Pith. sign in

REVIEW 3 major objections 4 minor 73 references

Hierarchical equivariant graph neural networks for forecasting collective motion in vortex clusters and microswimmers

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read This paper sets out to show that graph neural networks can forecast long-range collective motion when a hierarchy of local and global graphs is combined with rotation-translation equivariance.

desk verdict A fair, well-ablated empirical study of hierarchical equivariant GNNs for collective motion; the central claims hold up but need qualification about the lossy global graph and the interpolation framing of the phase transition. read the letter →

arxiv 2501.00626 v1 pith:SFWQJD26 submitted 2024-12-31 physics.flu-dyn

Alec J. Linot , Haotian Hang , Eva Kanso , Kunihiko Taira This is my paper · ORCID
classification physics.flu-dyn
keywords graphneuralnetworkscollectivemotionpointvorticesmicroswimmersequivariancehierarchicalgraphsHamiltonianconservationphasetransition
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 show that graph neural networks can forecast the collective motion of many interacting agents—vortex clusters and microswimmers—if two inductive biases are built in. First, instead of connecting every agent to every other, the network uses a local graph for short-range interactions and a global graph that compresses long-range interactions into aggregate nodes. Second, the graph is mapped into a rotation- and translation-invariant coordinate frame before the network acts, which guarantees equivariance by construction. On test cases the hierarchical equivariant network tracks cluster centroids, conserves the Hamiltonian of point vortices over long times, and reproduces the microswimmer phase transition from aggregation to swirling. A fully-connected baseline diverges, and removing either the hierarchy or equivariance creates spurious alignment in swimmers.

What carries the argument

The central object is a hierarchical graph pair, $G_l$ and $G_g$, combined with an equivariance-enforcing coordinate map. $G_l$ connects an agent to its short-range neighbors; $G_g$ links it to coarse-grained summaries of distant groups: for vortices, each outside cluster's centroid and summed circulation, and for swimmers, averaged position, orientation components, and speed in each of $S=12$ angular slices. Equivariance is enforced not by special layers but by mapping every input graph into a translation- and rotation-invariant subspace—centering about a reference point and rotating by a phase angle $\theta_e$ determined from the principal components of the local cluster (or the swimmer's own orientation), with sign ambiguity fixed by a modified PCA procedure—then rotating the network output back. The same graph operations run on both graphs, and the paper uses Chebyshev graph convolutions, so the hierarchy is the mechanism carrying the long-range information.

What would settle it

Take two vortex-cluster configurations with identical centroids and total circulations but different internal arrangements, such as one radially symmetric and one elongated, initialize the trained hierarchical equivariant GNN on both, and compare its forecasts to direct Biot-Savart simulations; if the true dynamics diverge while the model outputs identical motion, the coarse-grained global graph has lost information the dynamics use.

Watch

Extended reading notes

Core claim

The central discovery is that long-range order in collective dynamics is learnable when the graph encodes physical structure rather than raw connectivity. For point vortices, the local graph contains vortices within a cluster and the global graph represents each outside cluster by its centroid and total circulation; for microswimmers, the local graph covers a radius $R_1$ and the global graph averages the swimmers in $S=12$ angular slices. The paper demonstrates that a GNN built on this hierarchy, combined with an invariant-subspace mapping—centering and rotating by a PCA-derived angle—predicts the time derivatives of agent states well. Quantitatively, the hierarchical equivariant GNN keeps average Hamiltonian error near $1\%$ after clusters travel over six diameters and maintains low polar order for microswimmers over 500 time units, while fully-connected or non-equivariant models diverge or spuriously align.

Load-bearing premise

The load-bearing premise is that a faraway cluster or slice is completely described by its averaged properties—centroid and total circulation for vortices, or mean position, orientation, and speed for swimmers—so that the force on an agent never depends on details inside that distant group beyond those aggregates.

Editorial extensions

If this is right

  • For point vortices, the hierarchical equivariant GNN conserves the Hamiltonian over long forecast horizons: ensemble-averaged error is about $1\%$ after clusters move more than six diameters, where the fully-connected GNN diverges.
  • For microswimmers, the method reproduces the phase transition from aggregation to swirling as the rotational mobility $\nu$ crosses zero, and can even forecast the breakup of an initially aggregated state—a transition absent from the training data.
  • Removing either the hierarchy or the rotation equivariance degrades long-time behavior: microswimmers spuriously align and the polar order parameter rises, while the hierarchical equivariant GNN keeps it low.
  • Because the hierarchy separates short- and long-range information, the hierarchical equivariant GNN achieves better or equal accuracy than a larger-radius local model while taking less than half the compute time.

