Pith. sign in

REVIEW 2 major objections 4 minor 2 cited by

The extended adjoint state and nonlinearity in correlation-based passive imaging

T0 review · 2 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read The paper derives an explicit adjoint backpropagator for correlation-based passive imaging that halves the number of PDE solves, and proves a tangential-cone-like nonlinearity bound.

desk verdict Correct adjoint derivation and a partial TCC bound, but the headline solve-count reduction and 'any elliptic PDE' claim outrun what the paper proves. read the letter →

arxiv 2504.16797 v3 pith:LQON62YE submitted 2025-04-23 math.NA cs.NA

Tram Thi Ngoc Nguyen This is my paper · ORCID
classification math.NAcs.NA MSC 65M3265J2235R30
keywords passiveimagingcovariancemeasurementsextendedadjointstatemethodtangentialconeconditionellipticPDEiterativeregularizationparameteridentification
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

Correlation-based passive imaging infers an unknown medium from the covariance of ambient-noise records, but the covariance measurement squares both the data dimension and the nonlinearity. This paper establishes that, for affine-bilinear elliptic PDEs, the adjoint of the linearized parameter-to-covariance map has an explicit form in terms of an extended adjoint state defined on the doubled domain, and that this form halves the number of PDE solves needed in each gradient step. It also proves a linearization-error bound with the shape of the tangential cone condition, the standard nonlinearity condition that guarantees convergence of iterative regularization methods. If both claims hold, gradient-based reconstruction of media from noise correlations becomes markedly cheaper and has a clear path toward convergence guarantees.

What carries the argument

The extended adjoint state is the mechanism: it is a decomposable function in the Sobolev-Bochner space $L^2(\Omega;W)$ satisfying $D(\theta)^\star\Psi=y$ for almost every second variable, so computing it amounts to solving the standard adjoint PDE column-wise on the squared domain. The affine-bilinear structure $B(\theta,u)=Ku+B(u)\theta$ carries the argument, because linearity of $B$ in $u$ lets the J-average commute with $B$ and converts the covariance data into the operator $B_1^{\mathrm{cov}}(u,x')$ in the adjoint formula. This machinery turns a single nonlinear covariance adjoint into an integral of standard adjoint solves, which is what makes the computation explicit and parallelizable.

What would settle it

Compare the adjoint formula (4.3) against a finite-difference gradient of $\frac{1}{2}\|F(\theta)-y^\delta\|^2$ for the a,b,c problem on a fine grid; disagreement would disprove Theorem 3.5. A second test uses a state-nonlinear PDE, e.g. $-\Delta u+\theta u^3=f$, where the formula should fail exactly at the step that commutes $B$ with the J-average, confirming that affine-bilinearity is the boundary of the method. A code-level count of PDE solves in one Landweber step can check the halving claim directly.

Watch

Extended reading notes

Core claim

The central object is the extended adjoint state $\Psi\in L^2(\Omega;W)$, obtained by solving the adjoint PDE $D(\theta)^\star\Psi=y$ slice-wise on $\Omega\times\Omega$ for covariance data $y$. Theorem 3.5 states that the Hilbert space adjoint of the linearized parameter-to-covariance map is $F'(\theta)^*y = -2 I_X \Re \int_\Omega B_1^{\mathrm{cov}}(u,x')^\star \Psi(\cdot,x')\,dx'$, where $B_1^{\mathrm{cov}}$ is $B$ evaluated at the covariance-weighted state. The proof's key step is pulling $B$ through the J-average, which is exactly where the affine-bilinear form $B(\theta,u)=Ku+B(u)\theta$ is used; state-nonlinear PDEs are not covered. Under lifted regularity, Theorem 5.7 bounds the linearization error by $c_M\|\theta-\tilde\theta\|_{\widehat{X}}\|F(\theta)-F(\tilde\theta)\|_Y + \widehat{N}\|\theta-\tilde\theta\|_X\|\theta-\tilde\theta\|_{\widehat{X}}$, a tangential-cone-condition-like estimate that reduces to the TCC when the cross-covariance term vanishes.

Load-bearing premise

The load-bearing premise is affine-bilinearity: the PDE must take the form $B(\theta,u)=Ku+B(u)\theta$ with $B(u)$ linear in $u$, because the proof of Theorem 3.5 pulls $B$ through the J-average; for PDEs nonlinear in the state the adjoint formula collapses.

Editorial extensions

If this is right

  • Any gradient-based iterative regularization method for passive imaging can use Formula (3.11) with one block of adjoint PDE solves per residual, halving the PDE count relative to backpropagating through the covariance representation (2.14).
  • The extended adjoint solves are independent across slices, so the backpropagation step parallelizes across data columns.
  • Explicit backpropagators for the a,b,c elliptic model and the bi-Helmholtz model follow by inserting the adjoints of $D(\theta)$ and $B(u)$ into the general formula.
  • If the cross-covariance term $\mathrm{Cov}_J(u-v,u)$ vanishes, the linearization-error bound yields the tangential cone condition directly.
  • The TCC-like bound is the first step toward proving convergence of Landweber-type iteration for correlation-based passive imaging.

