pith. sign in

arxiv: 2607.00257 · v1 · pith:KC4VZGH4new · submitted 2026-06-30 · 💻 cs.LG · math.DS

Learning dynamical systems from noisy data with Weak-form Kernel Ridge Regression

Max Kreider , John Harlim , Daning Huang This is my paper

Pith reviewed 2026-07-02 19:24 UTC · model grok-4.3

classification 💻 cs.LG math.DS
keywords dynamical systemskernel ridge regressionweak formulationnoisy datasystem identificationchaosfluid dynamicsmachine learning
0
0 comments X

The pith

Weak-form kernel ridge regression identifies dynamical systems from noisy data by integrating residuals against test functions.

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

The paper establishes that a weak integral formulation, when embedded inside kernel ridge regression, produces a method that recovers the governing equations of dynamical systems even from noisy observations. The integration against smooth test functions damps high-frequency noise components before the regression step, yielding improved robustness without separate smoothing. The resulting WKRR framework remains accurate on clean data and is tested on chaotic benchmark systems up to 64 dimensions as well as 15,000-dimensional real fluid measurements. A reader would care because sensor data in scientific applications is routinely noisy yet accurate predictive models are needed for forecasting and control.

Core claim

The authors introduce Weak-form Kernel Ridge Regression (WKRR) that substitutes the weak form of the differential equations into the kernel ridge regression objective. This produces a linear system whose solution recovers the model parameters; the weak form supplies an inherent filtering mechanism that improves performance under noise. The approach is shown to outperform standard kernel ridge regression and other baselines on both synthetic chaotic attractors and high-dimensional fluid data.

What carries the argument

The weak-form residual, obtained by integrating the governing equation against compactly supported test functions, which supplies noise filtering before kernel ridge regression solves for the unknown coefficients.

If this is right

  • WKRR applies unchanged to both clean and noisy time-series measurements.
  • The method scales to chaotic systems in at least 64 dimensions and to fluid data with 15,000 variables.
  • A bias-variance decomposition explains why the weak form improves robustness.
  • Implementation reduces to solving a single linear system after forming the integrated kernel matrices.

Where Pith is reading between the lines

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

  • The same weak-form filtering idea could be inserted into other kernel or regression techniques used for inverse problems.
  • Direct comparison on data with structured or non-Gaussian noise would test whether the benefit is specific to the white-noise case examined here.
  • Extension to online or streaming settings might allow real-time model updates from noisy sensors.

Load-bearing premise

Integration of the residual against test functions attenuates noise enough to improve the accuracy of the subsequent kernel ridge regression step.

What would settle it

A side-by-side test on a known chaotic system with added noise where WKRR yields larger long-term trajectory errors than ordinary kernel ridge regression would refute the claimed noise-robustness benefit.

Figures

Figures reproduced from arXiv: 2607.00257 by Daning Huang, John Harlim, Max Kreider.

Figure 1
Figure 1. Figure 1: An illustration that appropriately chosen test functions can filter noisy data. A noisy signal (blue dots) is [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Top: Plots of the observed variance (42) as a function of h with the theoretical expression for V 2 ℓ = k ∗ from (35). In all cases, there is excellent agreement. Bottom: Plots of the observed bias (41) as a function of h with the approximation (39). There is good agreement for large L and small h. As h increases, the approximation worsens. 4 Proposed Approach: Weak-form Kernel Ridge Regression In this sec… view at source ↗
Figure 3
Figure 3. Figure 3: Left: A comparison between ground truth (black) and a typical WKRR forecast (dashed purple) for the L63 system under 5% noise corruption. The VPT, marked with a green line, is approximately 2.18. Right: The butterfly attractor formed from the ground truth (black) and the WKRR reconstruction (dashed purple) over 5000 timesteps. We now compare the following frameworks across noise intensities: (i) strong KRR… view at source ↗
Figure 4
Figure 4. Figure 4: Empirical VPT densities for the L63 system [PITH_FULL_IMAGE:figures/full_fig_p014_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Invariant measure comparison for the learned Lorenz-63 system [PITH_FULL_IMAGE:figures/full_fig_p015_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Ground truth data (right) and a typical WKRR forecast (middle) with 5% noise corruption. The error (right) [PITH_FULL_IMAGE:figures/full_fig_p016_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Empirical VPT densities for the KS system [PITH_FULL_IMAGE:figures/full_fig_p017_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Mean VPT statistics as a function of validation length [PITH_FULL_IMAGE:figures/full_fig_p018_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Empirical VPT densities over clean data for [PITH_FULL_IMAGE:figures/full_fig_p019_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: A visualization of the data and typical forecasts. The first four rows correspond to [PITH_FULL_IMAGE:figures/full_fig_p021_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: NMSE plots for WKRR (left), LSTM (middle), and a comparison of the two (right). Light gray lines [PITH_FULL_IMAGE:figures/full_fig_p021_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Typical data and forecast. The first five rows correspond to [PITH_FULL_IMAGE:figures/full_fig_p022_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: NMSE plots for WKRR (left), LSTM (middle), and a comparison of the two (right). Light gray lines [PITH_FULL_IMAGE:figures/full_fig_p022_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: A comparison of the filtering properties of wavelets (blue) and polynomial test functions (orange) with an [PITH_FULL_IMAGE:figures/full_fig_p024_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: A comparison of the filtering properties of wavelets (blue) and polynomial test functions (orange) with an [PITH_FULL_IMAGE:figures/full_fig_p025_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Typical coarse (left sub-columns) and fine (right sub-columns) validation landscapes for the L63 system [PITH_FULL_IMAGE:figures/full_fig_p027_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Typical coarse (left sub-columns) and fine (right sub-columns) validation landscapes for the KS system [PITH_FULL_IMAGE:figures/full_fig_p028_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: Typical coarse (left) and fine (right) validation landscapes for the Community Challenge fluid data with [PITH_FULL_IMAGE:figures/full_fig_p028_18.png] view at source ↗
read the original abstract