Reading between the lines

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

  • Editorial inference: the global graph is effectively a first-order moment approximation of each distant group, which suggests a testable upgrade—adding second-order moments such as internal variance or orientation spread inside each cluster or slice may extend accuracy to systems where the internal shape of distant groups matters.
  • Editorial inference: the same hierarchical coarse-graining recipe should transfer to other long-range interacting populations, such as gravitating bodies, charged colloids, or pedestrian crowds, where a fully-connected graph is prohibitive; that transfer is an implicit but untested claim of the paper.
  • Editorial inference: equivariance likely buys data efficiency more than representation power, so one could measure how much training data the hierarchical equivariant GNN needs to reach a fixed error relative to the non-equivariant hierarchical GNN, since the invariant mapping removes the need to learn all orientations.
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 / 4 minor

Summary. The paper proposes hierarchical equivariant graph neural networks (HE-GNNs) for forecasting the collective dynamics of point-vortex clusters and microswimmers. The method constructs a local graph for short-range interactions and a global graph for aggregated long-range interactions, then maps the input to a rotation-translation invariant subspace before applying graph convolutions and rotating the output back. For point vortices, the authors compare HE-GNN against fully-connected and hierarchical non-equivariant baselines, report improved centroid tracking, and report about 1% ensemble-averaged Hamiltonian error over long rollouts. For microswimmers, they train on 21 trajectories with rotational mobility coefficients ν evenly sampled in [−1,1], and show that HE-GNN reproduces qualitative aggregation/swirling behavior and polar-order statistics on new initial conditions, whereas local or non-equivariant hierarchical models develop spurious alignment.

Significance. The methodological idea of combining a two-scale graph hierarchy with an architecture-agnostic invariant-subspace map is attractive and potentially widely applicable. The empirical design is a genuine strength: all comparisons are on test data, the ablations isolate the contributions of hierarchy and equivariance, and the microswimmer results include three independently trained models per configuration. If the global-graph aggregation is accepted as a sufficient statistic, the point-vortex and microswimmer results provide a compelling demonstration that explicit inductive biases improve data-driven forecasting of collective motion. However, the paper does not establish that the coarse global features (centroid plus total circulation for vortex clusters; averaged slices for microswimmers) are sufficient for the long-range couplings at the separations used, and the phase-transition claim is stronger than the ν-interpolation experiment supports. These two issues are load-bearing for the central claims, so the significance is conditional pending additional analysis or rephrasing.

major comments (3)
  1. [§I C, Eq. (13), global-graph definition] The global graph represents every outside vortex cluster only by its unweighted average position and total circulation. The Biot-Savart velocity (Eq. 13) induced by that cluster at a target vortex is not determined by these quantities: the dipole and higher multipole moments contribute at the minimum cluster separation of 2.5 D0 used to generate the data, and two configurations with identical local graphs and identical aggregate-node features can have different true velocities. Since the GNN input is identical for such a pair, no trained model can be accurate for both configurations. The paper does not quantify this aggregation error or test the sufficiency of the representation, for example by adding circulation-weighted centroids or dipole moments, or by comparing against a global graph that retains fuller cluster information. The reported ~1% Hamiltonian error is an ensemble average over random disk-like clusters and therefore does not establish conservation for configurations with different internal cluster shapes. This issue is central to the claim that the hierarchy captures long-range interactions.
  2. [§I D (microswimmers) and §III D] The abstract and Sec. I D state that the model 'predicts the transition from aggregation to swirling,' but ν is an input feature and the 21 training trajectories sample ν evenly over [−1,1]; all test values shown in Figs. 7–9 (−0.5, 0, 0.2, 0.5) are among the training values. The model is therefore interpolating in ν rather than predicting the transition at unseen control parameters, and no quantitative critical-ν or bifurcation prediction is reported. The extrapolative experiment in Fig. 7(c), which starts from an aggregated configuration at ν=0.5, is a strong state-space generalization result, but it does not by itself support the phase-transition prediction claim. Please either soften the claim or add tests at unseen ν values and report a predicted transition curve against the true one.
  3. [§I C, Fig. 5(c)] The wording 'conserves the Hamiltonian over long times' is stronger than the evidence shown. Figure 5(c) reports a normalized ensemble-averaged Hamiltonian error of about 1% after trajectories have traveled roughly six diameters; this is approximate near-conservation, not exact conservation. The model is not symplectic, and the error is averaged over 50 initial conditions. To support the conservation claim, the paper should report the distribution or maximum error over individual trajectories, the time evolution of the error, and the behavior over the full T=200 test horizon. This matters because the Hamiltonian depends on all pair distances, including the cross-cluster interactions represented by the coarse global graph.
minor comments (4)
  1. [§I D, after Fig. 8] The sentence 'Here, we have shown the HE-GNN model accurately captures the phase transition,' ends with a comma and is incomplete; it should end with a period and be separated from the following paragraph.
  2. [§III D] The potential is referred to as 'Leonard Jones'; the standard spelling is 'Lennard-Jones.'
  3. [Eq. (9)] The definition of the prediction time T_p as 'argmax_{t_f} {t_f | ...}' is unusual; it should be stated as the largest t_f such that the condition holds for all times up to t_f, and the choice of the threshold ϵ should be justified.
  4. [§I C] For the vortex global graph, the text says the node stores 'average position' but does not specify whether this is the unweighted arithmetic mean or the circulation-weighted centroid; the distinction is important because the leading Biot-Savart monopole is centered at the circulation-weighted position.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the central claims are measured on held-out data and the control-parameter and self-citation concerns do not reduce the results to their inputs.

