Pith. sign in

REVIEW 4 major objections 4 minor 55 references

chebgreen: Learning and Interpolating Continuous Empirical Green's Functions from Data

T0 review · 4 major / 4 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read This paper establishes that learned continuous Green's functions for unknown one-dimensional linear operators can be interpolated to unseen control-parameter values with test errors under about 3 percent, including for non-self-adjoint…

desk verdict Genuine continuous extension of EGF interpolation with public code and a real error analysis, but the unseen-parameter claim needs an inconsistency fixed and a check on the injectivity radius before it fully lands. read the letter →

arxiv 2501.18715 v5 pith:ZTAPL7SY submitted 2025-01-30 cs.LG cs.NAmath.NA

Harshwardhan Praveen , Jacob Brown , Christopher Earls This is my paper · ORCID
classification cs.LGcs.NAmath.NA MSC 65D0565F2565R2068T07
keywords EmpiricalGreen'sfunctionoperatorlearningsingularvalueexpansionChebyshevpolynomialsquasimatricesStiefelmanifoldinterpolationrationalneuralnetworksparametriclinearPDEs
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 seeks to show that the solution operator of an unknown one-dimensional linear system, parameterized by a scalar control variable, can be learned from force-response data and then predicted at parameter values never seen in training. Its route is to learn a continuous empirical Green's function, the impulse response of the hidden operator, as a rational neural network, rewrite it as a singular value expansion in a Chebyshev basis, and interpolate the left and right singular functions through the tangent space of an infinite-dimensional Stiefel manifold. On the paper's benchmarks the interpolated Green's functions reach test errors of roughly 0.3 to 3 percent, the method handles non-self-adjoint operators such as advection-diffusion, and the data budget of about 100 input-output pairs is an order of magnitude smaller than in the earlier discrete approach. If the claim holds, unknown physical systems described by parametric linear PDEs can be modeled mesh-independently from modest, noisy measurements.

What carries the argument

The load-bearing object is the singular value expansion of the learned Green's function, $G_\theta(x,s) = U(s;\theta)\Sigma(\theta)V^*(x;\theta)$, where $U$ and $V$ are quasimatrices, meaning matrices whose columns are $L^2$ functions represented as Chebyshev interpolants, built by a two-dimensional Chebyshev factorization. Around this object the algorithm constructs an infinite-dimensional analogue of the Stiefel manifold, the space of quasimatrices with orthonormal columns. The essential mechanism is tangent-space interpolation: project each model's singular-function quasimatrix onto the tangent space at a reference parameter, interpolate those tangent points and the singular values with Lagrange polynomials, and retract via quasimatrix QR factorization. The paper supplies the projection formula and proves that it is orthogonal, and its error analysis identifies the QR retraction as the dominant error source.

What would settle it

Learn Green's functions for $d^2u/dx^2 + \theta^2 u = f$ at $\theta=4.9$ and $\theta=5.1$, where the paper shows the first two singular functions swap order, interpolate to $\theta^*=5.0$ with the paper's Algorithm 2, and compare against a directly learned target at 5.0; a test error far above the roughly 3 percent seen in the paper's Airy experiments would show that the sign- and order-corrected tangent-space interpolation does not track the crossing.

Watch

Extended reading notes

Core claim

The central discovery claimed is that a continuous empirical Green's function comes with a manifold-interpolation rule that lets it travel in parameter space: the singular functions of the learned Green's function for each sampled parameter $\theta_j$ are treated as points on the Stiefel manifold of quasimatrices, lifted to the tangent space at a reference parameter $\theta_0$ by the projection $\Psi \mapsto \Psi - \Phi\,\mathrm{sym}(\Phi^*\Psi)$, interpolated with Lagrange polynomials, and mapped back by quasimatrix QR factorization, while the singular values are interpolated separately. The paper proves that this projection is the orthogonal projection onto the tangent space, gives error bounds of cubic order for the projection and quadratic order for the QR retraction, and handles mode swaps and sign flips by matching inner products against the reference basis. In the paper's account the result is high-fidelity interpolation for the Poisson, advection-diffusion, Airy, and fractional-Laplacian operators, including extrapolation to $\theta=9$ from models at 6, 7, and 8, with the continuous Chebyshev representation making the pipeline independent of any particular spatial mesh.

