Pith. sign in

REVIEW 3 major objections 2 minor 57 references

PDE Identification Using Noise Adaptive Differentiation in Strong Form (S-IDENT)

T0 review · 3 major / 2 minor · reviewed 2026-07-01 · grok-4.3

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

Pith's one-line read S-IDENT identifies nonlinear PDEs from noisy single-trajectory data using strong-form dictionaries built from adaptive Savitzky-Golay derivatives.

desk verdict S-IDENT pairs SG differentiation with SURE window selection for strong-form PDE discovery and claims an edge at higher noise, but the abstract gives almost no quantitative detail to judge how well it holds up. read the letter →

arxiv 2606.31776 v1 pith:2VIZOIPH submitted 2026-06-30 math.NA cs.NA

Roy Y. He , Sung Ha Kang This is my paper · ORCID
classification math.NAcs.NA
keywords PDEidentificationstrongformSavitzky-GolaySUREnoisydatanonlinearPDEsdictionarylearningmodelselection
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper develops a method to recover both linear and nonlinear PDEs directly from noisy space-time observations by working in the strong form rather than the weak form. It replaces fixed-window differentiation with Savitzky-Golay smoothing whose window length is chosen automatically by Stein's Unbiased Risk Estimate, then applies trimming and residual-based model selection to the resulting dictionary. Numerical tests show that this approach recovers the correct PDE at noise levels where earlier strong-form techniques break down and produces accuracy comparable to state-of-the-art weak-form methods. The method also tolerates the larger, more ill-conditioned dictionaries that arise when all possible differential terms are retained.

What carries the argument

Savitzky-Golay differentiation with SURE-selected adaptive window length, used to build the strong-form differential dictionary before trimming and residual-based selection.

What would settle it

A controlled test in which the same noisy trajectory is fed to both S-IDENT and a known ground-truth PDE; if the selected equation differs from the true one at noise levels where the paper reports success, the method's reliability claim is refuted.

Watch

Extended reading notes

Core claim

By computing differential features with Savitzky-Golay differentiation whose window is chosen via SURE and then performing trimmed residual-based model selection on a strong-form dictionary, S-IDENT recovers the governing nonlinear PDE from noisy single-trajectory data at higher noise amplitudes than prior strong-form techniques while remaining competitive with weak-form approaches.

Load-bearing premise

The adaptive Savitzky-Golay step produces differential features whose bias and variance remain low enough that dictionary regression can still select the correct PDE terms without systematic error from the differentiation itself.

Editorial extensions

If this is right

  • Strong-form dictionaries become usable for general nonlinear PDEs without requiring integration against test functions.
  • Model selection remains stable even though the full differential dictionary is larger and more ill-conditioned than weak-form dictionaries.
  • Trimming and reduction-in-residual selection further reduce false positives in the identified terms.
  • The approach works on single space-time trajectories without needing multiple realizations or ensemble data.

Reading between the lines

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

  • The same adaptive differentiation pipeline could be tested on inverse problems where the PDE is known but coefficients or boundary conditions must be recovered.
  • If the SURE window choice proves robust across different equation types, it might reduce the need for hand-tuned regularization parameters in other data-driven PDE methods.
  • Extending the dictionary to include fractional or nonlocal operators would require only a change in the differentiation routine while keeping the rest of the pipeline intact.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 2 minor

Summary. The manuscript proposes S-IDENT, a strong-form dictionary method for identifying linear and nonlinear PDEs from noisy single-trajectory observations. Derivatives are obtained via Savitzky-Golay differentiation whose window length is chosen by Stein's Unbiased Risk Estimate (SURE); the resulting features are fed to a trimmed, reduction-in-residual model-selection procedure. Numerical experiments are reported to show that the method recovers nonlinear PDEs at higher noise levels than prior strong-form approaches while remaining competitive with weak-form methods, even though the strong-form dictionaries are larger and more ill-conditioned.