full rationale

I find no circular step in the derivation chain. The paper's central result is an architecture comparison: HE-GNN, H-GNN, L-GNN, and F-GNN are all trained on the same data and evaluated on held-out initial conditions; the reported improvements in centroid tracking, Hamiltonian error, polar order, and earth mover's distance are measured outputs, not quantities built into the loss or the graph construction. The vortex Hamiltonian (Eq. 7) is not used as a training target or input feature, and the enforced rotation/translation equivariance does not mathematically imply conservation of H, so the ~1% Hamiltonian error is an empirical result. The microswimmer 'phase transition' is conditioned on nu, which is an explicit input feature sampled across [-1,1] during training, so calling this 'prediction' is interpolation over a control parameter rather than discovery; however, this is not circular because the model still must learn the trajectory dynamics and the ablation in Fig. 9 shows that hierarchy and equivariance are needed to avoid spurious alignment. Self-citations (e.g., Nair and Taira [55,56]) are used only for dataset conventions and baseline context, not to justify the main claims. The global-graph aggregation to centroid/circulation or slice averages is a coarsening approximation whose information loss is a correctness/robustness concern, not a circular definition.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

No new physical entities are introduced. The modeling load is carried by the data-generating ODEs and by hand-selected graph construction choices, which are listed as free parameters.

free parameters (6)
  • Microswimmer local graph radius R1 = 20 (swept from 5 to 30+)
    Chosen by comparing training MSE, prediction time, and polar-order EMD; it directly controls which interactions are treated as local.
  • Microswimmer global graph outer radius R2 = 30 = R1 + 10
    Set as a fixed offset from R1 rather than optimized independently; long-range aggregation outside R2 is omitted.
  • Number of angular slices S in the microswimmer global graph = 12
    Hand-selected to summarize long-range interactions; no ablation over S is reported.
  • Edge-weight exponent in adjacency matrices = 2 (W = 1/||x - x'||^2)
    The authors report that an inverse-distance weight gave about two times greater error, so the squared inverse distance was retained.
  • Minimum initial separation between vortex clusters = 2.5 D0
    Set empirically to prevent cluster mixing over the forecast horizon, based on a known coherence distance of 2 D0.
  • GNN architecture and training hyperparameters = 5 Chebyshev layers with 3 polynomials, hidden width 64, Adam with piecewise learning rate 1e-3 to 1e-5
    Selected through parameter sweeps; these are standard but are part of the fitted pipeline and influence the reported accuracy.
assumptions (5)
  • domain assumption Point-vortex dynamics follow the Biot-Savart law (Eq. 13), with same-sign circulations and separated clusters that remain coherent for the forecast horizon.
    The GNN is trained and evaluated on simulations of this ODE; accuracy claims apply to this model system, not to general vortex flows.
  • domain assumption Microswimmer dynamics follow the potential-dipole ODE with Lennard-Jones repulsion (Eqs. 14-16) with fixed μ=0.9, U=1, σ=1.
    The aggregation-to-swirling transition is a property of this model from prior literature; the learned surrogate inherits that physics.
  • ad hoc to paper Compressing each outside vortex cluster to centroid plus total circulation, and the far annulus into S=12 averaged slices, is sufficient to represent long-range forces on a target agent.
    This is the central modeling premise of the hierarchy (Sec. I C and I D) and is not tested against finer-grained aggregation.
  • standard math Chebyshev graph convolutions approximate spectral graph convolutions, and the PCA-based sign disambiguation (Bro et al.) gives a unique invariant frame.
    These are standard tools from the cited literature and are not derived in the paper.
  • domain assumption Rotating inputs by the PCA phase angle and rotating outputs back enforces Eq. 4, assuming the phase angle is unique and continuous for the data.
    For compact clusters or a single central swimmer, the phase is well defined; degenerate eigenvalue cases could make the frame discontinuous.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hierarchical equivariant graph neural networks for forecasting collective motion in vortex clusters and microswimmers." pith.science (2026). https://pith.science/paper/SFWQJD26

