pith. sign in

arxiv: 2605.19896 · v1 · pith:ZVGC6IR2new · submitted 2026-05-19 · 🧮 math.NA · cs.NA· math.OC

Adaptive Reduced-Basis Trust-Region Methods for Defect Identification in Elastic Materials

Benedikt Klein , Mario Ohlberger , Thomas Schuster This is my paper

Pith reviewed 2026-05-20 01:47 UTC · model grok-4.3

classification 🧮 math.NA cs.NAmath.OC
keywords reduced basis methodstrust-region methodsdefect identificationelastic wave equationparameter identificationGauss-Newton methodhyperbolic systemsreduced order modeling
0
0 comments X

The pith

An adaptive reduced-basis trust-region framework yields reliable online-efficient surrogates for defect identification in hyperbolic elastic systems.

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

The paper addresses the computational demands of identifying defects in elastic materials from ultrasonic surface displacement measurements, where the governing hyperbolic wave equation leads to a high-dimensional inverse problem. Standard stabilization via the iteratively regularized Gauss-Newton method requires repeated expensive forward and adjoint solves. The proposed approach constructs adaptive reduced-basis spaces that simultaneously reduce the state and parameter spaces, producing fast surrogate models for these evaluations. These surrogates are embedded inside an adaptive trust-region framework that controls approximation accuracy during the iterations. The work extends earlier reduced-basis trust-region results from elliptic and parabolic regimes to the hyperbolic elastic case and validates the method with numerical defect-detection experiments.

Core claim

The authors establish that adaptively constructed reduced-basis spaces, when placed inside a trust-region framework, deliver online-efficient surrogate models for both forward and adjoint evaluations inside the iteratively regularized Gauss-Newton method while guaranteeing the reliability of the reduced-order approximations for defect identification in hyperbolic elastic systems.

What carries the argument

Adaptively constructed reduced-basis spaces for joint state and parameter reduction, controlled by an adaptive trust-region framework that enforces accuracy of the surrogates during IRGNM iterations.

If this is right

  • The method produces online-efficient surrogate models for forward and adjoint evaluations required by derivative-based optimization.
  • Reliability of the reduced-order approximations is guaranteed by the trust-region mechanism during iteration.
  • The technique successfully extends reduced-basis trust-region ideas from elliptic and parabolic problems to hyperbolic elastic systems.
  • Numerical experiments confirm reliability and effectiveness for defect detection in elastic materials.

Where Pith is reading between the lines

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

  • The same adaptive reduced-basis trust-region construction could be tested on inverse problems for other hyperbolic systems such as acoustic or electromagnetic waves.
  • If the online cost reduction is large enough, the approach could support near-real-time structural health monitoring from ultrasonic data.
  • Combining the adaptive basis construction with data-driven techniques might further accelerate basis updates for new material configurations.

Load-bearing premise

The adaptively constructed reduced-basis spaces, when embedded in the trust-region framework, maintain sufficient accuracy for the IRGNM iterations without introducing unacceptable errors in the hyperbolic setting.

What would settle it

Numerical runs in which the trust-region tolerances are satisfied yet the reconstructed defect locations deviate substantially from ground truth, or in which the IRGNM diverges due to accumulated reduced-order errors, would falsify the reliability claim.

read the original abstract

Monitoring the integrity of elastic structures using ultrasonic waves requires the efficient identification of material parameters from measured surface displacements. The displacement field is governed by Cauchy's equation of motion, i.e., an elastic wave equation. Consequently, defect localization leads to a high-dimensional spatial parameter identification problem for a hyperbolic system with given initial and boundary conditions. Stable parameter reconstructions typically rely on regularization techniques such as the iteratively regularized Gauss--Newton method (IRGNM). However, its practical application is computationally demanding due to the high-dimensional nature of the problem. To address this bottleneck, we propose a reduced-order modeling approach that simultaneously reduces the state and parameter spaces using adaptively constructed reduced-basis spaces. This yields online-efficient surrogate models for both the forward and adjoint evaluations required in derivative-based optimization. To ensure reliability, the IRGNM iteration is embedded into an adaptive, trust-region framework that provides accuracy of the reduced-order approximations. The approach extends our recent contributions, which focus on elliptic and parabolic problems, to the hyperbolic setting. We demonstrate the reliability and effectiveness of the method for defect detection through numerical experiments.