Significance. If the numerical claims hold under more detailed scrutiny, the work supplies a practical strong-form alternative that widens the range of noise levels at which dictionary-based PDE identification remains reliable. The explicit use of an external statistical criterion (SURE) for derivative estimation is a clear technical contribution that could be adopted more broadly.

major comments (3)
  1. [Abstract and results (presumably §5)] The abstract and results sections provide only high-level statements of success; they do not list the specific PDEs examined, the precise noise levels (e.g., SNR or variance values), dictionary cardinalities, or quantitative performance metrics (identification rate, coefficient error, residual norms). Without these data the central claim that S-IDENT succeeds at higher noise than existing strong-form methods cannot be evaluated.
  2. [§3] §3 (differentiation step): SURE is minimized on the smoothed field, yet the manuscript does not demonstrate that the selected window simultaneously controls bias in the first- and second-order derivatives that enter the nonlinear dictionary terms. A direct comparison of estimated versus analytic derivatives at the highest noise levels claimed would be required to rule out systematic error that could be absorbed by the subsequent trimming step.
  3. [§4] The trimming and reduction-in-residual selection procedure is described at a level that leaves open whether the thresholds or stopping criteria are fixed a priori or chosen after inspecting the target PDE. If the latter, the reported performance may not be reproducible on unseen data.
minor comments (2)
  1. Notation for the strong-form dictionary and the precise definition of the residual used in model selection should be written out explicitly (currently referenced only descriptively).
  2. A short table summarizing the PDEs, noise levels, and success metrics for each experiment would improve readability and allow direct comparison with the cited weak-form baselines.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments, which highlight opportunities to improve clarity, rigor, and reproducibility. We address each major comment below and will incorporate revisions accordingly.

read point-by-point responses

  1. Referee: [Abstract and results (presumably §5)] The abstract and results sections provide only high-level statements of success; they do not list the specific PDEs examined, the precise noise levels (e.g., SNR or variance values), dictionary cardinalities, or quantitative performance metrics (identification rate, coefficient error, residual norms). Without these data the central claim that S-IDENT succeeds at higher noise than existing strong-form methods cannot be evaluated.

    Authors: We agree that the abstract and the high-level summary in the results section would benefit from greater specificity. In the revised manuscript we will expand the abstract to name the PDEs tested, report noise levels in SNR or variance, state dictionary cardinalities, and include quantitative metrics such as identification rates, coefficient errors, and residual norms. A concise summary table will also be added to §5 to facilitate direct comparison with prior strong-form methods. revision: yes

  2. Referee: [§3] §3 (differentiation step): SURE is minimized on the smoothed field, yet the manuscript does not demonstrate that the selected window simultaneously controls bias in the first- and second-order derivatives that enter the nonlinear dictionary terms. A direct comparison of estimated versus analytic derivatives at the highest noise levels claimed would be required to rule out systematic error that could be absorbed by the subsequent trimming step.

    Authors: This observation is valid. While SURE is applied to the smoothed field, its simultaneous effect on first- and second-order derivative bias has not been explicitly verified. We will add a new figure and accompanying text in §3 that directly compares the SG-SURE estimated derivatives against analytic ground truth at the highest noise levels used in the experiments, thereby confirming that residual bias is not systematically absorbed by the trimming procedure. revision: yes

  3. Referee: [§4] The trimming and reduction-in-residual selection procedure is described at a level that leaves open whether the thresholds or stopping criteria are fixed a priori or chosen after inspecting the target PDE. If the latter, the reported performance may not be reproducible on unseen data.

    Authors: We clarify that all thresholds and stopping criteria are fixed a priori and independent of the target PDE; they were determined once via cross-validation on a separate set of synthetic trajectories and held constant thereafter. In the revised §4 we will explicitly list these fixed values together with the selection rationale, thereby ensuring full reproducibility on unseen data. revision: yes

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity in derivation or claims

full rationale

The paper proposes S-IDENT as a numerical procedure combining SG differentiation (with external SURE window selection), trimming, and residual-based model selection, then validates it via numerical experiments on synthetic data. No load-bearing step reduces by the paper's own equations to a quantity defined in terms of parameters fitted to the target PDE coefficients; SURE is an independent risk estimator, and the reported identification success is empirical rather than a self-referential derivation. No self-citation chains or ansatz smuggling appear in the provided text as central justifications.

