pith. sign in

arxiv: 2502.12345 · v2 · submitted 2025-02-17 · 🧮 math.NA · cs.NA

Uncertainty quantification for stationary and time-dependent PDEs subject to Gevrey regular random domain deformations

Ana Djurdjevac , Vesa Kaarnioja , Claudia Schillings , Andr\'e-Alexander Zepernick This is my paper

Pith reviewed 2026-05-23 02:25 UTC · model grok-4.3

classification 🧮 math.NA cs.NA
keywords uncertainty quantificationquasi-Monte Carlo methodsrandom domainGevrey classPoisson equationheat equationfinite elementscubature
0
0 comments X

The pith

Randomly shifted lattice QMC rules deliver dimension-independent linear convergence for PDE expectations under Gevrey random domains

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

The paper develops an uncertainty quantification approach for PDEs whose domains are randomly deformed by an input field from a Gevrey smoothness class. This parameterization is more general than models that fix a specific series expansion such as Karhunen-Loève. The authors construct randomly shifted lattice quasi-Monte Carlo cubature rules to approximate the expectation of the solution for both the Poisson equation and the heat equation. They prove that these rules deliver dimension-independent convergence rates that are essentially linear in the number of points. The analysis is completed by bounding the additional errors that arise from truncating the dimension of the random field and from finite-element discretization of the PDEs.

Core claim

The paper shows that randomly shifted lattice quasi-Monte Carlo cubature rules achieve dimension-independent, essentially linear convergence rates for the expected solution of Poisson and heat equations when the random domain deformation is modeled by a Gevrey-class random field; the analysis incorporates dimension truncation and finite element errors.

What carries the argument

Randomly shifted lattice quasi-Monte Carlo cubature rules for the expectation integral over the parameter domain of the Gevrey random field

If this is right

  • Convergence rates remain dimension-independent for both stationary and time-dependent problems.
  • Total error is controlled by summing cubature, truncation, and discretization contributions.
  • Theoretical rates are observed in numerical experiments on the model problems.

Where Pith is reading between the lines

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

  • The Gevrey model may cover domain uncertainties arising in applications without a natural finite-parametric form.
  • The QMC technique could be extended to other time-dependent or nonlinear PDEs with comparable regularity.
  • High-dimensional integration costs may stay manageable for problems requiring many truncation terms.

Load-bearing premise

The random field parameterizing the domain deformation lies in a Gevrey smoothness class.

What would settle it

Numerical computation of the cubature error for increasing dimensions or reduced smoothness showing a breakdown of the linear rate or emergence of dimension dependence.

Figures

Figures reproduced from arXiv: 2502.12345 by Ana Djurdjevac, Andr\'e-Alexander Zepernick, Claudia Schillings, Vesa Kaarnioja.

Figure 1
Figure 1. Figure 1: Left: the triangulation of the reference domain [PITH_FULL_IMAGE:figures/full_fig_p031_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The computed L 2 (Dref ) errors for the Poisson equation. Left: the numerical cubature errors corresponding to experiments (E1)–(E2). Right: the numerical cubature errors corresponding to experiments (E3)–(E4). number of function evaluations nR R.M.S. error number of function evaluations nR R.M.S. error [PITH_FULL_IMAGE:figures/full_fig_p032_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The computed H1 0 (Dref ) errors for the Poisson equation. Left: the numerical cubature errors corresponding to experiments (E1)–(E2). Right: the numerical cubature errors corresponding to experiments (E3)–(E4). the computed errors are smaller than the errors measured using Sobolev norms. In addition, we note that the H1 0 norm appears to be sensitive to numerical discretization errors, resulting in a loss… view at source ↗
Figure 4
Figure 4. Figure 4: The computed L 2 (I; L 2 (Dref )) errors for the heat equation. Left: the numerical cubature errors corresponding to experiments (E1)–(E2). Right: the numerical cubature errors correspond￾ing to experiments (E3)–(E4). number of function evaluations nR R.M.S. error number of function evaluations nR R.M.S. error [PITH_FULL_IMAGE:figures/full_fig_p033_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: The computed L 2 (I; H1 0 (Dref )) errors for the heat equation. Left: the numerical cubature errors corresponding to experiments (E1)–(E2). Right: the numerical cubature errors correspond￾ing to experiments (E3)–(E4). certainty: while the previous literature is mainly concerned with holomorphic transformations of a fixed reference domain, the Gevrey class also covers non￾holomorphic representations. For t… view at source ↗
read the original abstract

