pith. sign in

arxiv: 2605.27756 · v1 · pith:3BGOA3ZKnew · submitted 2026-05-26 · ⚛️ physics.flu-dyn · cs.LG· cs.NA· math.DS· math.NA

Sparse POD Mode Selection and Manifold Dimensionality Reduction with Neural Networks

Tomoki Koike , Prakash Mohan , Marc T. Henry de Frahan , Elizabeth Qian , Julie Bessac This is my paper

Pith reviewed 2026-06-29 14:58 UTC · model grok-4.3

classification ⚛️ physics.flu-dyn cs.LGcs.NAmath.DSmath.NA
keywords model order reductionproper orthogonal decompositionneural networkssparse mode selectionturbulent channel flowmanifold learningLassoNetdimensionality reduction
0
0 comments X

The pith

SparseModesNet uses LassoNet sparsity in a neural decoder to select informative POD modes and cut reconstruction error 51-78% in high-Reynolds turbulent channel flow.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces SparseModesNet, a model-order-reduction method that encodes flow data with standard POD modes but decodes them through a neural network whose sparsity is controlled by LassoNet. LassoNet enforces hierarchical sparsity on residual connections so the network automatically keeps the POD modes that matter most for accurate reconstruction, including low-energy modes that energy ranking would discard. On advection-dominated and chaotic test problems the method matches or beats existing polynomial-manifold approaches; on the Re_τ=5200 turbulent channel flow it lowers reconstruction error by 51-78% while the selected modes remain physically interpretable. Readers care because the same high-dimensional simulation data can now support faster inverse problems, control design, and real-time surrogates without losing dynamical fidelity.

Core claim

SparseModesNet employs linear encoding via POD modes and nonlinear NN decoding. The decoder leverages LassoNet, a method enforcing hierarchical sparsity through residual connections with linear skip layers, to simultaneously select informative POD modes and learn a nonlinear mapping that minimizes reconstruction error. On benchmark advection-dominated and chaotic flows, SparseModesNet matches or exceeds state-of-the-art performance. For turbulent channel flow at friction Reynolds number Re_τ=5200, we reduce reconstruction error by 51--78% compared to existing polynomial manifold methods while maintaining interpretability through physically meaningful mode selection.

What carries the argument

SparseModesNet, which encodes with POD modes and decodes with a LassoNet-enforced neural network that jointly selects sparse modes and learns the nonlinear reconstruction map.

If this is right

  • Fewer modes suffice for accurate reconstruction of advection-dominated and turbulent flows.
  • Downstream tasks such as inverse problems and control become cheaper because the reduced-order model is both accurate and interpretable.
  • Energy-based POD selection can be replaced by data-driven selection that keeps dynamically relevant but low-energy modes.
  • The same framework applies across multiple flow regimes without changing the nonlinear mapping form in advance.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The selected modes may supply new physical insight into which scales dominate transport in wall-bounded turbulence.
  • The method could be dropped into existing POD pipelines for any high-dimensional dynamical system that already stores snapshot data.
  • Because the decoder is a neural network, the same trained model might be queried for out-of-sample predictions or sensitivity analysis without retraining.
  • If the sparsity pattern proves stable across Reynolds numbers, it could guide adaptive mesh refinement or sensor placement in experiments.

Load-bearing premise

LassoNet's hierarchical sparsity on residual connections will reliably identify and retain the low-energy but dynamically important POD modes without requiring post-hoc tuning or validation on held-out data.

What would settle it

Running SparseModesNet on the same Re_τ=5200 channel-flow snapshots and obtaining reconstruction error no lower than the polynomial-manifold baselines, or obtaining mode selections that lack clear physical meaning in the flow.