@misc{pith2026250100626,
  author       = {Pith},
  title        = {Pith review of: Hierarchical equivariant graph neural networks for forecasting collective motion in vortex clusters and microswimmers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SFWQJD26}},
  note         = {Machine review of arXiv:2501.00626}
}
read the original abstract

Data-driven modeling of collective dynamics is a challenging problem because emergent phenomena in multi-agent systems are often shaped by long-range interactions among individuals. For example, in bird flocks and fish schools, long-range vision and flow coupling drive individual behaviors across the collective. Such collective motion can be modeled using graph neural networks (GNNs), but GNNs struggle when graphs become large and often fail to capture long-range interactions. Here, we construct hierarchical and equivariant GNNs, and show that these GNNs accurately predict local and global behavior in systems with collective motion. As representative examples, we apply this approach to simulations of clusters of point vortices and populations of microswimmers. For the point vortices, we define a local graph of vortices within a cluster and a global graph of interactions between clusters. For the microswimmers, we define a local graph around each microswimmer and a global graph that groups long-range interactions. We then combine this hierarchy of graphs with an approach that enforces equivariance to rotations and translations. This combination results in a significant improvement over a fully-connected GNN. For point vortices, our method conserves the Hamiltonian over long times, and, for microswimmers, our method predicts the transition from aggregation to swirling.

Figures

Figures reproduced from arXiv: 2501.00626 by the authors.

Figure 1
Figure 1. FIG. 1: Three styles of graphs. The fully-connected graph connects every node to every [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2: The present hierarchical and equivariant GNN method illustrated using the point [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3: Vortex dynamics given by (a) the reference solution, (b) the hierarchical and [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: FIG. 4: Trajectories of centroids for cases with (a) two, (b) three, (c) four, and (d) five [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5: Ensemble averaged tracking error of (a) centroids and (b) inidividual vortices. (c) [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6: (a) construction of the local (solid line) and global graphs (dotted line) for [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: FIG. 7: Snapshots of microswimmers at various rotational mobility coefficients. Test data is [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]
Figure 8
Figure 8. Figure 8: FIG. 8: PDFs of (a) velocity and (b) rotational activity parameter at various rotational [PITH_FULL_IMAGE:figures/full_fig_p017_8.png]
Figure 9
Figure 9. Figure 9: FIG. 9: Snapshots of microswimmers predicted using the (a) HE-GNN, (b) H-GNN, and [PITH_FULL_IMAGE:figures/full_fig_p018_9.png]
Figure 10
Figure 10. Figure 10: FIG. 10: Performance of models for varied [PITH_FULL_IMAGE:figures/full_fig_p020_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

73 extracted references · 67 canonical work pages

  1. [1]

    Vicsek and A

    T. Vicsek and A. Zafeiris, Collective motion, Physics Reports 517, 71 (2012)

    2012

  2. [2]

    Kudrolli, Concentration dependent diffusion of self-propelled rods, Phys

    A. Kudrolli, Concentration dependent diffusion of self-propelled rods, Phys. Rev. Lett. 104, 088001 (2010)

    work page 2010

  3. [3]

    N9UE8U0uRgPmA6iatrPUJD0YQe8=

    Predict dynamics with GNN operations <latexit sha1_base64="N9UE8U0uRgPmA6iatrPUJD0YQe8=">AAACF3icbVC7SgNBFJ31GeMramkzGAWrZVcFLQMWWoUI5gFJCLOzN8mQ2Qczd9VlyV/Y+Cs2ForYauffOHkUmnhg4HDOudy5x4ul0Og439bC4tLyympuLb++sbm1XdjZrekoURyqPJKRanhMgxQhVFGghEasgAWehLo3uBz59TtQWkThLaYxtAPWC0VXcIZG6hTsFsIDZqc2rSjwBUfqpyELBNf0XmCfXpXLNIpBjeN62CkUHdsZg84Td0qKZIpKp/DV8iOeBBAi...

    work page

  4. [4]

