics and Automation

    Nagabandi, A., Kahn, G., Fearing, R.S., Levine, S.: Neural network dynam- ics for model-based deep reinforcement learning with model-free fine-tuning. In: 2018 IEEE International Conference on Robotics and Automation. pp. 7559–7566 (2018). https://doi.org/10.1109/ICRA.2018.846...

    2018

  8. [16]

    In: Advances in Neural Information Processing Systems

    Rubanova, Y., Chen, R.T.Q., Duvenaud, D.: Latent odes for irregularly-sampled time series. In: Advances in Neural Information Processing Systems. vol. 32 (2019)

    2019

  9. [17]

    In: Proceedings of the 37th International Conference on Machine Learning

    Sanchez-Gonzalez, A., Godwin, J., Pfaff, T., Ying, R., Leskovec, J., Battaglia, P.W.: Learning to simulate complex physics with graph networks. In: Proceedings of the 37th International Conference on Machine Learning. vol. 119, pp. 8459–8468 (2020)

    2020

  10. [18]

    arXiv preprint arXiv:2201.10085 (2022)

    Sosanya, A., Greydanus, S.: Dissipative hamiltonian neural networks: Learning dissipative and conservative dynamics separately. arXiv preprint arXiv:2201.10085 (2022)

    2022 arXiv

  11. [19]

    In: Dynamical Systems and Turbulence, Warwick 1980, Lecture Notes in Mathematics, vol

    Takens, F.: Detecting strange attractors in turbulence. In: Dynamical Systems and Turbulence, Warwick 1980, Lecture Notes in Mathematics, vol. 898, pp. 366–381. Springer, Berlin, Heidelberg (1981). https://doi.org/10.1007/BFb0091924

    1981 doi

  12. [20]

    In: Advances in Neural Information Processing Systems

    Yildiz, C., Heinonen, M., Lahdesmaki, H.: Ode2vae: Deep generative second order odes with bayesian neural networks. In: Advances in Neural Information Processing Systems. vol. 32 (2019)

    2019

  13. [21]

    arXiv preprint arXiv:2509.13783 (2025)

    Zhang, T., Zhai, W., Yann, R., Gao, J., Cao, H., Xing, X.: Floating-body hydro- dynamic neural networks. arXiv preprint arXiv:2509.13783 (2025)

    2025

  14. [22]

    In: International Conference on Learning Representations (2020)

    Zhong, Y.D., Dey, B., Chakraborty, A.: Symplectic ode-net: Learning hamiltonian dynamics with control. In: International Conference on Learning Representations (2020)

    2020

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.