Accurate prediction of complex dynamical systems from noisy measurements remains a significant challenge in scientific computing. Kernel ridge regression learning strategies are often effective when applied to clean data, but have limited success with noisy data. Recent work has observed that a weak formulation can act to filter noisy data, and different learning strategies have achieved increased noise robustness with a weak-form framework. In this manuscript, we give an overview of the filtering mechanism behind the weak formulation and provide a bias-variance error decomposition. Using these insights, we combine a weak formulation with a kernel learning strategy to propose Weak-form Kernel Ridge Regression (WKRR) for learning dynamical systems. The proposed framework is simple to implement, effective for both clean and noisy data, and outperforms several baseline methods. We demonstrate the performance of WKRR on chaotic benchmark systems in up to 64 dimensions, as well as 15,000-dimensional real-world fluid data.

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

0 major / 3 minor

Summary. The manuscript proposes Weak-form Kernel Ridge Regression (WKRR) for identifying dynamical systems from noisy measurements. It reviews the noise-filtering property of weak formulations, derives a bias-variance error decomposition to motivate the approach, combines the weak form with kernel ridge regression, and reports that the resulting method is simple to implement, performs well on both clean and noisy data, and outperforms several baselines on chaotic benchmark systems (up to 64 dimensions) as well as 15,000-dimensional real-world fluid data.

Significance. If the central claims hold, the work would be a useful contribution to data-driven modeling of dynamical systems. The explicit bias-variance analysis and the combination of weak-form filtering with kernel methods address a practical limitation of standard KRR on noisy data. Demonstrations on high-dimensional chaotic systems and real fluid data provide evidence of scalability and applicability; the emphasis on simplicity and reproducibility is a positive feature.

minor comments (3)
  1. [Abstract] Abstract: the claim of outperformance is stated without naming the baseline methods or reporting quantitative metrics; a single sentence summarizing the key numerical improvements would strengthen the abstract.
  2. [Section 3 (or wherever the decomposition appears)] The bias-variance decomposition is central to the motivation; ensure that the decomposition is stated with explicit dependence on the weak-form test functions and the kernel hyperparameters so that readers can trace how the filtering effect enters the variance term.
  3. [Results section on fluid data] Figure captions and axis labels for the high-dimensional fluid example should explicitly state the noise level, the number of snapshots, and the dimension reduction (if any) used prior to applying WKRR.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive summary, significance assessment, and recommendation of minor revision. The feedback highlights the value of the bias-variance analysis and the practical combination of weak-form filtering with kernel methods. No specific major comments were provided in the report, so we have no points to address point-by-point. We will incorporate any minor suggestions during revision.