Assumptions & free parameters 0 free parameters · 1 assumptions · 0 invented entities

The central claim rests on standard numerical analysis results about polynomial smoothing accuracy and on the statistical properties of SURE; no new free parameters, axioms beyond standard math, or invented entities are introduced in the abstract.

assumptions (1)
  • standard math Savitzky-Golay differentiation supplies a guaranteed order of accuracy for the computed derivatives
    Explicitly stated in the abstract as a property of the chosen differentiation approach.

how reviews work

0 comments
Cite this review

Pith. "Pith review of PDE Identification Using Noise Adaptive Differentiation in Strong Form (S-IDENT)." pith.science (2026). https://pith.science/paper/2VIZOIPH

@misc{pith2026260631776,
  author       = {Pith},
  title        = {Pith review of: PDE Identification Using Noise Adaptive Differentiation in Strong Form (S-IDENT)},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2VIZOIPH}},
  note         = {Machine review of arXiv:2606.31776}
}
read the original abstract

We explore identifying partial differential equations (PDEs) from noisy observations of single time-space trajectories. Recent developments show the benefits of identifying PDEs in their weak forms. We investigate the use of differential Strong-form dictionaries for PDE IDENTification (S-IDENT), which enables finding more general linear and nonlinear PDEs. Building on an extensive exploration of integral-type denoised differentiation approaches, we propose to use Savitzky--Golay (SG) differentiation with an adaptive window length chosen based on Stein's Unbiased Risk Estimate (SURE). This offers a guaranteed order of accuracy while producing estimators with minimal variance. The identification process is further refined and stabilized through trimming and reduction-in-residual model selection. Numerical evidence shows that S-IDENT can successfully identify nonlinear PDEs at higher levels of noise than existing strong-form methods, while also yielding results comparable to weak-form approaches. We further verify the effectiveness of S-IDENT through comparisons with various strategies to approximate differential features. We provide numerical evidence that general differential-form dictionaries are larger and more ill-conditioned than those used for weak-form identification, yet S-IDENT does not significantly suffer from this combinatorial increase in dictionary size.

Figures

Figures reproduced from arXiv: 2606.31776 by the authors.

Figure 1
Figure 1. Noise amplification due to numerical differentiation. Estimation of the (a) first-order deriva [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The proposed S-IDENT framework (Subsection [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. SURE-SG: Stein’s Unbiased Risk Estimate (SURE) to automatically choose adaptive window [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (14 more)
Figure 4
Figure 4. Figure 4: Weight difference of various methods in terms of differentiation by integration. In the form of [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Relative magnitude response (19) of the representative methods shown in [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Comparison between SG differentiation and Successive Denoised Differentiation (SDD) [ [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: Comparison: S-IDENT, Robust-IDENT [21], and SINDy-PDE [46]. For each noise level, TPR and PPV are presented in the first row for support identification, and Ein and Eout in the second row for coefficient recovery. Each panel shows box plots from 50 independent experime…
Figure 8
Figure 8. Figure 8: Comparison: S-IDENT, Robust-IDENT [21], and SINDy-PDE [46]. Mean exact recovery rate (E.R.) (34) from 50 independent trials is presented for viscous Burgers (44), KdV (45), and Allen– Cahn (47). S-IDENT shows the best results. to be tuned for them to achieve exact reco…
Figure 9
Figure 9. Figure 9: Comparison: S-IDENT with Type-S (S-IDENT (S)), S-IDENT with Type-W (S-IDENT [PITH_FULL_IMAGE:figures/full_fig_p022_9.png]
Figure 10
Figure 10. Figure 10: Comparison: S-IDENT with Type-S (S-IDENT (S)), S-IDENT with Type-W (S-IDENT [PITH_FULL_IMAGE:figures/full_fig_p023_10.png]
Figure 11
Figure 11. Figure 11: SG polynomial degree vs S-IDENT results: Identification performance with the Type-S [PITH_FULL_IMAGE:figures/full_fig_p024_11.png]
Figure 12
Figure 12. Figure 12: Variant study of S1–S5 in Subsection 4.5: Exact recovery rate (E.R.) (34) from 50 indepen￾dent trials for identifying the viscous Burgers (44), KdV (45), and Allen–Cahn (47) equations using the feature approximation methods S1–S5. The proposed S-IDENT shows the best o…
Figure 13
Figure 13. Figure 13: Variant study of Direct (Proposed), Repeated, and Adaptive in Subsection [PITH_FULL_IMAGE:figures/full_fig_p026_13.png]
Figure 14
Figure 14. Figure 14: Correlation among feature pairs in the Type-W and Type-S dictionaries induced by (a) [PITH_FULL_IMAGE:figures/full_fig_p032_14.png]
Figure 15
Figure 15. Figure 15: Singular values of the feature matrices corresponding to the Type-W and Type-S dictionaries [PITH_FULL_IMAGE:figures/full_fig_p032_15.png]
Figure 16
Figure 16. Figure 16: Clean trajectory data of (A) the Harry Dym equation ( [PITH_FULL_IMAGE:figures/full_fig_p034_16.png]
Figure 17
Figure 17. Figure 17: Type-W dictionary, SG polynomial order vs. S-IDENT results: Identification performance [PITH_FULL_IMAGE:figures/full_fig_p036_17.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