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 proposes an adaptive reduced-basis trust-region framework to solve high-dimensional defect identification problems in elastic materials governed by the hyperbolic elastic wave equation. Reduced-basis surrogates are constructed adaptively for both the forward problem and the adjoint problem arising in the iteratively regularized Gauss-Newton method (IRGNM); these surrogates are embedded inside a trust-region loop whose acceptance criteria are intended to guarantee that the reduced-order approximations remain sufficiently accurate for reliable parameter updates. The method extends prior reduced-basis trust-region work from elliptic and parabolic regimes to the hyperbolic case and is illustrated by numerical experiments on defect localization from surface displacement data.

Significance. A reliable and online-efficient reduced-order approach for derivative-based inversion of hyperbolic systems would be a meaningful contribution to computational inverse problems in structural health monitoring. The combination of adaptive reduced-basis projection with trust-region control of approximation quality is a natural way to address the tension between computational cost and stability in IRGNM iterations. If the numerical experiments confirm that the trust-region mechanism prevents unacceptable accumulation of dispersion or phase errors, the work would strengthen the case for reduced-basis methods in time-dependent wave problems.

major comments (2)
  1. [§4] §4 (Trust-Region Algorithm): the acceptance criterion for the trust-region radius is stated in terms of a generic reduced-basis error indicator, but no explicit a-posteriori bound or stability estimate is given for the hyperbolic operator that would control the accumulation of dispersion errors over the propagation time interval; this bound is load-bearing for the claim that the framework guarantees reliability of the IRGNM iterates.
  2. [§3.2] §3.2 (Reduced-Basis Error Estimators): the a-posteriori estimators for the forward and adjoint reduced solutions are derived from the elliptic/parabolic setting; it is not shown that the same estimators remain rigorous or sufficiently sharp when applied to the second-order hyperbolic system, where small phase errors can corrupt the parameter update even if the L2-norm error appears controlled.
minor comments (2)
  1. [Numerical Experiments] The numerical experiments section would benefit from a table reporting the number of basis functions retained, the online speedup factor, and the final reconstruction error for each test case.
  2. [§2] Notation for the parameter-to-observable map and the reduced-basis projection operators should be introduced once and used consistently throughout the trust-region loop description.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments on our manuscript. The points raised regarding the theoretical underpinnings of the trust-region framework and error estimators in the hyperbolic setting are well taken. We address each major comment below and have made revisions to strengthen the presentation and clarify the reliability aspects.

read point-by-point responses

  1. Referee: [§4] §4 (Trust-Region Algorithm): the acceptance criterion for the trust-region radius is stated in terms of a generic reduced-basis error indicator, but no explicit a-posteriori bound or stability estimate is given for the hyperbolic operator that would control the accumulation of dispersion errors over the propagation time interval; this bound is load-bearing for the claim that the framework guarantees reliability of the IRGNM iterates.

    Authors: We agree that an explicit stability estimate controlling dispersion error accumulation would provide stronger theoretical support. The current acceptance criterion uses a residual-based reduced-basis error indicator that has proven effective in practice for the time-dependent wave problem. In the revised manuscript we have expanded the discussion in Section 4 to explain how the trust-region radius adaptation, combined with monitoring of the reduced objective-function change, limits the propagation of phase errors in the IRGNM updates. Additional numerical diagnostics have been included to illustrate that rejected steps correlate with regions where dispersion would otherwise degrade the gradient. A fully rigorous a-posteriori bound for arbitrary propagation times remains technically demanding and is noted as a direction for future analysis; the present framework relies on the combination of the indicator and trust-region safeguard for practical reliability. revision: yes

  2. Referee: [§3.2] §3.2 (Reduced-Basis Error Estimators): the a-posteriori estimators for the forward and adjoint reduced solutions are derived from the elliptic/parabolic setting; it is not shown that the same estimators remain rigorous or sufficiently sharp when applied to the second-order hyperbolic system, where small phase errors can corrupt the parameter update even if the L2-norm error appears controlled.

    Authors: The residual-based estimators in Section 3.2 are derived from the weak form of the governing equation and therefore carry over directly to the second-order hyperbolic system without modification of the residual computation. To address the specific concern about phase errors, the revised manuscript adds a short analysis subsection showing that the trust-region acceptance test, which compares the reduced and full-order objective values, rejects updates when phase discrepancies would materially affect the Gauss-Newton step. Numerical experiments in the paper already demonstrate that the IRGNM iterates converge to accurate defect locations with controlled L2 errors; we have augmented these results with a brief sensitivity study confirming that the estimators remain sufficiently sharp for the surface-displacement data considered. We believe this combination of residual control and trust-region filtering suffices for the application, while acknowledging that sharper hyperbolic-specific bounds would be desirable. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation combines standard components with explicit error control