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper's core contribution is the WKRR method obtained by pairing an existing weak-form filtering insight with kernel ridge regression, followed by a bias-variance analysis and empirical validation on benchmark systems and fluid data. No equations, parameter fits, or uniqueness claims in the abstract or summary reduce by construction to the inputs; the weak-form filtering is presented as an external observation rather than a self-derived tautology. The derivation chain therefore remains independent of the target results and does not exhibit any of the enumerated circularity patterns.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review supplies no equations or methods section, so no free parameters, axioms, or invented entities can be identified.

pith-pipeline@v0.9.1-grok · 5681 in / 1039 out tokens · 29885 ms · 2026-07-02T19:24:57.156746+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

86 extracted references · 7 canonical work pages · 2 internal anchors

  1. [1]

    Exploration and prediction of fluid dynamical systems using auto-encoder technology.Physics of Fluids, 32(6), 2020

    Lionel Agostini. Exploration and prediction of fluid dynamical systems using auto-encoder technology.Physics of Fluids, 32(6), 2020

    2020

  2. [2]

    Kernel ridge regression hybrid method for wheat yield prediction with satellite-derived predictors.Remote Sensing, 14(5):1136, 2022

    AA Masrur Ahmed, Ekta Sharma, S Janifer Jabin Jui, Ravinesh C Deo, Thong Nguyen-Huy, and Mumtaz Ali. Kernel ridge regression hybrid method for wheat yield prediction with satellite-derived predictors.Remote Sensing, 14(5):1136, 2022

    2022

  3. [3]

    Sampling procedures in function spaces and asymptotic equivalence with Shannon’ s sampling theory.Numerical functional analysis and optimization, 15(1-2):1–21, 1994

    Akram Aldroubi, Michael Unser, and Akram Aldroubi. Sampling procedures in function spaces and asymptotic equivalence with Shannon’ s sampling theory.Numerical functional analysis and optimization, 15(1-2):1–21, 1994

    1994

  4. [4]

    Complete ensemble empirical mode decomposition hybridized with random forest and kernel ridge regression model for monthly rainfall forecasts

    Mumtaz Ali, Ramendra Prasad, Yong Xiang, and Zaher Mundher Yaseen. Complete ensemble empirical mode decomposition hybridized with random forest and kernel ridge regression model for monthly rainfall forecasts. Journal of Hydrology, 584:124647, 2020

    2020

  5. [5]

    Dynamic data-driven local traffic state estimation and prediction.Transportation Research Part C: Emerging Technologies, 34:89–107, 2013

    Constantinos Antoniou, Haris N Koutsopoulos, and George Yannis. Dynamic data-driven local traffic state estimation and prediction.Transportation Research Part C: Emerging Technologies, 34:89–107, 2013

    2013

  6. [6]

    Data-driven analysis and forecasting of highway traffic dynamics.Nature communications, 11(1):2090, 2020

    Allan M Avila and Igor Mezi´ c. Data-driven analysis and forecasting of highway traffic dynamics.Nature communications, 11(1):2090, 2020

    2090

  7. [7]

    Approximation error for quasi-interpolators and (multi-) wavelet expansions

    Thierry Blu and Michael Unser. Approximation error for quasi-interpolators and (multi-) wavelet expansions. Applied and Computational Harmonic Analysis, 6(2):219–251, 1999

    1999

  8. [8]

    Weak form-based data-driven modeling: computationally efficient and noise robust equation learning and parameter inference

    David M Bortz, Daniel A Messenger, and April Tran. Weak form-based data-driven modeling: computationally efficient and noise robust equation learning and parameter inference. InHandbook of Numerical Analysis, volume 25, pages 53–82. Elsevier, 2024

    2024

  9. [9]

    John Wiley & Sons, 2015

    George EP Box, Gwilym M Jenkins, Gregory C Reinsel, and Greta M Ljung.Time series analysis: forecasting and control. John Wiley & Sons, 2015

    2015

  10. [10]

    Predicting regime changes and durations in Lorenz’ s atmospheric convection model.Chaos: An Interdisciplinary Journal of Nonlinear Science, 30(10), 2020

    Eduardo L Brugnago, Jason AC Gallas, and Marcus W Beims. Predicting regime changes and durations in Lorenz’ s atmospheric convection model.Chaos: An Interdisciplinary Journal of Nonlinear Science, 30(10), 2020

    2020

  11. [11]

    Cambridge University Press, 2022

    Steven L Brunton and J Nathan Kutz.Data-driven science and engineering: Machine learning, dynamical systems, and control. Cambridge University Press, 2022

    2022

  12. [12]

    Discovering governing equations from data by sparse identification of nonlinear dynamical systems.Proceedings of the national academy of sciences, 113(15):3932– 3937, 2016

    Steven L Brunton, Joshua L Proctor, and J Nathan Kutz. Discovering governing equations from data by sparse identification of nonlinear dynamical systems.Proceedings of the national academy of sciences, 113(15):3932– 3937, 2016