Load-bearing premise

The argument assumes that the left and right singular functions of the Green's function vary smoothly enough across the sampled parameter values for Lagrange interpolation inside a single tangent space to represent the unseen parameter, and the paper's error analysis establishes this only asymptotically as the parameter gap goes to zero rather than at the tested gaps such as Airy $\theta$ from 1 to 10.

Editorial extensions

If this is right

  • For one-dimensional linear operators with a scalar parameter, the pipeline yields a continuously queryable Green's function from about 100 force-response pairs without writing down the governing equation.
  • Interpolated test errors are 0.54 percent for advection-diffusion at $\theta=2.5$, 2.76 percent for Airy at $\theta=7$, and 0.99 percent for the fractional Laplacian at $\theta=0.85$; extrapolation to Airy $\theta=9$ gives 2.28 percent.
  • The method extends to non-self-adjoint and nonlocal operators, going beyond the self-adjoint case treated in the prior discrete empirical-Green's-function work.
  • Training with 50 percent output noise leaves interpolated test errors below about 3.3 percent on the three benchmarks, and the advection-diffusion test error stays below 5 percent even at 100 percent noise.
  • Because the representation lives in a Chebyshev basis rather than on a fixed sensor grid, the stored Green's function can be evaluated at arbitrary locations and kept to high precision.

Reading between the lines

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

  • Stress-test extension: interpolate across the singular-value crossing that the paper identifies for the Helmholtz operator at $\theta_{\mathrm{crit}}=\sqrt{5/2}\,\pi$; if the test error spikes at the crossing, the mode-matching assumption, rather than the manifold geometry, is the bottleneck.
  • Extension to real sensor data: because the singular value expansion is continuous, the pipeline should in principle accept force-response measurements taken at different sensor locations for different parameter values without first resampling to a common grid; the paper allows arbitrary sensor locations but does not demonstrate this cross-parameter mismatch.
  • Higher-dimensional extension: the same tangent-space interpolation would have to be retried on tensor decompositions such as chebfun3/Tucker or tensor-train forms, and the Eckart-Young-Mirsky obstruction the paper flags suggests a different low-rank manifold will be needed there.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The manuscript presents chebgreen, a Python library that learns continuous approximations of Green's functions of one-dimensional linear PDEs from input-output data and interpolates them to new values of a control parameter. The pipeline trains rational neural networks to represent the Green's function and homogeneous solution, converts the learned bivariate function into a Chebyshev-based singular value expansion using a new Python implementation of chebfun, and interpolates the resulting quasimatrices on the tangent space of an infinite-dimensional Stiefel manifold. Numerical experiments cover Poisson, advection-diffusion, Airy, and fractional Laplacian problems, with and without noise; the reported test errors for interpolated models are below about 3%.

Significance. If the results hold, the paper offers a meaningful step beyond the discrete empirical Green's function interpolation of Praveen et al. [32]: the representation is continuous and mesh-independent, it applies to non-self-adjoint operators, and the code is publicly available. The paper also provides a rare external check: for the advection-diffusion case the interpolated Green's function is compared to an analytical expression with relative error 0.26%. The theoretical error analysis (cubic projection error, quadratic retraction error) is elegant but only local; the main support for the central claim is the empirical test-error evidence, which is reasonably strong for the Airy and advection-diffusion experiments.