full rationale

The paper builds its adaptive reduced-basis trust-region framework by extending established reduced-basis projection, IRGNM, and trust-region concepts to the hyperbolic elastic setting. The trust-region loop is introduced specifically to enforce accuracy of the surrogates during iterations rather than presupposing or defining that accuracy. Self-citation to prior elliptic/parabolic work is present but serves only as background for the extension; no load-bearing uniqueness theorem, ansatz, or fitted input is reduced to a self-referential definition or prior result by the authors. The central reliability claim rests on the proposed adaptive construction and numerical experiments, which remain independent of any circular reduction in the provided text.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the standard modeling assumption that the displacement field obeys Cauchy's equation of motion together with the usual well-posedness of the forward and adjoint hyperbolic problems; no new free parameters or invented entities are introduced in the abstract.

axioms (1)
  • domain assumption The displacement field is governed by Cauchy's equation of motion (elastic wave equation) with given initial and boundary conditions.
    Explicitly stated in the abstract as the governing PDE for the forward problem.

pith-pipeline@v0.9.0 · 5728 in / 1332 out tokens · 40711 ms · 2026-05-20T01:47:42.834868+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

61 extracted references · 61 canonical work pages · 1 internal anchor

  1. [1]

    Computational Science and Engineering1(3) (2024) https://doi.org/10.1007/ s44207-024-00002-z 32

    Kartmann, M., Keil, T., Ohlberger, M., Volkwein, S., Kaltenbacher, B.: Adap- tive reduced basis trust region methods for parameter identification problems. Computational Science and Engineering1(3) (2024) https://doi.org/10.1007/ s44207-024-00002-z 32

    work page 2024

  2. [2]

    Inverse Problems41(12), 125006–36 (2025) https://doi.org/10.1088/1361-6420/ae2a66

    Kartmann, M., Klein, B., Ohlberger, M., Schuster, T., Volkwein, S.: Adap- tive reduced basis trust region methods for parabolic inverse problems. Inverse Problems41(12), 125006–36 (2025) https://doi.org/10.1088/1361-6420/ae2a66

    work page doi:10.1088/1361-6420/ae2a66 2025

  3. [3]

    Giurgiutiu, V.: Structural Health Monitoring with Piezoelectric Wafer Active Sensors, Second edition. edn. Academic Press Elsevier, Amsterdam (2014). https: //doi.org/10.1016/C2013-0-00155-7

    work page doi:10.1016/c2013-0-00155-7 2014

  4. [4]

    (eds.): Lamb-Wave Based Structural Health Monitoring in Polymer Composites

    Lammering, R., Grabbert, U., Sinapius, M., Schuster, T., Wierach, P. (eds.): Lamb-Wave Based Structural Health Monitoring in Polymer Composites. Research Topics in Aerospace. Springer, Cham (2018). https://doi.org/10.1007/ 978-3-319-49715-0

    work page 2018

  5. [5]

    Frontiers in Materials7, 148 (2020) https: //doi.org/10.3389/fmats.2020.00148

    Yang, S., Li, P., Guo, M., Liao, S., Wu, H.: Study on dynamic load monitoring of an enhanced stress absorption layer. Frontiers in Materials7, 148 (2020) https: //doi.org/10.3389/fmats.2020.00148

    work page doi:10.3389/fmats.2020.00148 2020

  6. [6]

    Inverse Problems in Science and Engineering 25(12), 1768–1787 (2017) https://doi.org/10.1080/17415977.2017.1289195

    Jadamba, B., Khan, A.A., Oberai, A.A., Sama, M.: First-order and second-order adjoint methods for parameter identification problems with an application to the elasticity imaging inverse problem. Inverse Problems in Science and Engineering 25(12), 1768–1787 (2017) https://doi.org/10.1080/17415977.2017.1289195

    work page doi:10.1080/17415977.2017.1289195 2017

  7. [7]

    Computers & Structures234(2020) https://doi.org/10.1016/j.compstruc.2020.106254

    Sun, Y., Luo, L., Chen, K., Qin, X., Zhang, Q.: A time-domain method for load identification using moving weighted least square technique. Computers & Structures234(2020) https://doi.org/10.1016/j.compstruc.2020.106254

    work page doi:10.1016/j.compstruc.2020.106254 2020

  8. [8]