read the original abstract

High-performance computing enables simulation of high-dimensional physical systems, but downstream analyses such as inverse problems and control remain computationally expensive, motivating model order reduction (MOR) to construct efficient low-dimensional surrogates. Proper Orthogonal Decomposition (POD), a widely adopted data-driven MOR method, projects dynamics onto linear subspaces spanned by the most energetic modes. However, POD struggles for problems with slowly decaying Kolmogorov \(n\)-widths, such as advection-dominated and turbulent flows, requiring many modes for accurate reconstruction. Moreover, energy-based selection can discard crucial low-energy modes needed to capture small-scale features. Recent nonlinear manifold methods using polynomial mappings with alternating or greedy mode selection achieve better reconstruction with fewer modes. However, these methods fix the nonlinear mapping form a priori, limiting expressivity. Conversely, neural network (NN) manifolds offer greater expressivity but employ energy-based selection. We present SparseModesNet, a dimensionality reduction framework that employs linear encoding via POD modes and nonlinear NN decoding. The decoder leverages LassoNet, a method enforcing hierarchical sparsity through residual connections with linear skip layers, to simultaneously select informative POD modes and learn a nonlinear mapping that minimizes reconstruction error. On benchmark advection-dominated and chaotic flows, SparseModesNet matches or exceeds state-of-the-art performance. For turbulent channel flow at friction Reynolds number \(Re_\tau=5200\), we reduce reconstruction error by 51--78\% compared to existing polynomial manifold methods while maintaining interpretability through physically meaningful mode selection.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces SparseModesNet, a model-order-reduction framework that performs linear encoding via POD modes followed by nonlinear decoding with LassoNet. LassoNet's hierarchical sparsity (via residual linear skip connections) is used to jointly select a sparse subset of POD modes and learn a nonlinear mapping that minimizes reconstruction error. The central empirical claim is that this yields reconstruction errors 51--78% lower than existing polynomial-manifold baselines on the Re_τ=5200 turbulent channel flow while preserving interpretability through physically meaningful mode selection; comparable or better performance is also reported on advection-dominated and chaotic benchmark flows.

Significance. If the performance claims are shown to be robust to hyperparameter choice and data partitioning, the work would supply a concrete, interpretable route to nonlinear manifold reduction that retains the physical meaning of POD modes while overcoming the limitations of purely energy-based or fixed-polynomial selection. The explicit use of a sparsity-inducing architecture to automate mode retention is a distinguishing feature relative to prior greedy or alternating-selection polynomial methods.

major comments (2)
  1. [Results, Re_τ=5200 experiment] Results section (turbulent channel flow at Re_τ=5200): the headline 51--78% error-reduction claim is load-bearing for the paper's contribution, yet the manuscript supplies no description of how the LassoNet sparsity regularization strength was chosen. If this hyperparameter was tuned on the same snapshots used for the reported test error (rather than via explicit cross-validation on held-out data or a quantified sensitivity study), the advantage over polynomial baselines cannot be unambiguously attributed to the method itself.
  2. [Method, LassoNet decoder] Method section (LassoNet decoder): the hierarchical sparsity mechanism is presented as automatically retaining low-energy but dynamically important modes, but no quantitative evidence (e.g., mode-energy ranking of selected versus discarded modes, or comparison against energy-based selection at the same cardinality) is provided to demonstrate that the retained modes are indeed the dynamically relevant ones rather than an artifact of the particular regularization schedule.
minor comments (2)
  1. [Abstract and Results tables] Abstract and results tables: quantitative error reductions are stated without accompanying standard deviations, number of independent trials, or explicit dataset-split protocol; adding these would strengthen reproducibility.
  2. [Method] Notation: the precise definition of the reconstruction error metric (L2 norm over which domain, normalization, etc.) should be stated once in the method section and used consistently in all figures and tables.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which highlight opportunities to strengthen the clarity and robustness of our claims. We address each major point below and will incorporate revisions to the manuscript.

read point-by-point responses

  1. Referee: [Results, Re_τ=5200 experiment] Results section (turbulent channel flow at Re_τ=5200): the headline 51--78% error-reduction claim is load-bearing for the paper's contribution, yet the manuscript supplies no description of how the LassoNet sparsity regularization strength was chosen. If this hyperparameter was tuned on the same snapshots used for the reported test error (rather than via explicit cross-validation on held-out data or a quantified sensitivity study), the advantage over polynomial baselines cannot be unambiguously attributed to the method itself.

    Authors: We agree that the manuscript lacks an explicit description of the LassoNet regularization strength (λ) selection procedure. In the original experiments, λ was selected via grid search over a held-out validation partition (20% of training snapshots) to balance reconstruction accuracy and sparsity before final test evaluation. We will add a new subsection detailing the grid range, selected λ values for each case, and a sensitivity plot of test error versus λ. This revision will directly address concerns about hyperparameter robustness and data partitioning. revision: yes

  2. Referee: [Method, LassoNet decoder] Method section (LassoNet decoder): the hierarchical sparsity mechanism is presented as automatically retaining low-energy but dynamically important modes, but no quantitative evidence (e.g., mode-energy ranking of selected versus discarded modes, or comparison against energy-based selection at the same cardinality) is provided to demonstrate that the retained modes are indeed the dynamically relevant ones rather than an artifact of the particular regularization schedule.