    1U3O6h5OY1EsngNYx1mnEaQztrg=

    Forecast in the original space <latexit sha1_base64="1U3O6h5OY1EsngNYx1mnEaQztrg=">AAACEXicbVDLSgMxFM34rPVVdekmWIWuyowUdFkQxGUF+4C2lEx624ZmkiG5I5ahv+DGX3HjQhG37tz5N6aPhbYeuHA4597k3hPGUlj0/W9vZXVtfWMzs5Xd3tnd288dHNasTgyHKtdSm0bILEihoIoCJTRiAywKJdTD4dXEr9+DsUKrOxzF0I5YX4me4Ayd1MkVWggPmJaK9Fob4MwiFYriAKg2oi8Uk9TGjMO4k8v7RX8KukyCOcmTOSqd3Ferq3kSgUIumbXNwI+xnT...

    work page

  5. [5]

    Ibele, T

    M. Ibele, T. E. Mallouk, and A. Sen, Schooling behavior of light-powered autonomous micro- motors in water, Angewandte Chemie International Edition 48, 3308 (2009)

    work page 2009

  6. [6]

    Wu and A

    X.-L. Wu and A. Libchaber, Particle diffusion in a quasi-two-dimensional bacterial bath, Phys. Rev. Lett. 84, 3017 (2000)

    work page 2000

  7. [7]

    Ballerini, N

    M. Ballerini, N. Cabibbo, R. Candelier, A. Cavagna, E. Cisbani, I. Giardina, A. Orlandi, G. Parisi, A. Procaccini, M. Viale, and V. Zdravkovic, Empirical investigation of starling flocks: a benchmark study in collective animal behaviour, Animal Behaviour 76, 201 (2008)

    work page 2008

  8. [8]

    Moussa ¨ ıd, D

    M. Moussa ¨ ıd, D. Helbing, and G. Theraulaz, How simple rules determine pedestrian behavior and crowd disasters, Proceedings of the National Academy of Sciences 108, 6884 (2011)

    work page 2011

Show all 73 references
  1. [9]

    Hochreiter and J

    S. Hochreiter and J. Schmidhuber, Long short-term memory, Neural Comput. 9, 1735–1780 (1997)

    1997

  2. [10]

    R. T. Chen, Y. Rubanova, J. Bettencourt, and D. K. Duvenaud, Neural ordinary differential equations, Advances in neural information processing systems 31 (2018)

    2018

  3. [11]

    Lukoˇ seviˇ cius and H

    M. Lukoˇ seviˇ cius and H. Jaeger, Reservoir computing approaches to recurrent neural network training, Computer Science Review 3, 127 (2009)

    2009

  4. [12]

    Vlachas, J

    P. Vlachas, J. Pathak, B. Hunt, T. Sapsis, M. Girvan, E. Ott, and P. Koumoutsakos, Backprop- agation algorithms and reservoir computing in recurrent neural networks for the forecasting of complex spatiotemporal dynamics, Neural Networks 126, 191 (2020)

    2020

  5. [13]

    Pathak, B

    J. Pathak, B. Hunt, M. Girvan, Z. Lu, and E. Ott, Model-Free Prediction of Large Spatiotem- porally Chaotic Systems from Data: A Reservoir Computing Approach, Physical Review Let- ters 120, 24102 (2018)

    2018

  6. [14]

    A. J. Linot and M. D. Graham, Deep learning to discover and predict dynamics on an inertial manifold, Phys. Rev. E 101, 062209 (2020). 29

    2020

  7. [15]

    A. J. Linot and M. D. Graham, Data-driven reduced-order modeling of spatiotemporal chaos with neural ordinary differential equations, Chaos: An Interdisciplinary Journal of Nonlinear Science 32, 073110 (2022)

    2022

  8. [16]

    Borrelli, L

    G. Borrelli, L. Guastoni, H. Eivazi, P. Schlatter, and R. Vinuesa, Predicting the temporal dynamics of turbulent channels through deep learning, International Journal of Heat and Fluid Flow 96, 109010 (2022)

    2022

  9. [17]

    A. J. Linot and M. D. Graham, Dynamics of a data-driven low-dimensional model of turbulent minimal Couette flow, Journal of Fluid Mechanics 973, A42 (2023)

    2023

  10. [18]

    Maulik, R

    R. Maulik, R. Egele, B. Lusch, and P. Balaprakash, Recurrent neural network architecture search for geophysical emulation, in Proceedings of the International Conference for High Per- formance Computing, Networking, Storage and Analysis, SC ’20 (IEEE Press, 2020)

    2020

  11. [19]

    Arcomano, I

    T. Arcomano, I. Szunyogh, A. Wikner, J. Pathak, B. R. Hunt, and E. Ott, A hybrid approach to atmospheric modeling that combines machine learning with a physics-based numerical model, Journal of Advances in Modeling Earth Systems 14, e2021MS002712 (2022)