    Sensors19(1), 103 (2019) https://doi

    Soman, R., Ostachowicz, W.: Kalman filter based load monitoring in beam like structures using fibre-optic strain sensors. Sensors19(1), 103 (2019) https://doi. org/10.3390/s19010103

    work page doi:10.3390/s19010103 2019

  9. [9]

    Patel, D., Tibrewala, R., Vega, A., Dong, L., Hugenberg, N., Oberai, A.A.: Cir- cumventing the solution of inverse problems in mechanics through deep learning: application to elasticity imaging. Comput. Methods Appl. Mech. Engrg.353, 448–466 (2019) https://doi.org/10.1016/j.cma.2019.04.045

    work page doi:10.1016/j.cma.2019.04.045 2019

  10. [10]

    Sensors and Actuators A: Physical281, 31–41 (2018) https://doi.org/10.1016/j.sna.2018.08.023

    Iele, A., Leone, M., Consales, M., Persiano, G.V., Brindisi, A., Ameduri, S., Concilio, A., Ciminello, M., Apicella, A., Bocchetto, F., Cusano, A.: Load mon- itoring of aircraft landing gears using fiber optic sensors. Sensors and Actuators A: Physical281, 31–41 (2018) https://doi.org/10.1016/j.sna.2018.08.023

    work page doi:10.1016/j.sna.2018.08.023 2018

  11. [11]

    Measurement89, 197–203 (2016) https://doi.org/10.1016/j.measurement.2016.04.013

    Tashakori, S., Baghalian, A., Unal, M., Fekrmandi, H., Seny¨ urek, V., McDaniel, D., Tansel, I.N.: Contact and non-contact approaches in load monitoring appli- cations using surface response to excitation method. Measurement89, 197–203 (2016) https://doi.org/10.1016/j.measurement.2016.04.013

    work page doi:10.1016/j.measurement.2016.04.013 2016

  12. [12]

    Inverse Problems 21(2) (2005) https://doi.org/10.1088/0266-5611/21/2/R01 33

    Bonnet, M., Constantinescu, A.: Inverse problems in elasticity. Inverse Problems 21(2) (2005) https://doi.org/10.1088/0266-5611/21/2/R01 33

    work page doi:10.1088/0266-5611/21/2/r01 2005

  13. [13]

    Continuum Mechan- ics and Thermodynamics36(6), 1413–1453 (2024) https://doi.org/10.1007/ s00161-024-01314-3

    Fedele, R., Placidi, L., Fabbrocino, F.: A review of inverse problems for gener- alized elastic media: formulations, experiments, synthesis. Continuum Mechan- ics and Thermodynamics36(6), 1413–1453 (2024) https://doi.org/10.1007/ s00161-024-01314-3

    work page 2024

  14. [14]

    Structural and Multidisciplinary Optimization37, 609–623 (2009) https://doi.org/10.1007/ s00158-008-0249-0

    Jankowski, L.: Off-line identification of dynamic loads. Structural and Multidisciplinary Optimization37, 609–623 (2009) https://doi.org/10.1007/ s00158-008-0249-0

    work page 2009

  15. [15]

    Structural and Multidisciplinary Optimization41, 243–253 (2010) https://doi

    Zhang, Q., Jankowski, L., Duan, Z.: Identification of coexistent load and damage. Structural and Multidisciplinary Optimization41, 243–253 (2010) https://doi. org/10.1007/s00158-009-0421-1

    work page doi:10.1007/s00158-009-0421-1 2010

  16. [16]

    Taddei, T., Penn, J.D., Yano, M., Patera, A.T.: Simulation-based classification; a model-order-reduction approach for structural health monitoring. Arch. Comput. Methods Eng.25(1), 23–45 (2018) https://doi.org/10.1007/s11831-016-9185-0

    work page doi:10.1007/s11831-016-9185-0 2018

  17. [17]

    Bigoni, C., Hesthaven, J.S.: Simulation-based anomaly detection and damage localization: an application to structural health monitoring. Comput. Methods Appl. Mech. Engrg.363, 112896–30 (2020) https://doi.org/10.1016/j.cma.2020. 112896

    work page doi:10.1016/j.cma.2020 2020

  18. [18]

    Mechanical Systems and Signal Processing197(2023) https://doi.org/ 10.1016/j.ymssp.2023.110376

    Torzoni, M., Manzoni, A., Mariani, S.: A multi-fidelity surrogate model for struc- tural health monitoring exploiting model order reduction and artificial neural networks. Mechanical Systems and Signal Processing197(2023) https://doi.org/ 10.1016/j.ymssp.2023.110376

    work page doi:10.1016/j.ymssp.2023.110376 2023

  19. [19]