major comments (4)
  1. [Section 3.3, Lemma 1] The combined error analysis is asymptotic and local, and the paper does not verify the conditions under which the bounds hold at the parameter spacings used in the experiments. In particular, Lemma 1 guarantees the exponential map is a local diffeomorphism only inside a geodesic ball of radius π, but the manuscript never checks that the learned singular quasimatrices at θ1=1 and θ3=10 in the Airy example lie within that ball around the reference θ0=5. Please either add a numerical check (e.g., compute the norm of the lifted tangent vectors ΓU_j and ΓV_j, or approximate the geodesic distances, and compare with π) or explicitly state that the theoretical bounds are asymptotic and do not cover the finite parameter gaps used in the experiments.
  2. [Section 4.4] The fractional Laplacian experiment is internally inconsistent: the text states 'interpolating to θ* = 0.9', but θ2 = 0.9 is one of the three training nodes, while the Fig. 16 caption reports 'Interpolated Green's function at θ* = 0.85'. Since the central claim concerns interpolation to unseen parameters, this example must be corrected and re-run; if the intended target is 0.85, the text and Table 1 should be updated consistently.
  3. [Section 4.2] The advection-diffusion experiment reports Ntrain = 237 input-output pairs, which conflicts with the default Nsamples = 100 (95% training) described in Section 4 and with the sample-size study in Section 4.1.1 that motivates using 100 samples. Please clarify the data budget for this experiment and reconcile the 'order of magnitude lower data' claim with the actual numbers.
  4. [Section 3.3] The theoretical error analysis covers only the interpolation steps for a given singular value expansion; it does not include the neural network approximation error or the Chebyshev truncation error. The statement that the quadratic retraction error 'dominates the error of the other two steps' is also not a complete error bound for the full pipeline. Please state clearly which sources of error are covered by the theory and which are only assessed empirically.
minor comments (4)
  1. [Equation (19)] The periodic boundary condition is written as u(−π/2) = u(−π/2) on both sides; it should presumably be u(−π/2) = u(π/2).
  2. [Algorithm 2, Step 3.2] The notation 'using {[ΓU_j, ΓU_j]}' appears to have a typo; the second entry should likely be ΓV_j.
  3. [Section 2.7] There is a typo: 'consiedered' should be 'considered'.
  4. [Section 4.1.1] The x-axis of Fig. 10A is labeled 'Noise level' but the study in Section 4.1.1 is about sample count; please ensure the figure labels match the discussion.

Circularity Check

1 steps flagged · score 3.0 of 10

No load-bearing circularity: the interpolation claim is tested against separately trained target models and one analytical Green's function; the only construction-level problem is the fractional-Laplacian demo, where the text interpolates to a training node (θ*=0.9=θ2), making that 'interpolated' model equal the trained model by construction, while the Fig. 16 caption reports θ*=0.85.

  1. self definitional [Section 4.4 (Fractional Laplacian), text vs. Fig. 16 caption; Algorithm 2 Step 3.2]
    "We follow the same procedure of approximating three Green’s functions at parameter values θ1 = 0.8, θ2 = 0.9, and θ3 = 0.95 and interpolating to θ∗ = 0.9. ... (Fig. 16 caption: 'Interpolated Green’s function at θ∗ = 0.85 ϵtest = 0.99%')"

    If the text is taken literally, the claimed 'unseen' target θ*=0.9 is exactly the training node θ2=0.9. In Algorithm 2 the base point is the closest node (θ0=θ2), the tangent-space lift of the base point is Γ=U(θ0)−U(θ0)sym(U*(θ0)U(θ0))=0, and Lagrange interpolation evaluated at a node returns that node's value by construction. Hence the 'interpolated' U(0.9), Σ(0.9), V(0.9) coincide with the trained model at θ2, so the reported ϵtest=0.99% merely re-measures the θ2 training model (Table 1 lists 0.95% for θ2) and no unseen parameter is tested. The Fig. 16 caption contradicts the text, so as written one of the three demonstrations reduces to a training input rather than a prediction.