    2022

  12. [20]

    Sanchez-Gonzalez, J

    A. Sanchez-Gonzalez, J. Godwin, T. Pfaff, R. Ying, J. Leskovec, and P. W. Battaglia, Learning to simulate complex physics with graph networks, in Proceedings of the 37th International Conference on Machine Learning, ICML’20 (JMLR.org, 2020)

    2020

  13. [21]

    Bhattoo, S

    R. Bhattoo, S. Ranu, and N. M. A. Krishnan, Learning the dynamics of particle-based systems with Lagrangian graph neural networks, Machine Learning: Science and Technology 4, 015003 (2023)

    2023

  14. [22]

    Li and A

    Z. Li and A. B. Farimani, Graph neural network-accelerated Lagrangian fluid simulation, Computers & Graphics 103, 201 (2022)

    2022

  15. [23]

    Choi and K

    Y. Choi and K. Kumar, Graph neural network-based surrogate model for granular flows, Computers and Geotechnics 166, 106015 (2024)

    2024

  16. [24]

    Xiong, X

    S. Xiong, X. He, Y. Tong, Y. Deng, and B. Zhu, Neural vortex method: From finite Lagrangian particles to infinite dimensional Eulerian dynamics, Computers & Fluids 258, 105811 (2023)

    2023

  17. [25]

    Battaglia, R

    P. Battaglia, R. Pascanu, M. Lai, D. J. Rezende, and K. Kavukcuoglu, Interaction networks for learning about objects, relations and physics, in Proceedings of the 30th International Conference on Neural Information Processing Systems, NIPS’16 (Curran Associates Inc., Red Hook,...

    2016

  18. [26]

    Pfaff, M

    T. Pfaff, M. Fortunato, A. Sanchez-Gonzalez, and P. W. Battaglia, Learning mesh-based sim- ulation with graph networks, in 9th International Conference on Learning Representations, ICLR 2021, Virtual Event, Austria, May 3-7, 2021(OpenReview.net, 2021)

    2021

  19. [27]

    Peng, Y.-Z

    J.-Z. Peng, Y.-Z. Wang, S. Chen, Z.-H. Chen, W.-T. Wu, and N. Aubry, Grid adaptive reduced- order model of fluid flow based on graph convolutional neural network, Physics of Fluids 34, 087121 (2022)

    2022

  20. [28]

    A. P. Willis, P. Cvitanovi´ c, and M. Avila, Revealing the state space of turbulent pipe flow by symmetry reduction, Journal of Fluid Mechanics 721, 514–540 (2013)

    2013

  21. [29]

    Thomas, T

    N. Thomas, T. E. Smidt, S. Kearnes, L. Yang, L. Li, K. Kohlhoff, and P. Riley, Tensor field networks: Rotation- and translation-equivariant neural networks for 3D point clouds., CoRR abs/1802.08219 (2018)

    2018 arXiv

  22. [30]

    F. B. Fuchs, D. E. Worrall, V. Fischer, and M. Welling, SE(3)-transformers: 3D roto- translation equivariant attention networks, inProceedings of the 34th International Conference on Neural Information Processing Systems, NIPS ’20 (Curran Associates Inc., Red Hook, NY, USA, 2020)

    2020

  23. [31]

    V. G. Satorras, E. Hoogeboom, and M. Welling, E(n) equivariant graph neural networks, in Proceedings of the 38th International Conference on Machine Learning, Proceedings of Machine Learning Research, Vol. 139, edited by M. Meila and T. Zhang (PMLR, 2021) pp. 9323–9332

    2021

  24. [32]

    Brandstetter, R

    J. Brandstetter, R. Hesselink, E. van der Pol, E. J. Bekkers, and M. Welling, Geometric and physical quantities improve E(3) equivariant message passing, in International Conference on Learning Representations(2022)

    2022

  25. [33]

    A. P. Toshev, G. Galletti, J. Brandstetter, S. Adami, and N. A. Adams, E(3) equivariant graph neural networks for particle-based fluid mechanics, ArXiv abs/2304.00150 (2023)

    2023 arXiv

  26. [34]

    M. Xu, J. Han, A. Lou, K. Azizzadenesheli, S. Ermon, and A. Anandkumar, Equivariant graph neural operator for modeling 3D dynamics (2024)

    2024

  27. [35]

    Y. Liu, J. Cheng, H. Zhao, T. Xu, P. Zhao, F. Tsung, J. Li, and Y. Rong, SEGNO: Generalizing equivariant graph neural networks with physical inductive biases, inThe Twelfth International Conference on Learning Representations(2024). 31