    Advanced Science10(18), 2300439 (2023) https: //doi.org/10.1002/advs.202300439

    Chen, C.-T., Gu, G.X.: Physics-informed deep-learning for elasticity: forward, inverse, and mixed problems. Advanced Science10(18), 2300439 (2023) https: //doi.org/10.1002/advs.202300439

    work page doi:10.1002/advs.202300439 2023

  20. [20]

    Mrs Bulletin46, 19–25 (2021) https://doi.org/10.1557/ s43577-020-00006-y

    Ni, B., Gao, H.: A deep learning approach to the inverse problem of modulus identification in elasticity. Mrs Bulletin46, 19–25 (2021) https://doi.org/10.1557/ s43577-020-00006-y

    work page 2021

  21. [21]

    arXiv preprint arXiv:2502.05463 (2025) https://doi.org/10.48550/arXiv.2502.05463

    Bhattacharya, K., Cao, L., Stepaniants, G., Stuart, A., Trautner, M.: Learning memory and material dependent constitutive laws. arXiv preprint arXiv:2502.05463 (2025) https://doi.org/10.48550/arXiv.2502.05463

    work page doi:10.48550/arxiv.2502.05463 2025

  22. [22]

    Bhattacharya, K., Cao, L., Stuart, A.: Optimal experimental design for reliable learning of history-dependent constitutive laws. Comput. Methods Appl. Mech. Engrg.457, 119022 (2026) https://doi.org/10.1016/j.cma.2026.119022

    work page doi:10.1016/j.cma.2026.119022 2026

  23. [23]

    Smart Materials and Structures 22(8) (2013) https://doi.org/10.1088/0964-1726/22/8/085014 34

    Ghajari, M., Sharif-Khodaei, Z., Aliabadi, M.H., Apicella, A.: Identification of impact force for smart composite stiffened panels. Smart Materials and Structures 22(8) (2013) https://doi.org/10.1088/0964-1726/22/8/085014 34

    work page doi:10.1088/0964-1726/22/8/085014 2013

  24. [25]

    Inverse Problems33(12), 124004 (2017) https://doi.org/10.1088/1361-6420/aa8d91

    Seydel, J., Schuster, T.: Identifying the stored energy of a hyperelastic struc- ture by using an attenuated landweber method. Inverse Problems33(12), 124004 (2017) https://doi.org/10.1088/1361-6420/aa8d91

    work page doi:10.1088/1361-6420/aa8d91 2017

  25. [26]

    Mathematical Methods in the Applied Sciences40(1), 183–204 (2017) https:// doi.org/10.1002/mma.3979

    Seydel, J., Schuster, T.: On the linearization of identifying the stored energy function of a hyperelastic material from full knowledge of the displacement field. Mathematical Methods in the Applied Sciences40(1), 183–204 (2017) https:// doi.org/10.1002/mma.3979

    work page doi:10.1002/mma.3979 2017

  26. [27]

    Applicable Analysis94(8), 1561–1593 (2015) https://doi.org/10.1080/00036811.2014.940519

    Woestehoff, A., Schuster, T.: Uniqueness and stability result for cauchy’s equation of motion for a certain class of hyperelastic materials. Applicable Analysis94(8), 1561–1593 (2015) https://doi.org/10.1080/00036811.2014.940519

    work page doi:10.1080/00036811.2014.940519 2015

  27. [28]

    Inverse Problems13(1), 79–95 (1997) https://doi.org/10.1088/0266-5611/13/1/007

    Hanke, M.: A regularizing Levenberg-Marquardt scheme with applications to inverse groundwater filtration problems. Inverse Problems13(1), 79–95 (1997) https://doi.org/10.1088/0266-5611/13/1/007

    work page doi:10.1088/0266-5611/13/1/007 1997

  28. [29]

    Langer, S., Hohage, T.: Convergence analysis of an inexact iteratively regularized Gauss-Newton method under general source conditions. J. Inverse Ill-Posed Probl. 15(3), 311–327 (2007) https://doi.org/10.1515/jiip.2007.017

    work page doi:10.1515/jiip.2007.017 2007

  29. [30]

    all-at-once formulations

    Kaltenbacher, B., Kirchner, A., Vexler, B.: Goal oriented adaptivity in the IRGNM for parameter identification in PDEs: II. all-at-once formulations. Inverse Problems30(2), 045002 (2014) https://doi.org/10.1088/0266-5611/30/4/045002

    work page doi:10.1088/0266-5611/30/4/045002 2014

  30. [31]

    reduced formulation

    Kaltenbacher, B., Kirchner, A., Veljovi´ c, S.: Goal oriented adaptivity in the IRGNM for parameter identification in PDEs: I. reduced formulation. Inverse Problems30(4), 045001 (2014) https://doi.org/10.1088/0266-5611/30/4/045001

    work page doi:10.1088/0266-5611/30/4/045001 2014

  31. [32]