    Authors: We concur that the current manuscript does not supply the requested quantitative comparisons. In the revised version we will add (i) a table or figure ranking the POD mode energies of retained versus discarded modes for the Re_τ=5200 case and (ii) a direct side-by-side reconstruction-error comparison between LassoNet-selected modes and the same number of highest-energy modes. These additions will demonstrate that the sparsity mechanism retains dynamically relevant modes beyond pure energy ranking. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical method with external benchmarks

full rationale

The paper presents SparseModesNet as a hybrid POD + LassoNet framework for mode selection and nonlinear manifold learning. All performance claims (51-78% error reduction on Re_τ=5200 channel flow, matching or exceeding polynomial baselines on advection/chaotic flows) rest on direct numerical comparisons against held-out or benchmark data rather than any derivation that reduces to its own fitted parameters or self-citations. No equations, uniqueness theorems, or ansatzes appear in the provided text; the method description simply composes two existing techniques (POD encoding + LassoNet sparsity) and reports reconstruction metrics. This is the standard non-circular case of an applied ML paper whose central results are falsifiable on external data.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

Abstract-only review; free parameters such as LassoNet regularization strength, number of retained modes, and NN architecture depth are not specified. Axioms include standard POD optimality for linear subspaces and the assumption that residual skip connections enforce useful hierarchical sparsity. No invented physical entities.

axioms (2)
  • standard math POD modes form an optimal linear basis for energy capture in the L2 sense
    Invoked implicitly when using POD as the encoder; standard result from snapshot method literature.
  • domain assumption LassoNet residual connections produce hierarchical sparsity that preserves dynamically important low-energy modes
    Central modeling choice stated in the method description; not independently verified in abstract.

pith-pipeline@v0.9.1-grok · 5819 in / 1413 out tokens · 32711 ms · 2026-06-29T14:58:51.760607+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

71 extracted references · 47 canonical work pages · 2 internal anchors

  1. [1]

    S. E. Ahmed and O. San , Breaking the Kolmogorov Barrier in Model Reduction of Fluid Flows , Fluids, 5 (2020), p. 26, https://doi.org/10.3390/fluids5010026

    work page doi:10.3390/fluids5010026 2020

  2. [2]

    A. C. Antoulas , Approximation of large-scale dynamical systems , SIAM, 2005

    2005

  3. [3]

    Nonlinear projection-based model order reduction with machine learning regression for closure error modeling in the latent space,

    S. Ares de Parga , R. Tezaur, C. G. Hern \'a ndez, and C. Farhat , Nonlinear projection-based model order reduction with machine learning regression for closure error modeling in the latent space , Computer Methods in Applied Mechanics and Engineering, 448 (2026), p. 118443, https://doi.org/10.1016/j.cma.2025.118443

    work page doi:10.1016/j.cma.2025.118443 2026

  4. [4]

    Balajewicz, I

    M. Balajewicz, I. Tezaur, and E. Dowell , Minimal subspace rotation on the stiefel manifold for stabilization and enhancement of projection-based reduced order models for the compressible navier--stokes equations , Journal of Computational Physics, 321 (2016), pp. 224--241

    2016

  5. [5]

    Barnett and C

    J. Barnett and C. Farhat , Quadratic approximation manifold for mitigating the Kolmogorov barrier in nonlinear projection-based model order reduction , Journal of Computational Physics, 464 (2022), p. 111348, https://doi.org/10.1016/j.jcp.2022.111348

    work page doi:10.1016/j.jcp.2022.111348 2022

  6. [6]

    Barnett, C

    J. Barnett, C. Farhat, and Y. Maday , Neural-network-augmented projection-based model order reduction for mitigating the Kolmogorov barrier to reducibility , Journal of Computational Physics, 492 (2023), p. 112420, https://doi.org/10.1016/j.jcp.2023.112420

    work page doi:10.1016/j.jcp.2023.112420 2023

  7. [7]

    A survey of projection-based model reduction methods for parametric dynamical systems,

    P. Benner, S. Gugercin, and K. Willcox , A Survey of Projection-Based Model Reduction Methods for Parametric Dynamical Systems , SIAM Review, 57 (2015), pp. 483--531, https://doi.org/10.1137/130932715

    work page doi:10.1137/130932715 2015

  8. [8]