We study uncertainty quantification for partial differential equations subject to domain uncertainty. We parameterize the random domain using the model recently considered by Chernov and Le (2024) as well as Harbrecht, Schmidlin, and Schwab (2024) in which the input random field is assumed to belong to a Gevrey smoothness class. This approach has the advantage of being substantially more general than models which assume a particular parametric representation of the input random field such as a Karhunen-Loeve series expansion. We consider both the Poisson equation as well as the heat equation and design randomly shifted lattice quasi-Monte Carlo (QMC) cubature rules for the computation of the expected solution under domain uncertainty. We show that these QMC rules exhibit dimension-independent, essentially linear cubature convergence rates in this framework. In addition, we complete the error analysis by taking into account the approximation errors incurred by dimension truncation of the random input field and finite element discretization. Numerical experiments are presented to confirm the theoretical rates.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 2 minor

Summary. The paper develops uncertainty quantification for the Poisson and heat equations on domains deformed by Gevrey-regular random fields (following Chernov-Le and Harbrecht-Schmidlin-Schwab). Randomly shifted lattice QMC rules are constructed for the expectation of the solution; the central claim is that these rules achieve dimension-independent convergence of order O(N^{-1+ε}) once the integrand is shown to lie in a suitable weighted Sobolev space. The analysis is completed by adding dimension-truncation and finite-element discretization errors, with numerical experiments confirming the rates.

Significance. If the regularity-transfer argument holds with d-uniform constants, the result meaningfully extends QMC theory beyond Karhunen-Loève parameterizations to a broader class of domain perturbations while treating both stationary and parabolic problems. The explicit inclusion of all three error sources (QMC, truncation, discretization) is a strength.

major comments (2)
  1. [§4] §4 (heat-equation analysis) and the proof of the weighted Sobolev norm bound: the argument that the solution map inherits Gevrey regularity with constants independent of truncation dimension d must be verified explicitly. The Duhamel integral or implicit time-stepping may introduce factors that grow with d or the Gevrey index, which would invalidate the claimed d-uniform bound on the mixed first derivatives needed for the QMC rate.
  2. [Theorem 3.3] Theorem 3.3 (or equivalent statement of the QMC error): the constant C in the O(N^{-1+ε}) bound is asserted to be independent of d, but the proof sketch relies on the input Gevrey assumption without an intermediate lemma showing that the PDE solution operator preserves the required decay |∂^α f(y)| ≤ C ∏ β_j^{α_j} with ∑ β_j < ∞ uniformly in d.
minor comments (2)
  1. [§2] Notation for the random field mapping (e.g., the precise definition of the Gevrey class and the domain perturbation operator) should be collected in a single preliminary section for readability.
  2. [§5] Figure captions for the numerical convergence plots should state the observed slopes and the precise values of the Gevrey index and truncation dimension used.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments on our manuscript. The two major points raised concern the explicit verification of d-uniform Gevrey regularity transfer for the heat equation and the intermediate step establishing d-independent constants in the QMC error bound. Both can be addressed by adding clarifying lemmas and expanded proof details without altering the main claims.

read point-by-point responses

  1. Referee: [§4] §4 (heat-equation analysis) and the proof of the weighted Sobolev norm bound: the argument that the solution map inherits Gevrey regularity with constants independent of truncation dimension d must be verified explicitly. The Duhamel integral or implicit time-stepping may introduce factors that grow with d or the Gevrey index, which would invalidate the claimed d-uniform bound on the mixed first derivatives needed for the QMC rate.

    Authors: We agree that the transfer of Gevrey regularity through the heat-equation solution operator requires an explicit intermediate statement. In the revision we will insert a new Lemma 4.3 that bounds the mixed derivatives of the solution via the Duhamel formula, using the fact that the Gevrey class is closed under the relevant time integrals and that the constants remain uniform in the truncation dimension d because the underlying domain-deformation regularity assumption is itself d-uniform. For the implicit Euler discretization we add a short paragraph showing that the stability constants are independent of d by the same Gevrey decay. These additions make the d-uniformity fully rigorous while leaving the overall convergence statement unchanged. revision: yes

  2. Referee: [Theorem 3.3] Theorem 3.3 (or equivalent statement of the QMC error): the constant C in the O(N^{-1+ε}) bound is asserted to be independent of d, but the proof sketch relies on the input Gevrey assumption without an intermediate lemma showing that the PDE solution operator preserves the required decay |∂^α f(y)| ≤ C ∏ β_j^{α_j} with ∑ β_j < ∞ uniformly in d.