full rationale

The central derivation is self-contained and not circular. The Green's function is learned from forcing–response pairs with rational neural networks [31] and validated externally: the Laplacian model against the analytical Green's function (ϵ=0.36%) and the interpolated advection–diffusion model against the closed-form solution (ϵ=0.26%). The manifold interpolation of Algorithm 2 is tested at genuinely unseen parameters in two demonstrations — Airy θ*=7 with interpolants at θ=1,5,10 and advection–diffusion θ*=2.5 with interpolants at θ=1,2,3 — where the target models are trained on data at θ* that are never used by the interpolation, so those checks are independent and the reported errors (2.76% and 0.54%) are not forced by construction. Citations to the authors' own prior work are routine and not load-bearing: [32] is a published, externally validated method that this paper extends with new proofs (Theorem 1, Propositions 1–2, Lemmas 1–2), and the external inputs to those proofs (Zimmermann–Stoye, Harms–Mennucci, Altmann et al.) are independent literature, not author-derived uniqueness theorems. The paper even candidly admits the gap in its theoretical support — 'no attention has been given to quantifying the size of the region about a base point by which these methods are valid' (Section 3) — which is an honest correctness risk (the injectivity radius π of Lemma 1 is never checked against the finite parameter separations used), but it is not circularity, since the paper presents asymptotic bounds rather than claiming a finite-gap theorem. The single construction-level defect is the fractional-Laplacian demonstration: the text says θ*=0.9, a training node where Lagrange interpolation returns the node's trained SVE by definition, while the figure caption says θ*=0.85; the paper must correct this inconsistency before that demonstration can count as interpolation to an unseen parameter. Because two of the three demonstrations plus the analytical comparison remain independent, the central claim does not reduce to its inputs; the score of 3 reflects one internally inconsistent demonstration and the acknowledged, non-circular limitation in the radius quantification, not load-bearing self-citation.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests on standard functional analysis results and on the reliability of two pre-existing tools (rational neural networks and Chebyshev approximation) plus a geometric interpolation assumption. No new physical entities are introduced.

free parameters (4)
  • GP length scale sigma = 0.01
    Hand-chosen in Section 2.1 to ensure forcing functions are resolved and the ensemble has full numerical rank; not estimated from data.
  • Number of training samples Ntrain = 100 (default), 237 (advection-diffusion)
    Set based on a plateau study in Section 4.1.1; the advection-diffusion value is not explained.
  • Chebyshev tolerance epsilon_cheb = 2.22e-16 (most), 1e-9 (fractional Laplacian)
    Set in Section 2.6 to control the truncation rank K; chosen per problem, not fitted.
  • Neural network architecture = 4 layers, 50 neurons each
    Chosen by hand; no sensitivity study provided.
assumptions (5)
  • standard math Hilbert-Schmidt theorem ensures a square-integrable Green's function admits an SVE
    Invoked in Section 2.2 to justify the SVE representation.
  • standard math Eckart-Young-Mirsky theorem gives optimality of the truncated SVE
    Used in Section 2.2 to argue the truncated representation is optimal.
  • domain assumption Quasimatrix QR factorization is a retraction on the infinite-dimensional Stiefel manifold
    Used in Section 2.7 as the exponential map approximation; cited from Altmann et al. [43].
  • domain assumption Rational NN approximation property from Boulle et al. [31]
    The method relies on rational NNs being able to approximate the Green's function from data.
  • domain assumption The forcing ensemble from a GP with squared-exponential kernel adequately spans the operator's range
    Used in Section 2.1; motivated by theoretical results [26].

how reviews work

0 comments
Cite this review

Pith. "Pith review of chebgreen: Learning and Interpolating Continuous Empirical Green's Functions from Data." pith.science (2026). https://pith.science/paper/ZTAPL7SY

@misc{pith2026250118715,
  author       = {Pith},
  title        = {Pith review of: chebgreen: Learning and Interpolating Continuous Empirical Green's Functions from Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZTAPL7SY}},
  note         = {Machine review of arXiv:2501.18715}
}
read the original abstract

In this work, we present a mesh-independent, data-driven library, chebgreen, to mathematically model one-dimensional systems, possessing an associated control parameter, and whose governing partial differential equation is unknown. The proposed method learns an Empirical Green's Function for the associated, but hidden, boundary value problem, in the form of a Rational Neural Network from which we subsequently construct a bivariate representation in a Chebyshev basis. We uncover the Green's function, at an unseen control parameter value, by interpolating the left and right singular functions within a suitable library, expressed as points on a manifold of Quasimatrices, while the associated singular values are interpolated with Lagrange polynomials.

Figures

Figures reproduced from arXiv: 2501.18715 by the authors.