    The proper orthogonal decomposition in the analysis of turbulent flows,

    G. Berkooz, P. Holmes, and J. L. Lumley , The Proper Orthogonal Decomposition in the Analysis of Turbulent Flows , Annual Review of Fluid Mechanics, 25 (1993), pp. 539--575, https://doi.org/10.1146/annurev.fl.25.010193.002543

    work page doi:10.1146/annurev.fl.25.010193.002543 1993

  9. [9]

    Boyd and L

    S. Boyd and L. Vandenberghe , Introduction to applied linear algebra: vectors, matrices, and least squares , Cambridge university press, 2018

    2018

  10. [10]

    Bredberg , On the wall boundary condition for turbulence models , Chalmers University of Technology, Department of Thermo and Fluid Dynamics

    J. Bredberg , On the wall boundary condition for turbulence models , Chalmers University of Technology, Department of Thermo and Fluid Dynamics. Internal Report 00/4. G oteborg, (2000), pp. 8--16

    2000

  11. [11]

    S. L. Brunton, J. L. Proctor, and J. N. Kutz , Discovering governing equations from data by sparse identification of nonlinear dynamical systems , Proceedings of the national academy of sciences, 113 (2016), pp. 3932--3937

    2016

  12. [12]

    Carlberg, M

    K. Carlberg, M. Barone, and H. Antil , Galerkin v. least-squares Petrov -- Galerkin projection in nonlinear model reduction , Journal of Computational Physics, 330 (2017), pp. 693--734, https://doi.org/10.1016/j.jcp.2016.10.033

    work page doi:10.1016/j.jcp.2016.10.033 2017

  13. [13]

    Champion, B

    K. Champion, B. Lusch, J. N. Kutz, and S. L. Brunton , Data-driven discovery of coordinates and governing equations , Proceedings of the National Academy of Sciences, 116 (2019), pp. 22445--22451

    2019

  14. [14]

    P. Chen, T. Hu, and J. Guilleminot , A nonlinear-manifold reduced-order model and operator learning for partial differential equations with sharp solution gradients , Computer Methods in Applied Mechanics and Engineering, 419 (2024), p. 116684, https://doi.org/10.1016/j.cma.2023.116684

    work page doi:10.1016/j.cma.2023.116684 2024

  15. [15]

    H. A. Chipman and H. Gu , Interpretable dimension reduction , Journal of applied statistics, 32 (2005), pp. 969--987

    2005

  16. [16]

    G. G. Chrysos, S. Moschoglou, G. Bouritsas, J. Deng, Y. Panagakis, and S. Zafeiriou , Deep Polynomial Neural Networks , IEEE Transactions on Pattern Analysis and Machine Intelligence, 44 (2022), pp. 4021--4034, https://doi.org/10.1109/TPAMI.2021.3058891

    work page doi:10.1109/tpami.2021.3058891 2022

  17. [17]

    Machine learning-based quadratic closures for non-intrusive reduced order models,

    G. Codega, A. Ivagnes, N. Demo, and G. Rozza , Machine Learning-Based Quadratic Closures for Non-Intrusive Reduced Order Models , SIAM Journal on Scientific Computing, (2026), pp. C505--C525, https://doi.org/10.1137/25M1766759

    work page doi:10.1137/25m1766759 2026

  18. [18]

    Cohen and R

    A. Cohen and R. DeVore , Kolmogorov widths under holomorphic mappings , IMA Journal of Numerical Analysis, 36 (2016), pp. 1--12, https://doi.org/10.1093/imanum/dru066

    work page doi:10.1093/imanum/dru066 2016

  19. [19]

    Csala and A

    H. Csala and A. Arzani , Decomposed sparse modal optimization: Interpretable reduced-order modeling of unsteady flows , International Journal of Heat and Fluid Flow, 117 (2026), p. 110124, https://doi.org/10.1016/j.ijheatfluidflow.2025.110124

    work page doi:10.1016/j.ijheatfluidflow.2025.110124 2026

  20. [20]

    Z. Dar, J. Baiges, and R. Codina , Artificial neural network based correction for reduced order models in computational fluid mechanics , Computer Methods in Applied Mechanics and Engineering, 415 (2023), p. 116232, https://doi.org/10.1016/j.cma.2023.116232

    work page doi:10.1016/j.cma.2023.116232 2023

  21. [21]