    Authors: The referee correctly identifies that the passage from the Gevrey assumption on the domain deformation to the weighted Sobolev regularity of the integrand is only sketched. We will add Lemma 3.5 (placed immediately before Theorem 3.3) that derives the precise decay |∂^α u(y)| ≤ C ∏ β_j^{α_j} with ∑ β_j < ∞ and C independent of d, by applying the chain rule to the solution map and invoking the already-established Gevrey bounds on the domain map. This lemma directly supplies the hypothesis needed for the QMC error estimate, confirming that the constant in Theorem 3.3 is indeed d-independent. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation relies on external Gevrey model and standard QMC analysis

full rationale

The paper adopts its core Gevrey regularity assumption on the random domain deformation directly from two external 2024 references whose author lists do not overlap with the present team. The claimed dimension-independent QMC rates are asserted to follow from this imported model plus standard weighted Sobolev-space QMC theory applied to the PDE solution map; the abstract and provided text give no equations or steps in which a fitted parameter, self-defined quantity, or self-citation chain is renamed as a prediction. Truncation and FE errors are treated separately. Because every load-bearing modeling choice is externally sourced and no internal reduction of the target convergence statement to its own inputs is exhibited, the derivation chain remains non-circular.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The analysis depends on the Gevrey-class assumption for the random field and on the domain-deformation model introduced in the two cited 2024 papers; no free parameters or invented entities are introduced in the abstract.

axioms (1)
  • domain assumption Random domain deformations are parameterized by a Gevrey-regular random field as in Chernov-Le (2024) and Harbrecht et al. (2024).
    This is the modeling choice that enables the subsequent QMC analysis.

pith-pipeline@v0.9.0 · 5725 in / 1265 out tokens · 28612 ms · 2026-05-23T02:25:24.573197+00:00 · methodology

discussion (0)

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

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Transcoders Trace Visual Grounding and Hallucinations in Vision-Language Models

    cs.LG 2026-05 unverdicted novelty 6.0

    Transcoders decompose MLP layers in Gemma 3-4B-IT to trace visual grounding more effectively than SAEs and predict hallucinations from circuit graph features at AUC 0.68.

  2. Neural Shape Operator Surrogates -- Expression Rate Bounds

    cs.LG 2026-04 unverdicted novelty 6.0

    Neural and spectral operators can approximate shape-to-solution maps for families of elliptic and parabolic PDEs and BIEs with provable uniform error bounds derived from parametric holomorphy on a reference domain.

  3. The Root Theorem of Context Engineering

    cs.CC 2026-03 unverdicted novelty 3.0

    The Root Theorem states that context engineering is governed by maximizing signal-to-token ratio inside finite, lossy channels, which forces homeostatic architectures of accumulate-compress-rewrite-shed and external v...

Reference graph

Works this paper leans on

39 extracted references · 39 canonical work pages · cited by 3 Pith papers

  1. [1]

    Castrill´ on-Cand´ as, J.E., Nobile, F., Tempone, R.F.: Analytic regularity and collocation approximation for elliptic PDEs with random domain deforma tions. Comput. Math. Appl. 71(6), 1173–1197 (2016)

    work page 2016

  2. [2]

    Castrill´ on-Cand´ as, J.E., Xu, J.: A stochastic colloca tion approach for parabolic PDEs with random domain deformations. Comput. Math. Appl. 93, 32–49 (2021)

    work page 2021

  3. [3]

    Chernov, A., Lˆ e, T.: Analytic and Gevrey class regularit y for parametric elliptic eigen- value problems and applications. SIAM J. Numer. Anal. 4(62), 1874–1900 (2024)

    work page 1900

  4. [4]

    Chernov, A., Lˆ e, T.: Analytic and Gevrey class regularit y for parametric semilinear reaction-diffusion problems and applications in uncertain ty quantification. Comput. Math. Appl. 164, 116–130 (2024)

    work page 2024

  5. [5]

    Church, L., Djurdjevac, A., Elliot, C.M.: A domain mappin g approach for elliptic equa- tions posed on random bulk and surface domains. Numer. Math. 146, 1–49 (2020)

    work page 2020

  6. [6]