    2024

  28. [36]

    Liao and T

    Y.-L. Liao and T. Smidt, Equiformer: Equivariant graph attention transformer for 3D atom- istic graphs, in The Eleventh International Conference on Learning Representations(2023)

    2023

  29. [37]

    Y.-L. Liao, B. M. Wood, A. Das, and T. Smidt, Equiformerv2: Improved equivariant trans- former for scaling to higher-degree representations, in The Twelfth International Conference on Learning Representations(2024)

    2024

  30. [38]

    M. Lino, S. Fotiadis, A. A. Bharath, and C. D. Cantwell, Multi-scale rotation-equivariant graph neural networks for unsteady Eulerian fluid dynamics, Physics of Fluids 34, 087110 (2022)

    2022

  31. [39]

    Shankar, S

    V. Shankar, S. Barwey, Z. Kolter, R. Maulik, and V. Viswanathan, Importance of equivariant and invariant symmetries for fluid flow modeling (2023), arXiv:2307.05486 [physics.flu-dyn]

    2023 arXiv

  32. [40]

    Z. Xiao, H. Lin, R. Li, H. Chao, and S. Ding, Endowing deep 3D models with rotation invari- ance based on principal component analysis, 2020 IEEE International Conference on Multi- media and Expo (ICME) , 1 (2019)

    2019

  33. [41]

    F. Li, K. Fujiwara, F. Okura, and Y. Matsushita, A closer look at rotation-invariant deep point cloud analysis, in 2021 IEEE/CVF International Conference on Computer Vision (ICCV) (2021) pp. 16198–16207

    2021

  34. [42]

    R. Wang, F. Fang, J. Cui, and W. Zheng, Learning self-driven collective dynamics with graph networks, Scientific Reports 12, 500 (2022)

    2022

  35. [43]

    Vicsek, A

    T. Vicsek, A. Czir´ ok, E. Ben-Jacob, I. Cohen, and O. Shochet, Novel type of phase transition in a system of self-driven particles, Phys. Rev. Lett. 75, 1226 (1995)

    1995

  36. [44]

    F. J. H. Heras, F. Romero-Ferrero, R. C. Hinz, and G. G. de Polavieja, Deep attention networks reveal the rules of collective motion in zebrafish, PLOS Computational Biology 15, 1 (2019)

    2019

  37. [45]

    Ha and H

    S. Ha and H. Jeong, Unraveling hidden interactions in complex systems with deep learning, Scientific Reports 11, 12804 (2021)

    2021

  38. [46]

    Veliˇ ckovi´ c, G

    P. Veliˇ ckovi´ c, G. Cucurull, A. Casanova, A. Romero, P. Li` o, and Y. Bengio, Graph Attention Networks, International Conference on Learning Representations (2018)

    2018

  39. [47]

    Diestel, Graph Theory, fifth edition ed., Graduate Texts in Mathematics, Vol

    R. Diestel, Graph Theory, fifth edition ed., Graduate Texts in Mathematics, Vol. 173 (Springer Berlin Heidelberg, Berlin, Heidelberg, 2017) pp. xviii–xviii

    2017

  40. [48]

    Defferrard, X

    M. Defferrard, X. Bresson, and P. Vandergheynst, Convolutional neural networks on graphs with fast localized spectral filtering, in Proceedings of the 30th International Conference on 32 Neural Information Processing Systems, NIPS’16 (Curran Associates Inc., Red Hook, NY, USA,...

    2016

  41. [49]

    Froehlich and P

    S. Froehlich and P. Cvitanovi´ c, Reduction of continuous symmetries of chaotic flows by the method of slices, Communications in Nonlinear Science and Numerical Simulation 17, 2074 (2012), special Issue: Mathematical Structure of Fluids and Plasmas

    2012

  42. [50]

    N. B. Budanur, P. Cvitanovi´ c, R. L. Davidchack, and E. Siminos, Reduction of SO(2) symmetry for spatially extended dynamical systems, Phys. Rev. Lett. 114, 084102 (2015)

    2015

  43. [51]

    Cortelezzi and A

    L. Cortelezzi and A. Leonard, Point vortex model of the unsteady separated flow past a semi- infinite plate with transverse motion, Fluid Dynamics Research 11, 263 (1993)

    1993

  44. [52]

    Chatelain, M

    P. Chatelain, M. Duponcheel, D.-G. Caprace, Y. Marichal, and G. Winckelmans, Vortex particle-mesh simulations of vertical axis wind turbine flows: from the airfoil performance to the very far wake, Wind Energy Science 2, 317 (2017)