    A. N. Diaz, J. T. Needels, I. K. Tezaur, and P. J. Blonigan , Kernel Manifolds : Nonlinear-Augmentation Dimensionality Reduction Using Reproducing Kernel Hilbert Spaces , International Journal for Numerical Methods in Engineering, 126 (2025), p. e70230, https://doi.org/10.1002/nme.70230

    work page doi:10.1002/nme.70230 2025

  22. [22]

    D \'i ez, A

    P. D \'i ez, A. Muix \'i , S. Zlotnik, and A. Garc \'i a-Gonz \'a lez , Nonlinear dimensionality reduction for parametric problems: A kernel proper orthogonal decomposition , International Journal for Numerical Methods in Engineering, 122 (2021), pp. 7306--7327, https://doi.org/10.1002/nme.6831

    work page doi:10.1002/nme.6831 2021

  23. [23]

    Eckart and G

    C. Eckart and G. Young , The approximation of one matrix by another of lower rank , Psychometrika, 1 (1936), pp. 211--218

    1936

  24. [24]

    J. H. Friedman, T. Hastie, and R. Tibshirani , Regularization paths for generalized linear models via coordinate descent , Journal of statistical software, 33 (2010), pp. 1--22

    2010

  25. [25]

    Fukami, T

    K. Fukami, T. Murata, K. Zhang, and K. Fukagata , Sparse identification of nonlinear dynamics with low-dimensionalized flow representations , Journal of Fluid Mechanics, 926 (2021), p. A10

    2021

  26. [26]

    Learning latent representations in high-dimensional state spaces using polynomial manifold constructions,

    R. Geelen, L. Balzano, and K. Willcox , Learning Latent Representations in High-Dimensional State Spaces Using Polynomial Manifold Constructions , in 2023 62nd IEEE Conference on Decision and Control ( CDC ), Dec. 2023, pp. 4960--4965, https://doi.org/10.1109/CDC49753.2023.10384209

    work page doi:10.1109/cdc49753.2023.10384209 2023

  27. [27]

    Learning physics-based reduced-order models from data using nonlinear manifolds,

    R. Geelen, L. Balzano, S. Wright, and K. Willcox , Learning physics-based reduced-order models from data using nonlinear manifolds , Chaos: An Interdisciplinary Journal of Nonlinear Science, 34 (2024), p. 033122, https://doi.org/10.1063/5.0170105

    work page doi:10.1063/5.0170105 2024

  28. [28]

    Geelen, S

    R. Geelen, S. Wright, and K. Willcox , Operator inference for non-intrusive model reduction with quadratic manifolds , Computer Methods in Applied Mechanics and Engineering, 403 (2023), p. 115717, https://doi.org/10.1016/j.cma.2022.115717

    work page doi:10.1016/j.cma.2022.115717 2023

  29. [29]

    Nonlinear model reduction by probabilistic manifold decomposition,

    J. Guo and D. Xiao , Nonlinear Model Reduction by Probabilistic Manifold Decomposition , SIAM Journal on Scientific Computing, 48 (2026), pp. A209--A235, https://doi.org/10.1137/25M1738863

    work page doi:10.1137/25m1738863 2026

  30. [30]

    E. T. Hale, W. Yin, and Y. Zhang , A fixed-point continuation method for _1 -regularized minimization with applications to compressed sensing , CAAM TR07-07, Rice University, 43 (2007), p. 2

    2007

  31. [31]

    E. T. Hale, W. Yin, and Y. Zhang , Fixed-point continuation for _1 -minimization: Methodology and convergence , SIAM Journal on Optimization, 19 (2008), pp. 1107--1130

    2008

  32. [32]

    K. He, X. Zhang, S. Ren, and J. Sun , Deep Residual Learning for Image Recognition , Dec. 2015, https://doi.org/10.48550/arXiv.1512.03385

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1512.03385 2015

  33. [33]

    Holmes , Turbulence, coherent structures, dynamical systems and symmetry , Cambridge university press, 2012

    P. Holmes , Turbulence, coherent structures, dynamical systems and symmetry , Cambridge university press, 2012

    2012

  34. [34]

    Jacot, F

    A. Jacot, F. Gabriel, and C. Hongler , Neural tangent kernel: Convergence and generalization in neural networks , Advances in neural information processing systems, 31 (2018)

    2018

  35. [35]

    S. Jain, P. Tiso, J. B. Rutzmoser, and D. J. Rixen , A quadratic manifold for model order reduction of nonlinear structural dynamics , Computers & Structures, 188 (2017), pp. 80--94, https://doi.org/10.1016/j.compstruc.2017.04.005

    work page doi:10.1016/j.compstruc.2017.04.005 2017

  36. [36]