Figure 1
Figure 1. Schematic of the method for approximating Green’s functions with [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Example of the approximate distance function for a domain [ [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Entries in the column, represented by dots, of a matrix (left) can be replaced by a [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (13 more)
Figure 4
Figure 4. Figure 4: Example of pivot locations on a grid The pivots are chosen at the locations where the absolute error between the current approximant and the bivariate function G(x, y) is maximum. Repeating this procedure iteratively for K steps, chebfun2 constructs an approximation fo…
Figure 5
Figure 5. Figure 5: Householder triangularization of a Quasimatrix [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: The points on the “Stiefel” manifold (represented by the orthonormal quasimatrices [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]
Figure 7
Figure 7. Figure 7: (A) Magnitude of the first two singular values of the Green’s function associated with the [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 8
Figure 8. Figure 8: (A) Approximation of the Green’s functions learned by [PITH_FULL_IMAGE:figures/full_fig_p023_8.png]
Figure 9
Figure 9. Figure 9: (A) Relative error for the Laplacian operator when trained with different number of samples, [PITH_FULL_IMAGE:figures/full_fig_p023_9.png]
Figure 10
Figure 10. Figure 10: (A) Relative error at different noise level, [PITH_FULL_IMAGE:figures/full_fig_p024_10.png]
Figure 11
Figure 11. Figure 11: (A-C) Approximation of the Green’s functions associated with the Advection-Diffusion prob [PITH_FULL_IMAGE:figures/full_fig_p025_11.png]
Figure 12
Figure 12. Figure 12: Relative error at different noise level, [PITH_FULL_IMAGE:figures/full_fig_p026_12.png]
Figure 13
Figure 13. Figure 13: (A) First and (B) Second learned Left Singular Function for the airy problem at [PITH_FULL_IMAGE:figures/full_fig_p027_13.png]
Figure 14
Figure 14. Figure 14: (A-C) Approximation of the Green’s functions associated with the Airy problem at [PITH_FULL_IMAGE:figures/full_fig_p028_14.png]
Figure 15
Figure 15. Figure 15: (A) Extrapolated Green’s function (ϵtest = 2.28%). (B) The target Green’s function approx￾imated by chebgreen at θ∗ = 9. 28 [PITH_FULL_IMAGE:figures/full_fig_p028_15.png]
Figure 16
Figure 16. Figure 16: (A-C) Approximation of the Green’s functions associated with the Fractional Laplacian [PITH_FULL_IMAGE:figures/full_fig_p029_16.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 42 canonical work pages

  1. [32]

    Praveen, N

    H. Praveen, N. Boull´ e, C. Earls, Principled interpolation of green’s functions learned from data, Computer Methods in Applied Mechanics and Engineering 409 (2023) 115971. doi:10.1016/J.CMA.2023.115971

    work page arXiv 2023

  2. [1]

    R. P. Feynman, The Character of Physical Laws, M.I.T. Press, 1967

    work page 1967

  3. [2]

    Raissi, Deep hidden physics models: Deep learning of nonlinear partial differential equations, J

    M. Raissi, Deep hidden physics models: Deep learning of nonlinear partial differential equations, J. Mach. Learn. Res. 19 (1) (2018) 932–955

    work page 2018

  4. [3]

    Raissi, G

    M. Raissi, G. E. Karniadakis, Hidden physics models: Machine learning of nonlinear partial differ- ential equations, J. Comput. Phys. 357 (2018) 125–141

    work page 2018

  5. [4]

    J. Berg, K. Nystr¨ om, Neural network augmented inverse problems for PDEs, arXiv preprint arXiv:1712.09685 (2017)

    arXiv 2017

  6. [5]

    L. Lu, X. Meng, Z. Mao, G. E. Karniadakis, Deepxde: A deep learning library for solving differential equations, SIAM Review 63 (2021) 208–228

    work page 2021

  7. [6]

    Sirignano, K

    J. Sirignano, K. Spiliopoulos, Dgm: A deep learning algorithm for solving partial differential equa- tions, Journal of Computational Physics 375 (2018) 1339–1364. doi:10.1016/J.JCP.2018.08.029

    doi:10.1016/j.jcp.2018.08.029 2018

  8. [7]