    Numerische Mathematik140, 449–478 (2018) https: //doi.org/10.1007/s00211-018-0971-5

    Kaltenbacher, B., Souza, M.L.: Convergence and adaptive discretization of the IRGNM Tikhonov and the IRGNM Ivanov method under a tangential cone con- dition in Banach space. Numerische Mathematik140, 449–478 (2018) https: //doi.org/10.1007/s00211-018-0971-5

    work page doi:10.1007/s00211-018-0971-5 2018

  32. [34]

    Springer, Heidelberg (2021)

    Kaltenbacher, B., Schuster, T., Wald, A.: Time-dependent Problems in Imaging and Parameter Identification. Springer, Heidelberg (2021). https://doi.org/10. 1007/978-3-030-57784-1 35

    work page 2021

  33. [35]

    Time-dependent Problems in Imaging and Parameter Identification, 377–412 (2021) https://doi.org/10.1007/978-3-030-57784-1 13

    Kaltenbacher, B., Nguyen, T.T.N., Wald, A., Schuster, T.: Parameter identifi- cation for the Landau–Lifshitz–Gilbert equation in magnetic particle imaging. Time-dependent Problems in Imaging and Parameter Identification, 377–412 (2021) https://doi.org/10.1007/978-3-030-57784-1 13

    work page doi:10.1007/978-3-030-57784-1 2021

  34. [36]

    SIAM Journal on Scientific Computing 32(5), 2523–2542 (2010) https://doi.org/10.1137/090775622

    Lieberman, C., Willcox, K., Ghattas, O.: Parameter and state model reduction for large-scale statistical inverse problems. SIAM Journal on Scientific Computing 32(5), 2523–2542 (2010) https://doi.org/10.1137/090775622

    work page doi:10.1137/090775622 2010

  35. [37]

    Himpe, C., Ohlberger, M.: Data-driven combined state and parameter reduction for inverse problems. Adv. Comput. Math.41(5), 1343–1364 (2015) https://doi. org/10.1007/s10444-015-9420-5

    work page doi:10.1007/s10444-015-9420-5 2015

  36. [38]

    Model reduction and approximation: theory and algorithms 15(1) (2017) https://doi.org/10.1137/1.9781611974829.ch1

    Gubisch, M., Volkwein, S.: Proper orthogonal decomposition for linear-quadratic optimal control. Model reduction and approximation: theory and algorithms 15(1) (2017) https://doi.org/10.1137/1.9781611974829.ch1

    work page doi:10.1137/1.9781611974829.ch1 2017

  37. [39]

    Acta Numer.30, 445–554 (2021) https://doi.org/10.1017/S0962492921000064

    Ghattas, O., Willcox, K.: Learning physics-based models from data: perspectives from inverse problems and model reduction. Acta Numer.30, 445–554 (2021) https://doi.org/10.1017/S0962492921000064

    work page doi:10.1017/s0962492921000064 2021

  38. [40]

    arXiv preprint arXiv:2512.14086 (2025) https://doi.org/10.48550/arXiv.2512.14086

    Yao, B., Luo, D., Cao, L., Kovachki, N., O’Leary-Roseberry, T., Ghattas, O.: Derivative-informed fourier neural operator: Universal approximation and appli- cations to PDE-constrained optimization. arXiv preprint arXiv:2512.14086 (2025) https://doi.org/10.48550/arXiv.2512.14086

    work page doi:10.48550/arxiv.2512.14086 2025

  39. [41]

    (eds.): Model Reduction and Approximation

    Benner, P., Cohen, A., Ohlberger, M., Willcox, K. (eds.): Model Reduction and Approximation. Theory and Algorithms. Computational Science & Engineer- ing, vol. 15, p. 412. Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA (2017). https://doi.org/10.1137/1.9781611974829

    work page doi:10.1137/1.9781611974829 2017

  40. [42]

    Springer, Cham (2015)

    Quarteroni, A., Manzoni, A., Negri, F.: Reduced Basis Methods for Partial Dif- ferential Equations: An Introduction. Springer, Cham (2015). https://doi.org/10. 1007/978-3-319-15431-2

    work page 2015

  41. [43]