57 extracted references · 57 canonical work pages

  1. [1]

    Global existence and decay for a nonlinear parabolic equation.Nonlinear Analysis: Theory, Methods & Applications, 19(8):763–768, 1992

    Matania Ben-Artzi. Global existence and decay for a nonlinear parabolic equation.Nonlinear Analysis: Theory, Methods & Applications, 19(8):763–768, 1992

    work page 1992

  2. [2]

    Higher order nonlinear degenerate parabolic equations

    Francisco Bernis and Avner Friedman. Higher order nonlinear degenerate parabolic equations. Journal of differential equations, 83(1):179–206, 1990

    work page 1990

  3. [3]

    Numerical differentiation of 2d functions from noisy data.Computers & Mathematics with Applications, 45(1-3):309–327, 2003

    M Bozzini and M Rossini. Numerical differentiation of 2d functions from noisy data.Computers & Mathematics with Applications, 45(1-3):309–327, 2003

    work page 2003

  4. [4]

    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

    work page 2016

  5. [5]

    The least-squares property of the Lanczos derivative.Mathematics Magazine, 78(5):368–378, 2005

    Nathanial Burch, Paul E Fishback, and Russell Gordon. The least-squares property of the Lanczos derivative.Mathematics Magazine, 78(5):368–378, 2005

    work page 2005

  6. [6]

    Identification of differential equations by dynamics-guided weighted weak form with voting.arXiv preprint arXiv:2506.03899, 2025

    Jiahui Cheng, Sung Ha Kang, Haomin Zhou, and Wenjing Liao. Identification of differential equations by dynamics-guided weighted weak form with voting.arXiv preprint arXiv:2506.03899, 2025

    work page arXiv 2025

  7. [7]

    Robust locally weighted regression and smoothing scatterplots.Journal of the American statistical association, 74(368):829–836, 1979

    William S Cleveland. Robust locally weighted regression and smoothing scatterplots.Journal of the American statistical association, 74(368):829–836, 1979

    work page 1979

  8. [8]

    Physics-informed genetic programming for discovery of partial differential equations from scarce and noisy data.Journal of Computational Physics, 514:113261, 2024

    Benjamin G Cohen, Burcu Beykal, and George M Bollas. Physics-informed genetic programming for discovery of partial differential equations from scarce and noisy data.Journal of Computational Physics, 514:113261, 2024

    work page 2024

Show all 57 references
  1. [9]

    Stoch-IDENT: New method and mathematical analysis for identifying SPDEs from data.arXiv preprint arXiv:2508.19177, 2025

    Jianbo Cui and Roy Y He. Stoch-IDENT: New method and mathematical analysis for identifying SPDEs from data.arXiv preprint arXiv:2508.19177, 2025

    2025 arXiv

  2. [10]

    Subspace pursuit for compressive sensing signal reconstruction

    Wei Dai and Olgica Milenkovic. Subspace pursuit for compressive sensing signal reconstruction. IEEE transactions on Information Theory, 55(5):2230–2249, 2009

    2009

  3. [11]

    Equations of Korteweg– de Vries type, and simple Lie algebras

    Vladimir Gershonovich Drinfeld and Vladimir Vyacheslavovich Sokolov. Equations of Korteweg– de Vries type, and simple Lie algebras. InDoklady Akademii Nauk, volume 258, pages 11–16. Russian Academy of Sciences, 1981

    1981

  4. [12]

    SIAM, 2019

    Qiang Du.Nonlocal Modeling, Analysis, and Computation: Nonlocal Modeling, Analysis, and Computation. SIAM, 2019

    2019

  5. [13]

    Non-parametric estimation of a multivariate probability density.Theory of Probability & Its Applications, 14(1):153–158, 1969

    Vassiliy A Epanechnikov. Non-parametric estimation of a multivariate probability density.Theory of Probability & Its Applications, 14(1):153–158, 1969

    1969

  6. [14]

    Local linear regression smoothers and their minimax efficiencies.The annals of Statistics, pages 196–216, 1993

    Jianqing Fan. Local linear regression smoothers and their minimax efficiencies.The annals of Statistics, pages 196–216, 1993