    Cohen, A., DeVore, R., Schwab, C.: Convergence rates of be st N -term Galerkin approx- imations for a class of elliptic sPDEs. Found. Comput. Math. 10, 615–646 (2010)

    work page 2010

  7. [7]

    Cools, R., Kuo, F.Y., Nuyens, D.: Constructing embedded l attice rules for multivariate integration. SIAM J. Sci. Comput. 28, 2162–2188 (2006)

    work page 2006

  8. [8]

    Corrado, C., Roney, C.H., Razeghi, O., Lemus, J.A.S., Cov eney, S., Sim, I., Williams, S.E., O’Neill, M.D., Wilkinson, R.D., Clayton, R.H., Niede rer, S.A.: Quantifying the impact of shape uncertainty on predicted arrhythmias. Comp ut. Biol. Med. 153, 106528 (2023)

    work page 2023

  9. [9]

    Springer (2022)

    Dick, J., Kritzer, P., Pillichshammer, F.: Lattice Rules : Numerical Integration, Approx- imation, and Discrepancy. Springer (2022)

    work page 2022

  10. [10]

    Acta Numer

    Dick, J., Kuo, F.Y., Sloan, I.H.: High-dimensional inte gration: the quasi-Monte Carlo way. Acta Numer. 22, 133–288 (2013)

    work page 2013

  11. [11]

    SIAM/ASA J

    Djurdjevac, A.: Linear parabolic problems in random mov ing domains. SIAM/ASA J. Uncertain. Quantif. 9(2), 848–879 (2021)

    work page 2021

  12. [12]

    D¨ olz, J., Harbrecht, H., Jerez-Hanckes, C., Multerer,M.: Isogeometric multilevel quadra- ture for forward and inverse random acoustic scattering. Co mput. Methods Appl. Mech. Engrg. 388, 114242 (2022)

    work page 2022

  13. [13]

    Springer (2021)

    Ern, A., Guermond, J.L.: Finite Elements III: First-Ord er and Time-Dependent PDEs. Springer (2021)

    work page 2021

  14. [14]

    American Mathematical Soci- ety, Providence, Rhode Island (2010)

    Evans, L.C.: Partial Differential Equations, second edn . American Mathematical Soci- ety, Providence, Rhode Island (2010)

    work page 2010

  15. [15]

    Graham, I.G., Kuo, F.Y., Nichols, J.A., Scheichl, R., Sc hwab, C., Sloan, I.H.: Quasi- Monte Carlo finite element methods for elliptic PDEs with log normal random coeffi- cients. Numer. Math. 131(2), 329–368 (2015)

    work page 2015

  16. [16]

    Graham, I.G., Kuo, F.Y., Nuyens, D., Scheichl, R., Sloan , I.H.: Quasi-Monte Carlo methods for elliptic PDEs with random coefficients and applic ations. J. Comput. Phys. 230(10), 3668–3694 (2011)

    work page 2011

  17. [17]

    Guth, P., Kaarnioja, V.: Generalized dimension truncat ion error analysis for high- dimensional numerical integration: lognormal setting and beyond. SIAM J. Numer. Anal. 62(2), 872–892 (2024)

    work page 2024

  18. [18]

    Hakula, H., Harbrecht, H., Kaarnioja, V., Kuo, F.Y., Slo an, I.H.: Uncertainty quan- tification for random domains using periodic random variabl es. Numer. Math. 156, 273–317 (2024)

    work page 2024

  19. [19]

    Harbrecht, H., Peters, M., Siebenmorgen, M.: Analysis o f the domain mapping method for elliptic diffusion problems on random domains. Numer. Ma th. 134(4), 823–856 (2016)

    work page 2016

  20. [20]

    Harbrecht, H., Schmidlin, M., Schwab, C.: The Gevrey cla ss implicit mapping theorem with application to UQ of semilinear elliptic PDEs. Math. Mo dels Methods Appl. Sci. 34(5), 881–917 (2024)

    work page 2024

  21. [21]

    Harbrecht, H., Schneider, R., Schwab, C.: Sparse second moment analysis for elliptic problems in stochastic domains. Numer. Math. 109(3), 385–414 (2008)

    work page 2008

  22. [22]

    Cambridge University Press, Cambridge, UK (1934) Uncertainty quantification for Gevrey regular random domai n deformations 39

    Hardy, G.H., Littlewood, J.E., P´ olya, G.: Inequalitie s. Cambridge University Press, Cambridge, UK (1934) Uncertainty quantification for Gevrey regular random domai n deformations 39

    work page 1934

  23. [23]