    2016

  13. [13]

    Sparse identification of nonlinear dynamics with control (SINDYc).IFAC-PapersOnLine, 49(18):710–715, 2016

    Steven L Brunton, Joshua L Proctor, and J Nathan Kutz. Sparse identification of nonlinear dynamics with control (SINDYc).IFAC-PapersOnLine, 49(18):710–715, 2016

    2016

  14. [14]

    An intelligent system for financial time series prediction combining dynami- cal systems theory, fractal theory, and statistical methods

    Oscar Castillo and Patricia Melin. An intelligent system for financial time series prediction combining dynami- cal systems theory, fractal theory, and statistical methods. InProceedings of 1995 Conference on Computational Intelligence for Financial Engineering (CIFEr), pages 151–155. IEEE, 1995

    1995

  15. [15]

    Neural ordinary differential equations.Advances in neural information processing systems, 31, 2018

    Ricky TQ Chen, Yulia Rubanova, Jesse Bettencourt, and David K Duvenaud. Neural ordinary differential equations.Advances in neural information processing systems, 31, 2018

    2018

  16. [16]

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

    Sibo Cheng, César Quilodrán-Casas, Said Ouala, Alban Farchi, Che Liu, Pierre Tandeo, Ronan Fablet, Di- dier Lucor, Bertrand Iooss, Julien Brajard, et al. Machine learning with data assimilation and uncertainty quantification for dynamical systems: a review.IEEE/CAA Journal of Automatica Sinica, 10(6):1361–1387, 2023

    2023

  17. [17]

    From reliable weather forecasts to skilful climate response: A dynamical systems approach.Quarterly Journal of the Royal Meteorological Society, 145(720):1052–1069, 2019

    Hannah M Christensen and Judith Berner. From reliable weather forecasts to skilful climate response: A dynamical systems approach.Quarterly Journal of the Royal Meteorological Society, 145(720):1052–1069, 2019

    2019

  18. [18]

    Diffusion maps.Applied and computational harmonic analysis, 21(1):5–30, 2006

    Ronald R Coifman and Stéphane Lafon. Diffusion maps.Applied and computational harmonic analysis, 21(1):5–30, 2006

    2006

  19. [19]

    Graph Laplacian tomography from unknown random projections.IEEE Transactions on Image Processing, 17(10):1891–1899, 2008

    Ronald R Coifman, Yoel Shkolnisky, Fred J Sigworth, and Amit Singer. Graph Laplacian tomography from unknown random projections.IEEE Transactions on Image Processing, 17(10):1891–1899, 2008

    2008

  20. [20]

    The mpedmd algorithm for data-driven computations of measure-preserving dynamical systems.SIAM Journal on Numerical Analysis, 61(3):1585–1608, 2023

    Matthew J Colbrook. The mpedmd algorithm for data-driven computations of measure-preserving dynamical systems.SIAM Journal on Numerical Analysis, 61(3):1585–1608, 2023

    2023

  21. [21]

    Residual dynamic mode decomposition: robust and verified Koopmanism.Journal of Fluid Mechanics, 955:A21, 2023

    Matthew J Colbrook, Lorna J Ayton, and Máté Sz˝ oke. Residual dynamic mode decomposition: robust and verified Koopmanism.Journal of Fluid Mechanics, 955:A21, 2023. 29 APREPRINT- JULY2, 2026

    2023

  22. [22]

    Lyapunov exponents of the Kuramoto–Sivashinsky PDE.The ANZIAM Journal, 61(3):270–285, 2019

    Russell A Edson, Judith E Bunder, Trent W Mattner, and Anthony J Roberts. Lyapunov exponents of the Kuramoto–Sivashinsky PDE.The ANZIAM Journal, 61(3):270–285, 2019

    2019

  23. [23]

    Combining physics-based and data-driven modeling in well construction: Hybrid fluid dynamics modeling.Journal of Natural Gas Science and Engineering, 97:104348, 2022

    Oney Erge and Eric Van Oort. Combining physics-based and data-driven modeling in well construction: Hybrid fluid dynamics modeling.Journal of Natural Gas Science and Engineering, 97:104348, 2022

    2022

  24. [24]

    Nonlinear forecasting with many predictors using kernel ridge regression.International Journal of Forecasting, 32(3):736–753, 2016