    1993

  7. [15]

    Routledge, 2018

    Jianqing Fan.Local polynomial modelling and its applications: monographs on statistics and applied probability 66. Routledge, 2018

    2018

  8. [16]

    Local poly- nomial regression: Optimal kernels and asymptotic minimax efficiency.Annals of the Institute of Statistical Mathematics, 49(1):79–99, 1997

    Jianqing Fan, Theo Gasser, Ir` ene Gijbels, Michael Brockmann, and Joachim Engel. Local poly- nomial regression: Optimal kernels and asymptotic minimax efficiency.Annals of the Institute of Statistical Mathematics, 49(1):79–99, 1997

    1997

  9. [17]

    Generation of finite difference formulas on arbitrarily spaced grids.Mathematics of computation, 51(184):699–706, 1988

    Bengt Fornberg. Generation of finite difference formulas on arbitrarily spaced grids.Mathematics of computation, 51(184):699–706, 1988. 28

    1988

  10. [18]

    General least-squares smoothing and differentiation by the convolution (Savitzky- Golay) method.Analytical Chemistry, 62(6):570–573, 1990

    Peter A Gorry. General least-squares smoothing and differentiation by the convolution (Savitzky- Golay) method.Analytical Chemistry, 62(6):570–573, 1990

    1990

  11. [19]

    Lanczos’ generalized derivative.The American mathematical monthly, 105(4):320–326, 1998

    Charles W Groetsch. Lanczos’ generalized derivative.The American mathematical monthly, 105(4):320–326, 1998

    1998

  12. [20]

    Chapter 4 - recent advances in identi- fication of differential equations from noisy data: IDENT review

    Roy Yuchen He, Hao Liu, Wenjing Liao, and Sung Ha Kang. Chapter 4 - recent advances in identi- fication of differential equations from noisy data: IDENT review. In Andreas Hauptmann, Bangti Jin, and Carola-Bibiane Sch¨ onlieb, editors,Machine Learning Solutions for Inverse Pro...

    2025

  13. [21]

    Robust identification of differential equations by numerical techniques from a single set of noisy observation.SIAM Journal on Scientific Computing, 44(3):A1145–A1175, 2022

    Yuchen He, Sung-Ha Kang, Wenjing Liao, Hao Liu, and Yingjie Liu. Robust identification of differential equations by numerical techniques from a single set of noisy observation.SIAM Journal on Scientific Computing, 44(3):A1145–A1175, 2022

    2022

  14. [22]

    Group projected subspace pursuit for identification of variable coefficient differential equations (GP-IDENT).Journal of Computational Physics, 494:112526, 2023

    Yuchen He, Sung Ha Kang, Wenjing Liao, Hao Liu, and Yingjie Liu. Group projected subspace pursuit for identification of variable coefficient differential equations (GP-IDENT).Journal of Computational Physics, 494:112526, 2023

    2023

  15. [23]

    Asymptotic the- ory ofℓ 1-regularized PDE identification from a single noisy trajectory.SIAM/ASA Journal on Uncertainty Quantification, 10(3):1012–1036, 2022

    Yuchen He, Namjoon Suh, Xiaoming Huo, Sung Ha Kang, and Yajun Mei. Asymptotic the- ory ofℓ 1-regularized PDE identification from a single noisy trajectory.SIAM/ASA Journal on Uncertainty Quantification, 10(3):1012–1036, 2022

    2022

  16. [24]

    How much can one learn a partial differential equation from its solution?Foundations of Computational Mathematics, 24(5):1595–1641, 2024

    Yuchen He, Hongkai Zhao, and Yimin Zhong. How much can one learn a partial differential equation from its solution?Foundations of Computational Mathematics, 24(5):1595–1641, 2024

    2024

  17. [25]

    Derivation and implicit solution of the Harry Dym equation and its connections with the Korteweg-de Vries equation

    Willy Hereman, Partha P Banerjee, and Monish Ranjan Chatterjee. Derivation and implicit solution of the Harry Dym equation and its connections with the Korteweg-de Vries equation. Journal of Physics A: Mathematical and General, 22(3):241–255, 1989

    1989

  18. [26]

    Finite differences in forward and inverse imaging problems: Maxpol design.SIAM Journal on Imaging Sciences, 10(4):1963–1996, 2017