    Weinan, B

    E. Weinan, B. Yu, The deep ritz method: A deep learning-based numerical algorithm for solving variational problems, Communications in Mathematics and Statistics 6 (2018) 1–12

    work page 2018

Show all 55 references
  1. [8]

    Raissi, P

    M. Raissi, P. Perdikaris, G. E. Karniadakis, Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equa- tions, J. Comput. Phys. 378 (2019) 686–707

    2019

  2. [9]

    Schaeffer, Learning partial differential equations via data discovery and sparse optimization, Proc

    H. Schaeffer, Learning partial differential equations via data discovery and sparse optimization, Proc. R. Soc. A 473 (2197) (2017) 20160446

    2017

  3. [10]

    S. L. Brunton, J. L. Proctor, J. N. Kutz, Discovering governing equations from data by sparse identification of nonlinear dynamical systems, Proc. Natl. Acad. Sci. U.S.A. 113 (15) (2016) 3932– 3937

    2016

  4. [11]

    Champion, B

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

    2019

  5. [12]

    S. H. Rudy, S. L. Brunton, J. L. Proctor, J. N. Kutz, Data-driven discovery of partial differential equations, Sci. Adv. 3 (4) (2017) e1602614

    2017

  6. [13]

    Bonneville, C

    C. Bonneville, C. Earls, Bayesian deep learning for partial differential equation parameter discovery with sparse and noisy data, J. Comput. Phys.: X 16 (2022) 100115

    2022

  7. [14]

    Stephany, C

    R. Stephany, C. Earls, PDE-READ: Human-readable partial differential equation discovery using deep learning, Neural Netw. 154 (2022) 360–382

    2022

  8. [15]

    Stephany, C

    R. Stephany, C. Earls, Pde-learn: Using deep learning to discover partial differential equations from noisy, limited data, Neural Networks 174 (2024) 106242

    2024

  9. [16]

    Stephany, C

    R. Stephany, C. Earls, Weak-pde-learn: A weak form based approach to discovering pdes from noisy, limited data, Journal of Computational Physics 506 (6 2024)

    2024

  10. [17]

    Boull´ e, A

    N. Boull´ e, A. Townsend, A mathematical guide to operator learning, Handbook of Numerical Anal- ysis 25 (2024) 83–125. doi:10.1016/BS.HNA.2024.05.003

    2024 doi

  11. [18]

    L. Lu, P. Jin, G. Pang, Z. Zhang, G. E. Karniadakis, Learning nonlinear operators via deeponet based on the universal approximation theorem of operators, Nature Machine Intelligence 3 (2021) 218–229

    2021

  12. [19]

    Z. Li, N. Kovachki, K. Azizzadenesheli, B. Liu, K. Bhattacharya, A. Stuart, A. Anandkumar, Fourier neural operator for parametric partial differential equations, Advances in Neural Information Processing Systems 33 (2020) 6755–6766

    2020

  13. [20]

    Z. Li, N. Kovachki, K. Azizzadenesheli, B. Liu, K. Bhattacharya, A. Stuart, A. Anandkumar, Neural operator: Graph kernel network for partial differential equations, arXiv preprint arXiv:2003.03485 (3 2020)

    2003 arXiv

  14. [21]

    Z. Li, N. Kovachki, K. Azizzadenesheli, B. Liu, K. Bhattacharya, A. Stuart, A. Anandkumar, Multipole graph neural operator for parametric partial differential equations, Proceedings of the 34th International Conference on Neural Information Processing Systems (2020)

    2020

  15. [22]

    Z. Li, N. Kovachki, C. Choy, Boyi, J. Kossaifi, S. P. Otta, M. A. Nabian, M. Stadler, C. Hundt, , K. Azizzadenesheli, A. Stuart, A. Anandkumar, Geometry-informed neural operator for large- scale 3d pdes, Proceedings of the 37th International Conference on Neural Information Pr...