    Peter Exterkate, Patrick JF Groenen, Christiaan Heij, and Dick van Dijk. Nonlinear forecasting with many predictors using kernel ridge regression.International Journal of Forecasting, 32(3):736–753, 2016

    2016

  25. [25]

    Data-driven discovery of intrinsic dynamics.Nature Machine Intelli- gence, 4(12):1113–1120, 2022

    Daniel Floryan and Michael D Graham. Data-driven discovery of intrinsic dynamics.Nature Machine Intelli- gence, 4(12):1113–1120, 2022

    2022

  26. [26]

    Next generation reservoir computing

    Daniel J Gauthier, Erik Bollt, Aaron Griffith, and Wendson AS Barbosa. Next generation reservoir computing. Nature communications, 12(1):5564, 2021

    2021

  27. [27]

    Data-driven prediction in dynamical systems: recent developments

    Amin Ghadami and Bogdan I Epureanu. Data-driven prediction in dynamical systems: recent developments. Philosophical transactions. Series A, Mathematical, physical, and engineering sciences, 380(2229):20210213, 2022

    2022

  28. [28]

    Learning dynamics from large biological data sets: machine learning meets systems biology.Current Opinion in Systems Biology, 22:1–7, 2020

    William Gilpin, Yitong Huang, and Daniel B Forger. Learning dynamics from large biological data sets: machine learning meets systems biology.Current Opinion in Systems Biology, 22:1–7, 2020

    2020

  29. [29]

    PhD thesis, University of the Balearic Islands (UIB); Institute for Cross-Disciplinary Physics and Complex Systems, 2024

    Àlex Giménez-Romero.Theoretical and data-driven models in Ecology. PhD thesis, University of the Balearic Islands (UIB); Institute for Cross-Disciplinary Physics and Complex Systems, 2024

    2024

  30. [30]

    Gaussian process priors with uncertain inputs application to multiple-step ahead time series forecasting.Advances in neural information processing systems, 15, 2002

    Agathe Girard, Carl Rasmussen, Joaquin Q Candela, and Roderick Murray-Smith. Gaussian process priors with uncertain inputs application to multiple-step ahead time series forecasting.Advances in neural information processing systems, 15, 2002

    2002

  31. [31]

    Georg A Gottwald and Sebastian Reich. Combining machine learning and data assimilation to forecast dynamical systems from noisy partial observations.Chaos: An Interdisciplinary Journal of Nonlinear Science, 31(10), 2021

    2021

  32. [32]

    Supervised learning from noisy observations: Combining machine- learning techniques with data assimilation.Physica D: Nonlinear Phenomena, 423:132911, 2021

    Georg A Gottwald and Sebastian Reich. Supervised learning from noisy observations: Combining machine- learning techniques with data assimilation.Physica D: Nonlinear Phenomena, 423:132911, 2021

    2021

  33. [33]

    Neural ODEs with irregular and noisy data.arXiv preprint arXiv:2205.09479, 2022

    Pawan Goyal and Peter Benner. Neural ODEs with irregular and noisy data.arXiv preprint arXiv:2205.09479, 2022

    work page arXiv 2022

  34. [34]

    Diffusion maps kernel ridge regression.arXiv preprint, in preparation, 2026

    John Harlim, Daning Huang, Jiwoo Song, and Alex Townsend. Diffusion maps kernel ridge regression.arXiv preprint, in preparation, 2026

    2026

  35. [35]

    Long short-term memory.Neural computation, 9(8):1735–1780, 1997

    Sepp Hochreiter and Jürgen Schmidhuber. Long short-term memory.Neural computation, 9(8):1735–1780, 1997

    1997

  36. [36]

    Learning vector fields of differential equations on manifolds with geometrically constrained operator-valued kernels

    Daning Huang, Hanyang He, John Harlim, and Yan Li. Learning vector fields of differential equations on manifolds with geometrically constrained operator-valued kernels. InThe Thirteenth International Conference on Learning Representations, 2025

    2025

  37. [37]

    A dynamic neural network architecture with immunology inspired optimization for weather data forecasting.Big data research, 14:81–92, 2018

    Abir Jaafar Hussain, Panos Liatsis, Mohammed Khalaf, Hissam Tawfik, and Haya Al-Asker. A dynamic neural network architecture with immunology inspired optimization for weather data forecasting.Big data research, 14:81–92, 2018

    2018

  38. [38]

    SINDy-PI: a robust algorithm for parallel implicit sparse identification of nonlinear dynamics.Proceedings