    Kadeethum, F

    T. Kadeethum, F. Ballarin, D. O'Malley, Y. Choi, N. Bouklas, and H. Yoon , Reduced order modeling for flow and transport problems with Barlow Twins self-supervised learning , Scientific Reports, 12 (2022), p. 20654, https://doi.org/10.1038/s41598-022-24545-3

    work page doi:10.1038/s41598-022-24545-3 2022

  37. [37]

    Koike, P

    T. Koike, P. Mohan, M. T. H. de Frahan, J. Bessac, and E. Qian , Streaming Operator Inference for Model Reduction of Large-Scale Dynamical Systems , Feb. 2026, https://doi.org/10.48550/arXiv.2601.12161

    work page doi:10.48550/arxiv.2601.12161 2026

  38. [38]

    Kramer, B

    B. Kramer, B. Peherstorfer, and K. E. Willcox , Learning Nonlinear Reduced Models from Data with Operator Inference , Annual Review of Fluid Mechanics, 56 (2024), p. null, https://doi.org/10.1146/annurev-fluid-121021-025220

    work page doi:10.1146/annurev-fluid-121021-025220 2024

  39. [39]

    M. B. Kuhn, M. T. Henry de Frahan , P. Mohan, G. Deskos, M. Churchfield, L. Cheung, A. Sharma, A. Almgren, S. Ananthan, M. J. Brazell, L. A. Mart \'i nez-Tossas , R. Thedin, J. Rood, P. Sakievich, G. Vijayakumar, W. Zhang, and M. Sprague , AMR-Wind : A Performance-Portable , High-Fidelity Flow Solver for Wind Farm Simulations , Wind Energy, 28 (2025), p. ...

    work page doi:10.1002/we.70010 2025

  40. [40]

    Kunisch and S

    K. Kunisch and S. Volkwein , Galerkin proper orthogonal decomposition methods for a general equation in fluid dynamics , SIAM Journal on Numerical analysis, 40 (2002), pp. 492--515

    2002

  41. [41]

    Kuramoto , Diffusion-induced chaos in reaction systems , Progress of Theoretical Physics Supplement, 64 (1978), pp

    Y. Kuramoto , Diffusion-induced chaos in reaction systems , Progress of Theoretical Physics Supplement, 64 (1978), pp. 346--367

    1978

  42. [42]

    Model reduction of dynamical systems on nonlinear manifolds using deep convolutional autoencoders,

    K. Lee and K. T. Carlberg , Model reduction of dynamical systems on nonlinear manifolds using deep convolutional autoencoders , Journal of Computational Physics, 404 (2020), p. 108973, https://doi.org/10.1016/j.jcp.2019.108973

    work page doi:10.1016/j.jcp.2019.108973 2020

  43. [43]

    Lemhadri, F

    I. Lemhadri, F. Ruan, L. Abraham, and R. Tibshirani , LassoNet : A Neural Network with Feature Sparsity , Journal of Machine Learning Research, 22 (2021), pp. 1--29

    2021

  44. [44]

    M. Ma, L. Fu, X. Guo, and Z. Zhai , Incorporating lasso regression to physics-informed neural network for inverse pde problem. , CMES-Computer Modeling in Engineering & Sciences, 141 (2024)

    2024

  45. [45]

    Mirsky , Symmetric gauge functions and unitarily invariant norms , The quarterly journal of mathematics, 11 (1960), pp

    L. Mirsky , Symmetric gauge functions and unitarily invariant norms , The quarterly journal of mathematics, 11 (1960), pp. 50--59

    1960

  46. [46]

    Parish, A

    E. Parish, A. Gruber, P. Blonigan, and I. Tezaur , NN-OpInf : An operator inference approach using structure-preserving composable neural networks , Mar. 2026, https://doi.org/10.48550/arXiv.2603.08488

    work page doi:10.48550/arxiv.2603.08488 2026

  47. [47]

    M. Y. Park and T. Hastie , l_1 -regularization path algorithm for generalized linear models , Journal of the Royal Statistical Society Series B: Statistical Methodology, 69 (2007), pp. 659--677

    2007

  48. [48]