    2023

  16. [23]

    Evans, Partial differential equations, 2nd Edition, American Mathematical Society, 2010

    L. Evans, Partial differential equations, 2nd Edition, American Mathematical Society, 2010

    2010

  17. [24]

    Khanfer, Applied functional analysis, Applied Functional Analysis (2024) 1–367doi:10.1007/978- 1-4612-0821-1/COVER

    A. Khanfer, Applied functional analysis, Applied Functional Analysis (2024) 1–367doi:10.1007/978- 1-4612-0821-1/COVER

    2024 doi

  18. [25]

    Fasshauer, M

    G. Fasshauer, M. McCourt, Kernel-based Approximation Methods using MATLAB, Vol. 19, WORLD SCIENTIFIC, 2015. doi:10.1142/9335. 31

    2015 doi

  19. [26]

    Boull´ e, A

    N. Boull´ e, A. Townsend, Learning elliptic partial differential equations with randomized linear algebra, Found. Comput. Math. (2022) 1–31

    2022

  20. [27]

    Boulle, A

    N. Boulle, A. Townsend, A generalization of the randomized singular value decomposition, in: International Conference on Learning Representations, 2022

    2022

  21. [28]

    Boull´ e, S

    N. Boull´ e, S. Kim, T. Shi, A. Townsend, Learning green’s functions associated with time- dependent partial differential equations, The Journal of Machine Learning Research 23 (2022) 1–34. doi:10.5555/3586589.3586807

    2022

  22. [29]

    C. R. Gin, D. E. Shea, S. L. Brunton, J. N. Kutz, DeepGreen: Deep learning of Green’s functions for nonlinear boundary value problems, Sci. Rep. 11 (1) (2021) 1–14

    2021

  23. [30]

    Boulle, Y

    N. Boulle, Y. Nakatsukasa, A. Townsend, Rational neural networks, Advances in Neural Information Processing Systems 33 (2020) 14243–14253

    2020

  24. [31]

    Boull´ e, C

    N. Boull´ e, C. J. Earls, A. Townsend, Data-driven discovery of Green’s functions with human- understandable deep learning, Sci. Rep 12 (1) (2022) 1–9

    2022

  25. [33]

    Amsallem, C

    D. Amsallem, C. Farhat, Interpolation method for adapting reduced-order models and application to aeroelasticity, AIAA Journal 46 (2008) 1803–1813. doi:10.2514/1.35374. URL https://arc.aiaa.org/doi/abs/10.2514/1.35374

    2008 doi

  26. [34]

    Townsend, L

    A. Townsend, L. N. Trefethen, An extension of chebfun to two dimensions, SIAM Journal on Scientific Computing 35 (12 2013). doi:10.1137/130908002

    2013 doi

  27. [35]

    T. A. Driscoll, N. Hale, L. N. Trefethen, Chebfun Guide, Pafnuty Publications, 2014

    2014

  28. [36]

    Z. Li, N. B. Kovachki, K. Azizzadenesheli, B. liu, K. Bhattacharya, A. Stuart, A. Anandkumar, Fourier neural operator for parametric partial differential equations, in: International Conference on Learning Representations, 2021

    2021

  29. [37]

    G. W. Stewart, J. guang. Sun, Matrix perturbation theory, Academic Press, 1990

    1990

  30. [38]

    Sukumar, A

    N. Sukumar, A. Srivastava, Exact imposition of boundary conditions with distance functions in physics-informed deep neural networks, Computer Methods in Applied Mechanics and Engineering 389 (2022) 114333. doi:10.1016/J.CMA.2021.114333

    2022

  31. [39]

    V. L. Rvachev, T. I. Sheiko, V. Shapiro, I. Tsukanov, Transfinite interpolation over implicitly defined sets, Computer Aided Geometric Design 18 (2001) 195–220. doi:10.1016/S0167-8396(01)00015-2

    2001 doi

  32. [40]

    Biswas, V

    A. Biswas, V. Shapiro, Approximate distance fields with non-vanishing gradients, Graphical Models 66 (2004) 133–159. doi:10.1016/J.GMOD.2004.01.003

    2004 doi

  33. [41]