    John Wiley & Sons, Chichester, England (2002)

    Holzapfel, G.A.: Nonlinear Solid Mechanics: A Continuum Approach for Engi- neering Science. John Wiley & Sons, Chichester, England (2002)

    work page 2002

  42. [44]

    Mathematical Modelling and Numerical Analysis42(2), 277–302 (2008) https://doi.org/10.1051/m2an: 2008001

    Haasdonk, B., Ohlberger, M.: Reduced basis method for finite volume approx- imations of parametrized linear evolution equations. Mathematical Modelling and Numerical Analysis42(2), 277–302 (2008) https://doi.org/10.1051/m2an: 2008001

    work page doi:10.1051/m2an: 2008

  43. [46]

    SIAM Journal on Scientific Computing40(5), 3267–3292 (2018) https://doi.org/10.1137/16M1085413

    Himpe, C., Leibner, T., Rave, S.: Hierarchical approximate proper orthogonal decomposition. SIAM Journal on Scientific Computing40(5), 3267–3292 (2018) https://doi.org/10.1137/16M1085413

    work page doi:10.1137/16m1085413 2018

  44. [47]

    SIAM Journal on Scientific Computing38, 194–216 (2016) https://doi.org/10.1137/15M1026614

    Milk, R., Rave, S., Schindler, F.: pyMOR – Generic algorithms and interfaces for model order reduction. SIAM Journal on Scientific Computing38, 194–216 (2016) https://doi.org/10.1137/15M1026614

    work page doi:10.1137/15m1026614 2016

  45. [48]

    Journal of Numerical Mathematics32(4), 369–380 (2024) https://doi.org/10.1515/jnma-2024-0137

    Africa, P.C., Arndt, D., Bangerth, W., Blais, B., Fehling, M., Gassm¨ oller, R., Heister, T., Heltai, L., Kinnewig, S., Kronbichler, M., Maier, M., Munch, P., Schreter-Fleischhacker, M., Thiele, J.P., Turcksin, B., Wells, D., Yushutin, V.: The deal.II library, version 9.6. Journal of Numerical Mathematics32(4), 369–380 (2024) https://doi.org/10.1515/jnma-...

    work page doi:10.1515/jnma-2024-0137 2024

  46. [49]

    Inverse Problems31(2), 025006 (2015) https://doi.org/10.1088/0266-5611/31/2/ 025006

    Binder, F., Sch¨ opfer, F., Schuster, T.: Defect localization in fibre-reinforced com- posites by computing external volume forces from surface sensor measurements. Inverse Problems31(2), 025006 (2015) https://doi.org/10.1088/0266-5611/31/2/ 025006

    work page doi:10.1088/0266-5611/31/2/ 2015

  47. [50]

    Cambridge Studies in Advanced Math- ematics

    Wloka, J.: Partial Differential Equations. Cambridge Studies in Advanced Math- ematics. Cambridge University Press, Cambridge, UK (1987). https://doi.org/10. 1017/CBO9781139171755

    work page 1987

  48. [51]

    Inverse Problems25(6), 065003 (2009) https://doi.org/ 10.1088/0266-5611/25/6/065003

    Kaltenbacher, B., Sch¨ opfer, F., Schuster, T.: Iterative methods for nonlinear ill- posed problems in banach spaces: convergence and applications to parameter identification problems. Inverse Problems25(6), 065003 (2009) https://doi.org/ 10.1088/0266-5611/25/6/065003

    work page doi:10.1088/0266-5611/25/6/065003 2009

  49. [52]

    Nocedal, J

    Nocedal, J., Wright, S.J.: Numerical Optimization, 2nd edn. Springer Series in Operations Research and Financial Engineering. Springer, New York (2006). https://doi.org/10.1007/978-0-387-40065-5

    work page doi:10.1007/978-0-387-40065-5 2006

  50. [53]

    International Journal of Computational Fluid Dynamics34(2), 139–146 (2020) https://doi.org/10.1080/10618562.2019.1686486

    Glas, S., Patera, A.T., Urban, K.: A reduced basis method for the wave equation. International Journal of Computational Fluid Dynamics34(2), 139–146 (2020) https://doi.org/10.1080/10618562.2019.1686486

    work page doi:10.1080/10618562.2019.1686486 2020

  51. [54]

    Mathematical Models and Methods in Applied Sciences15(02), 199–225 (2005) https://doi.org/10.1142/S0218202505000339