    Breaking the Kolmogorov barrier with nonlinear model reduction,

    B. Peherstorfer , Breaking the Kolmogorov Barrier with Nonlinear Model Reduction , Notices of the American Mathematical Society, 69 (2022), p. 1, https://doi.org/10.1090/noti2475

    work page doi:10.1090/noti2475 2022

  49. [49]

    Peherstorfer and K

    B. Peherstorfer and K. Willcox , Data-driven operator inference for nonintrusive projection-based model reduction , Computer Methods in Applied Mechanics and Engineering, 306 (2016), pp. 196--215, https://doi.org/10.1016/j.cma.2016.03.025

    work page doi:10.1016/j.cma.2016.03.025 2016

  50. [50]

    H. Qiao, P. Zhang, D. Wang, and B. Zhang , An Explicit Nonlinear Mapping for Manifold Learning , IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society, 43 (2012), https://doi.org/10.1109/TSMCB.2012.2198916

    work page doi:10.1109/tsmcb.2012.2198916 2012

  51. [51]

    Quarteroni, A

    A. Quarteroni, A. Manzoni, and F. Negri , Reduced basis methods for partial differential equations: an introduction , vol. 92, Springer, 2015

    2015

  52. [52]

    Rahaman, A

    N. Rahaman, A. Baratin, D. Arpit, F. Draxler, M. Lin, F. Hamprecht, Y. Bengio, and A. Courville , On the spectral bias of neural networks , in International conference on machine learning, PMLR, 2019, pp. 5301--5310

    2019

  53. [53]

    Raissi, P

    M. Raissi, P. Perdikaris, and G. E. Karniadakis , Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations , Journal of Computational Physics, 378 (2019), pp. 686--707, https://doi.org/10.1016/j.jcp.2018.10.045

    work page doi:10.1016/j.jcp.2018.10.045 2019

  54. [54]

    S. K. Robinson , Coherent Motions in the Turbulent Boundary Layer , Annual Review of Fluid Mechanics, 23 (1991), pp. 601--639, https://doi.org/10.1146/annurev.fl.23.010191.003125

    work page doi:10.1146/annurev.fl.23.010191.003125 1991

  55. [55]

    Romor, G

    F. Romor, G. Stabile, and G. Rozza , Non-linear Manifold Reduced-Order Models with Convolutional Autoencoders and Reduced Over-Collocation Method , Journal of Scientific Computing, 94 (2023), p. 74, https://doi.org/10.1007/s10915-023-02128-2

    work page doi:10.1007/s10915-023-02128-2 2023

  56. [56]

    C. W. Rowley, T. Colonius, and R. M. Murray , Model reduction for compressible flows using POD and Galerkin projection , Physica D: Nonlinear Phenomena, 189 (2004), pp. 115--129, https://doi.org/10.1016/j.physd.2003.03.001

    work page doi:10.1016/j.physd.2003.03.001 2004

  57. [57]

    J. B. Rutzmoser, D. J. Rixen, P. Tiso, and S. Jain , Generalization of quadratic manifolds for reduced order modeling of nonlinear structural dynamics , Computers & Structures, 192 (2017), pp. 196--209, https://doi.org/10.1016/j.compstruc.2017.06.003

    work page doi:10.1016/j.compstruc.2017.06.003 2017

  58. [58]

    Empirical sparse regression on quadratic manifolds,

    P. Schwerdtner, S. Gugercin, and B. Peherstorfer , Empirical Sparse Regression on Quadratic Manifolds , SIAM Journal on Scientific Computing, 47 (2025), pp. A3085--A3107, https://doi.org/10.1137/24M1717270

    work page doi:10.1137/24m1717270 2025

  59. [59]

    Online learning of quadratic manifoldsfromstreamingdatafornonlineardimensionalityreductionandnonlinearmodelreduction,

    P. Schwerdtner, P. Mohan, A. Pachalieva, J. Bessac, D. O'Malley, and B. Peherstorfer , Online learning of quadratic manifolds from streaming data for nonlinear dimensionality reduction and nonlinear model reduction , Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 481 (2025), p. 20240670, https://doi.org/10.1098/rspa.2024.0670

    work page doi:10.1098/rspa.2024.0670 2025

  60. [60]

    Greedy construction of quadratic manifolds for nonlinear dimensionality reduction and nonlinear model reduction,

    P. Schwerdtner and B. Peherstorfer , Greedy construction of quadratic manifolds for nonlinear dimensionality reduction and nonlinear model reduction , Mar. 2024, https://doi.org/10.48550/arXiv.2403.06732

    work page doi:10.48550/arxiv.2403.06732 2024

  61. [61]