    Kadierdan Kaheman, J Nathan Kutz, and Steven L Brunton. SINDy-PI: a robust algorithm for parallel implicit sparse identification of nonlinear dynamics.Proceedings. Mathematical, physical, and engineering sciences, 476(2242):20200279, 2020

    2020

  39. [39]

    A new approach to linear filtering and prediction problems.Transactions of the ASME–Journal of Basic Engineering, 1960

    Rudolph Emil Kalman. A new approach to linear filtering and prediction problems.Transactions of the ASME–Journal of Basic Engineering, 1960

    1960

  40. [40]

    Fourth-order time-stepping for stiff PDEs.SIAM Journal on Scientific Computing, 26(4):1214–1233, 2005

    Aly-Khan Kassam and Lloyd N Trefethen. Fourth-order time-stepping for stiff PDEs.SIAM Journal on Scientific Computing, 26(4):1214–1233, 2005

    2005

  41. [41]

    SIAM, 2016

    J Nathan Kutz, Steven L Brunton, Bingni W Brunton, and Joshua L Proctor.Dynamic mode decomposition: data-driven modeling of complex systems. SIAM, 2016

    2016

  42. [42]

    The Lorenz system: hidden boundary of practical stability and the Lyapunov dimension.Nonlinear Dyn, 102:713–732, 2020

    NV Kuznetsov, TN Mokaev, OA Kuznetsova, and EV Kudryashova. The Lorenz system: hidden boundary of practical stability and the Lyapunov dimension.Nonlinear Dyn, 102:713–732, 2020

    2020

  43. [43]

    Qianxiao Li, Felix Dietrich, Erik M Bollt, and Ioannis G Kevrekidis. Extended dynamic mode decomposition with dictionary learning: A data-driven adaptive spectral decomposition of the Koopman operator.Chaos: An Interdisciplinary Journal of Nonlinear Science, 27(10), 2017. 30 APREPRINT- JULY2, 2026

    2017

  44. [44]

    A Weak Penalty Neural ODE for Learning Chaotic Dynamics from Noisy Time Series

    Xuyang Li, John Harlim, Dibyajyoti Chakraborty, and Romit Maulik. A weak penalty neural ODE for learning chaotic dynamics from noisy time series.arXiv preprint arXiv:2511.06609, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  45. [45]

    A survey on long short-term memory networks for time series prediction.Procedia Cirp, 99:650–655, 2021

    Benjamin Lindemann, Timo Müller, Hannes Vietz, Nasser Jazdi, and Michael Weyrich. A survey on long short-term memory networks for time series prediction.Procedia Cirp, 99:650–655, 2021

    2021

  46. [46]

    On a measure of lack of fit in time series models.Biometrika, 65(2):297–303, 1978

    Greta M Ljung and George EP Box. On a measure of lack of fit in time series models.Biometrika, 65(2):297–303, 1978

    1978

  47. [47]

    Hybridnet: integrating model-based and data-driven learning to predict evolution of dynamical systems

    Yun Long, Xueyuan She, and Saibal Mukhopadhyay. Hybridnet: integrating model-based and data-driven learning to predict evolution of dynamical systems. InConference on robot learning, pages 551–560. PMLR, 2018

    2018

  48. [48]

    Deterministic nonperiodic flow 1

    Edward N Lorenz. Deterministic nonperiodic flow 1. InUniversality in Chaos, 2nd edition, pages 367–378. Routledge, 2017

    2017

  49. [49]

    Ecological forecasting and data assimilation in a data-rich era.Ecological Applications, 21(5):1429– 1442, 2011

    Yiqi Luo, Kiona Ogle, Colin Tucker, Shenfeng Fei, Chao Gao, Shannon LaDeau, James S Clark, and David S Schimel. Ecological forecasting and data assimilation in a data-rich era.Ecological Applications, 21(5):1429– 1442, 2011

    2011

  50. [50]

    Academic press, 1982

    Peter S Maybeck.Stochastic models, estimation, and control, volume 3. Academic press, 1982

    1982

  51. [51]

    Weak SINDy for partial differential equations.Journal of Computational Physics, 443:110525, 2021

    Daniel A Messenger and David M Bortz. Weak SINDy for partial differential equations.Journal of Computational Physics, 443:110525, 2021

    2021

  52. [52]

    Weak SINDy: Galerkin-based data-driven model selection.Multiscale Modeling & Simulation, 19(3):1474–1497, 2021

    Daniel A Messenger and David M Bortz. Weak SINDy: Galerkin-based data-driven model selection.Multiscale Modeling & Simulation, 19(3):1474–1497, 2021