    2017

  45. [53]

    Samarbakhsh and N

    S. Samarbakhsh and N. Kornev, Simulation of a free circular jet using the vortex particle intensified les (vπles), International Journal of Heat and Fluid Flow 80, 108489 (2019)

    2019

  46. [54]

    J. D. Eldredge, Examples of two-dimensional flow modeling, in Mathematical Modeling of Unsteady Inviscid Flows(Springer International Publishing, Cham, 2019) pp. 341–367

    2019

  47. [55]

    Mimeau and I

    C. Mimeau and I. Mortazavi, A review of vortex methods and their applications: From creation to recent advances, Fluids 6, 10.3390/fluids6020068 (2021)

    2021 doi

  48. [56]

    P. G. Saffman, Singular distributions of vorticity, in Vortex Dynamics, Cambridge Monographs on Mechanics (Cambridge University Press, 1993) p. 20–45

    1993

  49. [57]

    A. G. Nair and K. Taira, Network-theoretic approach to sparsified discrete vortex dynamics, Journal of Fluid Mechanics 768, 549–571 (2015)

    2015

  50. [58]

    Gopalakrishnan Meena, A

    M. Gopalakrishnan Meena, A. G. Nair, and K. Taira, Network community-based model re- duction for vortical flows, Phys. Rev. E 97, 063103 (2018)

    2018

  51. [59]

    J. F. Brady, G. Bossis, et al., Stokesian dynamics, Annual Review of Fluid Mechanics 20, 111 (1988)

    1988

  52. [60]

    G. J. Elfring and J. F. Brady, Active Stokesian dynamics, Journal of Fluid Mechanics 952, A19 (2022)

    2022

  53. [61]

    A. U. Oza, L. Ristroph, and M. J. Shelley, Lattices of hydrodynamically interacting flapping swimmers, Physical Review X 9, 041024 (2019). 33

    2019

  54. [62]

    Heydari, H

    S. Heydari, H. Hang, and E. Kanso, Mapping spatial patterns to energetic benefits in groups of flow-coupled swimmers, Elife , 2024 (2024)

    2024

  55. [63]

    Cheng Hou Tsang and E

    A. Cheng Hou Tsang and E. Kanso, Flagella-induced transitions in the collective behavior of confined microswimmers, Phys. Rev. E 90, 021001 (2014)

    2014

  56. [64]

    A. C. H. Tsang and E. Kanso, Density shock waves in confined microswimmers, Physical Review Letters 116, 048101 (2016)

    2016

  57. [65]

    Takagi, J

    D. Takagi, J. Palacci, A. B. Braunschweig, M. J. Shelley, and J. Zhang, Hydrodynamic capture of microswimmers into sphere-bound orbits, Soft Matter 10, 1784 (2014)

    2014

  58. [66]

    I. D. Couzin, J. Krause, et al., Self-organization and collective behavior in vertebrates, Ad- vances in the Study of Behavior 32, 10 (2003)

    2003

  59. [67]

    Filella, F

    A. Filella, F. Nadal, C. Sire, E. Kanso, and C. Eloy, Model of collective fish behavior with hydrodynamic interactions, Physical Review Letters 120, 198101 (2018)

    2018

  60. [68]

    Rubner, C

    Y. Rubner, C. Tomasi, and L. Guibas, A metric for distributions with applications to image databases, in Sixth International Conference on Computer Vision (IEEE Cat. No.98CH36271) (1998) pp. 59–66

    1998

  61. [69]

    T. N. Kipf and M. Welling, Semi-supervised classification with graph convolutional networks, in International Conference on Learning Representations (ICLR)(2017)

    2017

  62. [70]

    R. Bro, E. Acar, and T. G. Kolda, Resolving the sign ambiguity in the singular value decom- position, Journal of Chemometrics 22, 135 (2008)

    2008

  63. [71]

    Tao, Explicit symplectic approximation of nonseparable hamiltonians: Algorithm and long time performance, Phys

    M. Tao, Explicit symplectic approximation of nonseparable hamiltonians: Algorithm and long time performance, Phys. Rev. E 94, 043303 (2016)

    2016

  64. [72]

    Fey and J

    M. Fey and J. E. Lenssen, Fast graph representation learning with PyTorch Geometric, in ICLR Workshop on Representation Learning on Graphs and Manifolds(2019)

    2019

  65. [73]

    Kanso and A

    E. Kanso and A. C. H. Tsang, Pursuit and synchronization in hydrodynamic dipoles, Journal of Nonlinear Science 25, 1141 (2015). 34

    2015

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.