    S¨ uli, D

    E. S¨ uli, D. F. Mayers, An introduction to numerical analysis, An Introduction to Numerical Analysis (8 2003). doi:10.1017/CBO9780511801181

    2003 doi

  34. [42]

    Richardson, Chebpy - a python implementation of chebfun

    M. Richardson, Chebpy - a python implementation of chebfun. URL https://github.com/chebpy/chebpy

  35. [43]

    Altmann, D

    R. Altmann, D. Peterseim, T. Stykel, Energy-adaptive riemannian optimization on the stiefel man- ifold (2022). arXiv:2108.09831

    2022 arXiv

  36. [44]

    L. N. Trefethen, Householder triangularization of a quasimatrix, IMA Journal of Numerical Analysis 30 (2010) 887–897. doi:10.1093/IMANUM/DRP018

    2010 doi

  37. [45]

    Absil, R

    P.-A. Absil, R. Mahony, R. Sepulchre, Optimization Algorithms on Matrix Manifolds, Princeton University Press, Princeton, NJ, 2008

    2008

  38. [46]

    Sternfels, C

    R. Sternfels, C. J. Earls, Reduced-order model tracking and interpolation to solve pde-based bayesian inverse problems, Inverse Problems 29 (7) (2013) 075014

    2013

  39. [47]

    Harms, A

    P. Harms, A. C. Mennucci, Geodesics in infinite dimensional stiefel and grassmann manifolds, Comptes Rendus Mathematique 350 (15–16) (2012) 773–776. doi:10.1016/j.crma.2012.08.010

    2012 doi

  40. [48]

    Stoye, R

    J. Stoye, R. Zimmermann, On the injectivity radius of the stiefel manifold: Numerical investigations and an explicit construction of a cut point at short distance (2024). arXiv:2403.03782. URL https://arxiv.org/abs/2403.03782

    2024 arXiv

  41. [49]

    Biliotti, Exponential map of a weak riemannian hilbert manifold (2006)

    L. Biliotti, Exponential map of a weak riemannian hilbert manifold (2006). arXiv:math/0402207. URL https://arxiv.org/abs/math/0402207

    2006 arXiv

  42. [50]

    Boumal, An introduction to optimization on smooth manifolds, Cambridge University Press, 2023

    N. Boumal, An introduction to optimization on smooth manifolds, Cambridge University Press, 2023

    2023

  43. [51]

    L. N. Trefethen, Chapter 13. Equispaced Points, Runge Phenomenon, SIAM, 2019, pp. 95–102. doi:10.1137/1.9781611975949.ch13

    2019 doi

  44. [52]

    Hashemi, L

    B. Hashemi, L. N. Trefethen, Chebfun in three dimensions, SIAM Journal on Scientific Computing 39 (2017) C341–C363. doi:10.1137/16M1083803;JOURNAL:JOURNAL:SJOCE3;PAGE:STRING:ARTICLE/CHAPTER. 32

    2017 doi

  45. [53]

    Dolgov, D

    S. Dolgov, D. Kressner, C. Str¨ ossner, Functional tucker approximation using cheby- shev interpolation, SIAM Journal on Scientific Computing 43 (2021) A2190–A2210. doi:10.1137/20M1356944;WGROUP:STRING:PUBLICATION. URL /doi/pdf/10.1137/20M1356944?download=true

    2021 doi

  46. [54]

    Gorodetsky, S

    A. Gorodetsky, S. Karaman, Y. Marzouk, A continuous analogue of the tensor-train de- composition, Computer Methods in Applied Mechanics and Engineering 347 (2019) 59–84. doi:10.1016/J.CMA.2018.12.015. URL https://www.sciencedirect.com/science/article/abs/pii/S0045782518306133

    2019 doi

  47. [55]

    Brivio, S

    S. Brivio, S. Fresca, N. R. Franco, A. Manzoni, Error estimates for pod-dl-roms: a deep learning framework for reduced order modeling of nonlinear parametrized pdes enhanced by proper orthog- onal decomposition, Advances in Computational Mathematics 50 (2024) 1–46. doi:10.1007...

    2024 doi

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.