    2021

  53. [53]

    Asymptotic consistency of the WSINDy algorithm in the limit of continuum data.IMA Journal of Numerical Analysis, 45(6):3264–3312, 2025

    Daniel A Messenger and David M Bortz. Asymptotic consistency of the WSINDy algorithm in the limit of continuum data.IMA Journal of Numerical Analysis, 45(6):3264–3312, 2025

    2025

  54. [54]

    The weak form is stronger than you think

    Daniel A Messenger, April Tran, Vanja Dukic, and David M Bortz. The weak form is stronger than you think. arXiv preprint arXiv:2409.06751, 2024

    work page arXiv 2024

  55. [55]

    On numerical approximations of the Koopman operator.Mathematics, 10(7):1180, 2022

    Igor Mezi´ c. On numerical approximations of the Koopman operator.Mathematics, 10(7):1180, 2022

    2022

  56. [56]

    Springer, 2021

    Kohei Nakajima and Ingo Fischer.Reservoir computing. Springer, 2021

    2021

  57. [57]

    Data-driven methods for weather forecast

    Elias David Nino-Ruiz and Felipe J Acevedo García. Data-driven methods for weather forecast. InInternational Conference on Computational Science, pages 326–336. Springer, 2021

    2021

  58. [58]

    A review of data-driven discovery for dynamic systems.International Statistical Review, 91(3):464–492, 2023

    Joshua S North, Christopher K Wikle, and Erin M Schliep. A review of data-driven discovery for dynamic systems.International Statistical Review, 91(3):464–492, 2023

    2023

  59. [59]

    Comprehensive review of neural differential equations for time series analysis, 2025

    YongKyung Oh, Seungsu Kam, Jonghun Lee, Dong-Young Lim, Sungil Kim, and Alex Bui. Comprehensive review of neural differential equations for time series analysis.arXiv preprint arXiv:2502.09885, 2025

    work page arXiv 2025

  60. [60]

    Data-driven discovery of dynamical models in biology.arXiv preprint arXiv:2509.06735, 2025

    Bartosz Prokop and Lendert Gelens. Data-driven discovery of dynamical models in biology.arXiv preprint arXiv:2509.06735, 2025

    work page internal anchor Pith review arXiv 2025

  61. [61]

    Smoothing and differentiation of data by simplified least squares procedures.Analytical chemistry, 36(8):1627–1639, 1964

    Abraham Savitzky and Marcel JE Golay. Smoothing and differentiation of data by simplified least squares procedures.Analytical chemistry, 36(8):1627–1639, 1964

    1964

  62. [62]

    Data-driven reduced-complexity modeling of fluid flows: A community challenge.arXiv preprint arXiv:2601.06183, 2026

    Oliver T Schmidt, Aaron Towne, Adrian Lozano-Duran, Scott Dawson, and Ricardo Vinuesa. Data-driven reduced-complexity modeling of fluid flows: A community challenge.arXiv preprint arXiv:2601.06183, 2026

    work page arXiv 2026

  63. [63]

    Communication in the presence of noise.Proceedings of the IRE, 37(1):10–21, 1949

    Claude E Shannon. Communication in the presence of noise.Proceedings of the IRE, 37(1):10–21, 1949

    1949

  64. [64]

    Application of dynamic data driven application system in environmental science.Environmental Reviews, 22(3):287–297, 2014

    Jingwei Song, Bo Xiang, Xinyuan Wang, Li Wu, and Chun Chang. Application of dynamic data driven application system in environmental science.Environmental Reviews, 22(3):287–297, 2014

    2014

  65. [65]

    Learning solution operator of dynamical systems with diffusion maps kernel ridge regression.arXiv preprint arXiv:2512.17203, 2025

    Jiwoo Song, Daning Huang, and John Harlim. Learning solution operator of dynamical systems with diffusion maps kernel ridge regression.arXiv preprint arXiv:2512.17203, 2025

    work page arXiv 2025

  66. [66]

    Recent advances in physical reservoir computing: A review

    Gouhei Tanaka, Toshiyuki Yamane, Jean Benoit Héroux, Ryosho Nakane, Naoki Kanazawa, Seiji Takeda, Hidetoshi Numata, Daiju Nakano, and Akira Hirose. Recent advances in physical reservoir computing: A review. Neural Networks, 115:100–123, 2019

    2019

  67. [67]

    Sampling-50 years after Shannon.Proceedings of the IEEE, 88(4):569–587, 2002

    Michael Unser. Sampling-50 years after Shannon.Proceedings of the IEEE, 88(4):569–587, 2002