    Bernardi, C., S¨ uli, E.: Time and space adaptivity for the second-order wave equation. Mathematical Models and Methods in Applied Sciences15(02), 199–225 (2005) https://doi.org/10.1142/S0218202505000339

    work page doi:10.1142/s0218202505000339 2005

  52. [55]

    SIAM Journal on Scientific Computing39(5), 434–460 (2017) https://doi.org/10.1137/16M1081981

    Qian, E., Grepl, M., Veroy, K., Willcox, K.: A certified trust region reduced basis approach to PDE-constrained optimization. SIAM Journal on Scientific Computing39(5), 434–460 (2017) https://doi.org/10.1137/16M1081981

    work page doi:10.1137/16m1081981 2017

  53. [56]

    Klein, B., Ohlberger, M.: Multi-fidelity learning of reduced order models for parabolic PDE constrained optimization. Adv. Comput. Math.52(2), 19–36 (2026) https://doi.org/10.1007/s10444-026-10296-6 37

    work page doi:10.1007/s10444-026-10296-6 2026

  54. [57]

    Yue, Y., Meerbergen, K.: Accelerating optimization of parametric linear systems by model order reduction. SIAM J. Optim.23(2), 1344–1370 (2013) https://doi. org/10.1137/120869171

    work page doi:10.1137/120869171 2013

  55. [59]

    ESAIM: Mathematical Modelling and Numerical Analysis58(1), 79–105 (2024) https://doi.org/10.1051/m2an/ 2023089

    Keil, T., Ohlberger, M.: A relaxed localized trust-region reduced basis approach for optimization of multiscale problems. ESAIM: Mathematical Modelling and Numerical Analysis58(1), 79–105 (2024) https://doi.org/10.1051/m2an/ 2023089

    work page doi:10.1051/m2an/ 2024

  56. [60]

    arXiv preprint arXiv:2012.11653 (2020) https://doi.org/10.48550/arXiv.2012.11653

    Banholzer, S., Keil, T., Mechelli, L., Ohlberger, M., Schindler, F., Volkwein, S.: An adaptive projected newton non-conforming dual approach for trust-region reduced basis approximation of PDE-constrained parameter optimization. arXiv preprint arXiv:2012.11653 (2020) https://doi.org/10.48550/arXiv.2012.11653

    work page doi:10.48550/arxiv.2012.11653 2012

  57. [61]

    Inverse Problems in Science and Engineer- ing25(1), 2–26 (2017) https://doi.org/10.1080/17415977.2015.1132713

    Lechleiter, A., Schlasche, J.W.: Identifying Lam´ e parameters from time- dependent elastic wave measurements. Inverse Problems in Science and Engineer- ing25(1), 2–26 (2017) https://doi.org/10.1080/17415977.2015.1132713

    work page doi:10.1080/17415977.2015.1132713 2017

  58. [62]

    Do large language models perform latent multi-hop reasoning without exploiting shortcuts? abs/2411.16679, 2024

    Azmi, B., Bernreuther, M.: On the nonmonotone linesearch for a class of infinite- dimensional nonsmooth problems. arXiv (2023) https://doi.org/10.48550/arXiv. 2303.01878

    work page internal anchor Pith review doi:10.48550/arxiv 2023

  59. [63]

    Adaptive Reduced-Basis Trust-Region Methods for Defect Identification in Elastic Materials

    Klein, B., Ohlberger, M., Schuster, T.: Source code to “Adaptive Reduced-Basis Trust-Region Methods for Defect Identification in Elastic Materials”. Zenodo (2026). https://doi.org/10.5281/zenodo.20274435

    work page doi:10.5281/zenodo.20274435 2026

  60. [64]

    Applied Mathematics Letters96, 216–222 (2019) https://doi.org/10.1016/j.aml

    Greif, C., Urban, K.: Decay of the Kolmogorov N-width for wave problems. Applied Mathematics Letters96, 216–222 (2019) https://doi.org/10.1016/j.aml. 2019.05.013

    work page doi:10.1016/j.aml 2019

  61. [65]

    Arbes, F., Greif, C., Urban, K.: The Kolmogorov N-width for linear trans- port: Exact representation and the influence of the data. arXiv preprint arXiv:2305.00066 (2023) https://doi.org/10.48550/arXiv.2305.00066 38 A Proofs A.1 Proof of Theorem 3.2 Proof.The proof follows a standard discrete energy argument; see, e.g., [53, 54]. For brevity, we introduce...

    work page doi:10.48550/arxiv.2305.00066 2023