    Mahdi S Hosseini and Konstantinos N Plataniotis. Finite differences in forward and inverse imaging problems: Maxpol design.SIAM Journal on Imaging Sciences, 10(4):1963–1996, 2017

    1963

  19. [27]

    Derivative error bounds for Lagrange interpolation: An extension of Cauchy’s bound for the error of Lagrange interpolation.Journal of approximation theory, 67(2):164–173, 1991

    Gary W Howell. Derivative error bounds for Lagrange interpolation: An extension of Cauchy’s bound for the error of Lagrange interpolation.Journal of approximation theory, 67(2):164–173, 1991

    1991

  20. [28]

    IDENT: Identifying differential equations with numerical time evolution.Journal of Scientific Computing, 87(1), 2021

    Sung Ha Kang, Wenjing Liao, and Yingjie Liu. IDENT: Identifying differential equations with numerical time evolution.Journal of Scientific Computing, 87(1), 2021

    2021

  21. [29]

    Dynamic scaling of growing interfaces

    Mehran Kardar, Giorgio Parisi, and Yi-Cheng Zhang. Dynamic scaling of growing interfaces. Physical Review Letters, 56(9):889, 1986

    1986

  22. [30]

    Nonlinear wave equations

    Martin Kruskal. Nonlinear wave equations. InDynamical Systems, Theory and Applications: Battelle Seattle 1974 Rencontres, pages 310–354. Springer, 2005

    1974

  23. [31]

    Applied analysis prentice hall.New York, 1956

    C Lanczos. Applied analysis prentice hall.New York, 1956

    1956

  24. [32]

    Evaluation of noisy data.Journal of the Society for Industrial and Applied Mathematics, Series B: Numerical Analysis, 1(1):76–85, 1964

    Cornelius Lanczos. Evaluation of noisy data.Journal of the Society for Industrial and Applied Mathematics, Series B: Numerical Analysis, 1(1):76–85, 1964

    1964

  25. [33]

    Weighted essentially non-oscillatory schemes

    Xu-Dong Liu, Stanley Osher, and Tony Chan. Weighted essentially non-oscillatory schemes. Journal of computational physics, 115(1):200–212, 1994

    1994

  26. [34]

    Savitzky–Golay smoothing and differentiation filter for even number data.Signal processing, 85(7):1429–1434, 2005

    Jianwen Luo, Kui Ying, and Jing Bai. Savitzky–Golay smoothing and differentiation filter for even number data.Signal processing, 85(7):1429–1434, 2005

    2005

  27. [35]

    Properties of Savitzky–Golay digital differentia- tors.Digital Signal Processing, 15(2):122–136, 2005

    Jianwen Luo, Kui Ying, Ping He, and Jing Bai. Properties of Savitzky–Golay digital differentia- tors.Digital Signal Processing, 15(2):122–136, 2005

    2005

  28. [36]

    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. 29

    2021

  29. [37]

    A new regularization approach for numerical differentiation.Inverse Problems in Science and Engineering, 28(12):1747–1772, 2020

    Abinash Nayak. A new regularization approach for numerical differentiation.Inverse Problems in Science and Engineering, 28(12):1747–1772, 2020

    2020

  30. [38]

    Long-scale evolution of thin liquid films.Reviews of modern physics, 69(3):931, 1997

    Alexander Oron, Stephen H Davis, and S George Bankoff. Long-scale evolution of thin liquid films.Reviews of modern physics, 69(3):931, 1997

    1997

  31. [39]

    Chebyshev approximation for nonrecursive digital filters with linear phase.IEEE Transactions on circuit theory, 19(2):189–194, 1972

    T Parks and James McClellan. Chebyshev approximation for nonrecursive digital filters with linear phase.IEEE Transactions on circuit theory, 19(2):189–194, 1972

    1972

  32. [40]

    Wiley-Interscience, 1987

    Thomas W Parks and C Sidney Burrus.Digital filter design. Wiley-Interscience, 1987

    1987

  33. [41]

    Yogeshwari F Patel and Mohammad Izadi. Analytical study of the fractional-order Drinfeld– Sokolov–Wilson system for nonlinear wave dynamics in dispersive media with modulation insta- bility analysis.Arabian Journal of Mathematics, pages 1–18, 2025

    2025

  34. [42]

    Smoothing by Savitzky-Golay and Legendre filters

    Per-Olof Persson and Gilbert Strang. Smoothing by Savitzky-Golay and Legendre filters. In Mathematical systems theory in biology, communications, computation, and finance, pages 301–