    2002

  68. [68]

    A general sampling theory for nonideal acquisition devices.IEEE Transactions on Signal Processing, 42(11):2915–2925, 2002

    Michael Unser and Akram Aldroubi. A general sampling theory for nonideal acquisition devices.IEEE Transactions on Signal Processing, 42(11):2915–2925, 2002

    2002

  69. [69]

    Polynomial spline signal approximations: filter design and asymptotic equivalence with Shannon’ s sampling theorem.IEEE Transactions on Information Theory, 38(1):95–103, 2002

    Michael Unser, Akram Aldroubi, and Murray Eden. Polynomial spline signal approximations: filter design and asymptotic equivalence with Shannon’ s sampling theorem.IEEE Transactions on Information Theory, 38(1):95–103, 2002. 31 APREPRINT- JULY2, 2026

    2002

  70. [70]

    A generalized sampling theory without band-limiting constraints.IEEE transactions on circuits and systems II: analog and digital signal processing, 45(8):959–969, 2002

    Michael Unser and Josiane Zerubia. A generalized sampling theory without band-limiting constraints.IEEE transactions on circuits and systems II: analog and digital signal processing, 45(8):959–969, 2002

    2002

  71. [71]

    Pantelis R Vlachas, Wonmin Byeon, Zhong Y Wan, Themistoklis P Sapsis, and Petros Koumoutsakos. Data- driven forecasting of high-dimensional chaotic systems with long short-term memory networks.Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 474(2213):20170844, 2018

    2018

  72. [72]

    Kernel ridge regression

    Vladimir Vovk. Kernel ridge regression. InEmpirical inference: Festschrift in honor of vladimir n. vapnik, pages 105–116. Springer, 2013

    2013

  73. [73]

    Data-driven neural modeling and chaos control in fractional-order financial dynamical systems.AIP Advances, 16(1), 2026

    Shamaila Waheed, Mubashir Qayyum, Omar Khan, and Gilbert Chambashi. Data-driven neural modeling and chaos control in fractional-order financial dynamical systems.AIP Advances, 16(1), 2026

    2026

  74. [74]

    Gaussian process dynamical models.Advances in neural information processing systems, 18, 2005

    Jack Wang, Aaron Hertzmann, and David J Fleet. Gaussian process dynamical models.Advances in neural information processing systems, 18, 2005

    2005

  75. [75]

    A data–driven approximation of the Koopman operator: Extending dynamic mode decomposition.Journal of Nonlinear Science, 25(6):1307–1346, 2015

    Matthew O Williams, Ioannis G Kevrekidis, and Clarence W Rowley. A data–driven approximation of the Koopman operator: Extending dynamic mode decomposition.Journal of Nonlinear Science, 25(6):1307–1346, 2015

    2015

  76. [76]

    Introduction to ‘communication in the presence of noise’ by CE Shannon

    Aaron D Wyner and Shlomo Shamai. Introduction to ‘communication in the presence of noise’ by CE Shannon. Proc. IEEE, 86(2):442–446, 1998

    1998

  77. [77]

    Reconstructing data-driven governing equations for cell phenotypic transitions: integration of data science and systems biology.Physical Biology, 19(6):061001, 2022

    Jianhua Xing. Reconstructing data-driven governing equations for cell phenotypic transitions: integration of data science and systems biology.Physical Biology, 19(6):061001, 2022

    2022

  78. [78]

    Big data driven mobile traffic understanding and forecasting: A time series approach.IEEE transactions on services computing, 9(5):796–805, 2016

    Fengli Xu, Yuyun Lin, Jiaxin Huang, Di Wu, Hongzhi Shi, Jeungeun Song, and Yong Li. Big data driven mobile traffic understanding and forecasting: A time series approach.IEEE transactions on services computing, 9(5):796–805, 2016

    2016

  79. [79]

    Emerging opportunities and challenges for the future of reservoir computing.Nature Communications, 15(1):2056, 2024

    Min Yan, Can Huang, Peter Bienstman, Peter Tino, Wei Lin, and Jie Sun. Emerging opportunities and challenges for the future of reservoir computing.Nature Communications, 15(1):2056, 2024

    2056

  80. [80]

    Gaussian process for long-term time-series forecasting

    Weizhong Yan, Hai Qiu, and Ya Xue. Gaussian process for long-term time-series forecasting. In2009 interna- tional joint conference on neural networks, pages 3420–3427. IEEE, 2009

    2009

Showing first 80 references.