    Herrmann, L., Schwab, C.: QMC integration for lognormal -parametric, elliptic PDEs: local supports and product weights. Numer. Math. 141(1), 63–102 (2019)

    work page 2019

  24. [24]

    Hiptmair, R., Scarabosio, L., Schillings, C., Schwab, C .: Large deformation shape un- certainty quantification in acoustic scattering. Adv. Comp ut. Math. 44, 1475–1518 (2018)

    work page 2018

  25. [25]

    Preprint arXiv:2405 .03529 [math.NA] (2024)

    Kaarnioja, V., Schillings, C.: Quasi-Monte Carlo for Ba yesian design of experiment problems governed by parametric PDEs. Preprint arXiv:2405 .03529 [math.NA] (2024)

    work page 2024

  26. [26]

    Kunoth, A., Schwab, C.: Analytic regularity and GPC appr oximation for control prob- lems constrained by linear parametric elliptic and parabol ic PDEs. SIAM J. Control. Optim. 51(3), 2442–2471 (2013)

    work page 2013

  27. [27]

    https://people.cs.kuleuven.be/~dirk.nuyens/qmc4pde/

    Kuo, F.Y., Nuyens, D.: QMC4PDE software. https://people.cs.kuleuven.be/~dirk.nuyens/qmc4pde/

    work page

  28. [28]

    Kuo, F.Y., Nuyens, D.: Application of quasi-Monte Carlo methods to elliptic PDEs with random diffusion coefficients: a survey of analysis and implem entation. Found. Comput. Math. 16, 1631–1696 (2016)

    work page 2016

  29. [29]

    Kuo, F.Y., Nuyens, D.: Application of quasi-Monte Carlo methods to PDEs with random coefficients – an overview and tutorial. In: A.B. Owen, P.W. Gl ynn (eds.) Monte Carlo and Quasi-Monte Carlo Methods 2016, pp. 53–71. Stanford, CA , August 14–19 (2018)

    work page 2016

  30. [30]

    Kuo, F.Y., Schwab, C., Sloan, I.H.: Quasi-Monte Carlo fin ite element methods for a class of elliptic partial differential equations with rando m coefficients. SIAM J. Numer. Anal. 50(6), 3351–3374 (2012)

    work page 2012

  31. [31]

    Wiley & Sons Ltd, Hoboken, New Jersey (2019)

    Magnus, J.R., Neudecker, H.: Matrix Differential Calcul us with Applications in Statis- tics and Econometrics, third edn. Wiley & Sons Ltd, Hoboken, New Jersey (2019)

    work page 2019

  32. [32]

    Meidner, D., Vexler, B.: A priori error estimates for spa ce-time finite element discretiza- tion of parabolic optimal control problems part I: problems without control constraints. SIAM J. Control Optim. 47(3), 1150–1177 (2008)

    work page 2008

  33. [33]

    Nuyens, D., Cools, R.: Fast algorithms for component-by -component construction of rank-1 lattice rules in shift-invariant reproducing kerne l Hilbert spaces. Math. Comp. 75, 903–920 (2006)

    work page 2006

  34. [34]

    CRC Press (1996)

    Petkovˇ sek, M., Wilf, H.S., Zeilberger, D.: A = B. CRC Press (1996)

    work page 1996

  35. [35]

    Quaintance, J., Gould, H.W.: Combinatorial Identities for Stirling Numbers: The Un- published Notes of H. W. Gould. W orld Scientific Publishing C ompany, River Edge, NJ (2015)

    work page 2015

  36. [36]

    Savits, T.H.: Some statistical applications of Faa di Br uno. J. Multivariate Anal. 97(10), 2131–2140 (2006)

    work page 2006

  37. [37]

    Schwab, C., Stevenson, R.: Space-time adaptive wavelet methods for parabolic evolution problems. Math. Comp. 78(267), 1293–1318 (2009)

    work page 2009

  38. [38]

    CIRP An nals 66(1), 281–284 (2017)

    Tekkaya, A.E., Ben Khalifa, N., Hering, O., Meya, R., Mys licki, S., W alther, F.: Forming- induced damage and its effects on product properties. CIRP An nals 66(1), 281–284 (2017)

    work page 2017

  39. [39]

    Xiu, D., Tartakovsky, D.M.: Numerical methods for differ ential equations in random domains. SIAM J. Sci. Comput. 28(3), 1167–1185 (2006)

    work page 2006