    Sharma, M

    A. Sharma, M. J. Brazell, G. Vijayakumar, S. Ananthan, L. Cheung, N. deVelder , M. T. Henry de Frahan , N. Matula, P. Mullowney, J. Rood, P. Sakievich, A. Almgren, P. S. Crozier, and M. Sprague , ExaWind : Open-source CFD for hybrid- RANS / LES geometry-resolved wind turbine simulations in atmospheric flows , Wind Energy, 27 (2024), pp. 225--257, https://...

    work page doi:10.1002/we.2886 2024

  62. [62]

    G. I. Sivashinsky , Nonlinear analysis of hydrodynamic instability in laminar flames—i. derivation of basic equations , Acta astronautica, 4 (1977), pp. 1207--1221

    1977

  63. [63]

    Solera-Rico , C

    A. Solera-Rico , C. Sanmiguel Vila, M. G \'o mez-L \'o pez , Y. Wang, A. Almashjary, S. T. M. Dawson, and R. Vinuesa , - Variational autoencoders and transformers for reduced-order modelling of fluid flows , Nature Communications, 15 (2024), p. 1361, https://doi.org/10.1038/s41467-024-45578-4

    work page doi:10.1038/s41467-024-45578-4 2024

  64. [64]

    Somasekharan and S

    N. Somasekharan and S. Pan , Beyond the Kolmogorov barrier: A learnable weighted hybrid autoencoder for model order reduction , Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 481 (2025), p. 20250159, https://doi.org/10.1098/rspa.2025.0159

    work page doi:10.1098/rspa.2025.0159 2025

  65. [65]

    Sudjianto and A

    A. Sudjianto and A. Zhang , Designing Inherently Interpretable Machine Learning Models , Nov. 2021, https://doi.org/10.48550/arXiv.2111.01743

    work page doi:10.48550/arxiv.2111.01743 2021

  66. [66]

    Tencer and K

    J. Tencer and K. Potter , A Tailored Convolutional Neural Network for Nonlinear Manifold Learning of Computational Physics Data Using Unstructured Spatial Discretizations , SIAM Journal on Scientific Computing, 43 (2021), pp. A2581--A2613, https://doi.org/10.1137/20M1344263

    work page doi:10.1137/20m1344263 2021

  67. [67]

    Wavelet-Based Observables for Koopman Analysis: An Extended Dynamic Mode Decomposition Framework

    C. Tilki and S. Gugercin , Wavelet- Based Observables for Koopman Analysis : An Extended Dynamic Mode Decomposition Framework , May 2026, https://doi.org/10.48550/arXiv.2605.14224

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2605.14224 2026

  68. [68]

    J. M. Wallace, H. Eckelmann, and R. S. Brodkey , The wall region in turbulent shear flow , Journal of Fluid Mechanics, 54 (1972), pp. 39--48, https://doi.org/10.1017/S0022112072000515

    work page doi:10.1017/s0022112072000515 1972

  69. [69]

    Y. Xia, Q. Lu, J. Feng, and H.-Y. Bae , An Explicit Sparse Mapping for Nonlinear Dimensionality Reduction , in Rough Sets and Knowledge Technology , D. Miao, W. Pedrycz, D. \'S l e zak, G. Peters, Q. Hu, and R. Wang, eds., Cham, 2014, Springer International Publishing, pp. 149--157, https://doi.org/10.1007/978-3-319-11740-9_15

    work page doi:10.1007/978-3-319-11740-9_15 2014

  70. [70]

    Zhang , Nonlinear mode decomposition via physics-assimilated convolutional autoencoder for unsteady flows over an airfoil , Physics of Fluids, 35 (2023), p

    B. Zhang , Nonlinear mode decomposition via physics-assimilated convolutional autoencoder for unsteady flows over an airfoil , Physics of Fluids, 35 (2023), p. 095115, https://doi.org/10.1063/5.0164250

    work page doi:10.1063/5.0164250 2023

  71. [71]

    write newline

    " write newline "" before.all 'output.state := FUNCTION fin.entry add.period write newline FUNCTION new.block output.state before.all = 'skip after.block 'output.state := if FUNCTION not #0 #1 if FUNCTION and 'skip pop #0 if FUNCTION or pop #1 'skip if FUNCTION new.block.checka empty 'skip 'new.block if FUNCTION field.or.null duplicate empty pop "" 'skip ...