  35. [43]

    Homogeneous differentiator design using implicit Lyapunov function method

    Andrey Polyakov, Denis Efimov, and Wilfrid Perruquetti. Homogeneous differentiator design using implicit Lyapunov function method. In2014 European Control Conference (ECC), pages 288–293. IEEE, 2014

    2014

  36. [44]

    Rangarajan and Sudarshan P

    S.K. Rangarajan and Sudarshan P. Purushothaman. Lanczos’ generalized derivative for higher orders.Journal of computational and applied mathematics, 177(2):461–465, 2005

    2005

  37. [45]

    Wiley New York, 1973

    Calyampudi Radhakrishna Rao.Linear statistical inference and its applications, volume 2. Wiley New York, 1973

    1973

  38. [46]

    Data-driven discovery of partial differential equations.Science advances, 3(4):e1602614, 2017

    Samuel H Rudy, Steven L Brunton, Joshua L Proctor, and J Nathan Kutz. Data-driven discovery of partial differential equations.Science advances, 3(4):e1602614, 2017

    2017

  39. [47]

    Multivariate locally weighted least squares regression.The annals of statistics, pages 1346–1370, 1994

    David Ruppert and Matthew P Wand. Multivariate locally weighted least squares regression.The annals of statistics, pages 1346–1370, 1994

    1994

  40. [48]

    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

  41. [49]

    Maximally flat low-pass FIR filters with reduced delay

    Ivan W Selesnick and C Sidney Burrus. Maximally flat low-pass FIR filters with reduced delay. IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing, 45(1):53–68, 1998

    1998

  42. [50]

    Estimation of the mean of a multivariate normal distribution.The annals of Statistics, pages 1135–1151, 1981

    Charles M Stein. Estimation of the mean of a multivariate normal distribution.The annals of Statistics, pages 1135–1151, 1981

    1981

  43. [51]

    WG-IDENT: Weak group identification of pdes with varying coefficients.Journal of Computational Physics, page 114454, 2025

    Cheng Tang, Roy Y He, and Hao Liu. WG-IDENT: Weak group identification of pdes with varying coefficients.Journal of Computational Physics, page 114454, 2025

    2025

  44. [52]

    PriorIDENT: Prior-informed PDE identification from noisy data.arXiv preprint arXiv:2603.05946, 2026

    Cheng Tang, Hao Liu, and Dong Wang. PriorIDENT: Prior-informed PDE identification from noisy data.arXiv preprint arXiv:2603.05946, 2026

    2026

  45. [53]

    WeakIdent: Weak formulation for identifying differential equation using narrow-fit and trimming.Journal of Computational Physics, 483:112069, 2023

    Mengyi Tang, Wenjing Liao, Rachel Kuske, and Sung Ha Kang. WeakIdent: Weak formulation for identifying differential equation using narrow-fit and trimming.Journal of Computational Physics, 483:112069, 2023

    2023

  46. [54]

    Fourier features for identifying differ- ential equations (FourierIdent).Journal of Scientific Computing, 106(47), 2026

    Mengyi Tang, Hao Liu, Wenjing Liao, and Sung Ha Kang. Fourier features for identifying differ- ential equations (FourierIdent).Journal of Scientific Computing, 106(47), 2026

    2026

  47. [55]

    Regression shrinkage and selection via the lasso.Journal of the Royal Statistical Society Series B: Statistical Methodology, 58(1):267–288, 1996

    Robert Tibshirani. Regression shrinkage and selection via the lasso.Journal of the Royal Statistical Society Series B: Statistical Methodology, 58(1):267–288, 1996

    1996

  48. [56]

    B-spline signal processing

    Michael Unser, Akram Aldroubi, and Murray Eden. B-spline signal processing. ii. efficiency design and applications.IEEE transactions on signal processing, 41(2):834–848, 2002

    2002

  49. [57]

    Piecewise polynomial, positive definite and compactly supported radial func- tions of minimal degree.Advances in computational Mathematics, 4(1):389–396, 1995

    Holger Wendland. Piecewise polynomial, positive definite and compactly supported radial func- tions of minimal degree.Advances in computational Mathematics, 4(1):389–396, 1995. 30 Algorithm 1Proposed identification method: S-IDENT Input :Noisy samples{˜u i}N i=1 on a space–tim...

    1995

Pith tools

Reviewed July 1, 2026 · model on record in the stance chip above.