Reading between the lines

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

  • The same derivation should transfer to affine-bilinear time-harmonic wave equations and other linear non-elliptic settings whenever $D(\theta)$ is boundedly invertible, since the proof uses only that structure.
  • The halving claim suggests a practical scaling rule: for sensor arrays, backpropagation cost becomes comparable to one forward solve rather than two, altering the tradeoff in large-scale helioseismic inversions.
  • A natural test is to verify numerically that the formula agrees with finite-difference gradients for a non-elliptic model; failure would pinpoint exactly where the affine-bilinear assumption is load-bearing.
  • For state-nonlinear PDEs, where the covariance representation (2.14) no longer holds, the covariance adjoint may still be approximated by solving the linearized and adjoint equations around the current state; the paper leaves this open.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 4 minor

Summary. The paper develops a backpropagation framework for correlation-based passive imaging, where the observable is the sample covariance Cov_J(u,u) of the solution of an elliptic PDE depending on an unknown parameter. The forward map is F = Cov_J ∘ S, with S the parameter-to-state map for an affine-bilinear PDE model B(θ,u)=Ku+B(u)θ. The central theoretical result, Theorem 3.5, gives an explicit formula for the adjoint F'(θ)^*: the backpropagation is performed by solving an 'extended adjoint PDE' on the product domain Ω×Ω, slice by slice, and then applying a covariance-weighted adjoint of B. The formula is worked out in detail for the a,b,c-problem and for a bi-Helmholtz example. The second main contribution is Theorem 5.7, which derives a two-term bound on the linearization error of F under a bounded-inverse assumption and a lifted-regularity assumption; the bound has a structure resembling the tangential cone condition, though it does not by itself establish the TCC. The paper also claims in the abstract that the approach halves the number of required PDE solves.

Significance. If the claims are fully supported, the paper makes a useful contribution to PDE-constrained inverse problems with covariance data. The explicit adjoint formula gives a systematic and apparently novel way to backpropagate covariance residuals, avoiding the need to form the derivative of the covariance operator explicitly, and the a,b,c and bi-Helmholtz examples show that the computation is transparent in practice. The nonlinearity analysis is also valuable: Theorem 5.7 is derived from first principles, with assumptions stated explicitly, and it is a concrete step toward convergence guarantees for iterative regularization in passive imaging. The paper is careful to mark the gap between its linearization bound and the full tangential cone condition, which is a strength. The main weaknesses are that the headline efficiency claim ('halves the number of PDE solves') is not backed by an operation count, and the abstract's universality wording overstates the affine-bilinear scope of the derivation.

major comments (2)
  1. [Section 3.2, Remark 7(ii) and Algorithm 3.1] The claim that the extended adjoint state 'halves the number of required PDE solves' is not supported by a rigorous operation count, and as stated it appears to compare only the backpropagation block rather than the full Landweber iteration. In Algorithm 3.1, step S.1 evaluates F(θ_k)=Cov_J(u,u) via (2.14), which for covariance-only data requires applying D(θ_k)^{-1} and (D(θ_k)^{-1})^* to the source covariance—already two blocks of PDE solves—while step S.2 adds N slice-wise adjoint solves for Ψ. Direct backpropagation through (2.14) may reuse quantities from S.1, so the total number of solves depends on the rank J, the discretization dimension N, and the implementation; in a low-rank/no-sample regime with J≪N the direct approach can be cheaper than the proposed one. Since the abstract calls the factor 2 'crucial,' the practical motivation needs either a quantitative comparison for the full iteration or a substantially qualified statement.
  2. [Abstract and Section 1, in view of Eq. (2.2) and Discussion 4] The abstract's phrase 'suitable for any elliptic PDE' overstates the scope of the derivation. Theorem 3.5 is proved under the affine-bilinear structure B(θ,u)=Ku+B(u)θ with B(u) linear in u, and Discussion 4 explicitly states that nonlinearity in the state is a nontrivial obstacle for which the argument in Theorem 3.5 is not available. The introduction's phrasing 'any linear elliptic PDE' is closer to the actual hypotheses, but the abstract should be revised so that the universality claim matches the affine-bilinear framework treated in the paper.
minor comments (4)
  1. [Abstract and Section 1] There are typographical errors in the abstract ('This articles investigates') and in the outline ('the extend adjoint state'); these should be corrected.
  2. [Section 5.2 heading] The heading 'TTC-like linearization error' uses 'TTC' instead of 'TCC'; the abbreviation is defined earlier as tangential cone condition.
  3. [Remark 4] The remark refers to 'the discretization dimension N of Ω' without defining N; a sentence clarifying that N is the number of slice points used in the discretization would avoid ambiguity.
  4. [Section 4, Eq. (4.5)] In the displayed formula for Bcov_1(u)^*Ψ there is a parenthesis mismatch ('Bcov_1(u))⋆Ψ'); the notation should be made uniform with Theorem 3.5.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the adjoint derivation and TCC-like bound are self-contained and do not reduce to their inputs.

full rationale

The paper's central results are derived from first principles rather than imported from a fitted value or a self-citation. Theorem 3.5 computes the adjoint of the linearized parameter-to-covariance map by differentiating the composition F = Cov_J ∘ S, pairing with covariance data, introducing the extended adjoint state Ψ through the PDE D(θ)*Ψ = y, and using only the assumed affine-bilinear structure B(θ,u) = Ku + B(u)θ together with the stated invertibility of D(θ). The proof does not assume the formula it derives, and every intermediate equality is justified by the displayed chain rule, the Hermitian symmetry of covariance data, and the definition of B_cov^1. Lemma 3.4 constructs Ψ as a finite sum of standard adjoint states, so the extended adjoint state is not defined in terms of the target adjoint F'(θ)*. Similarly, the nonlinearity analysis in Theorem 5.7 proves a genuine linearization-error bound under the explicitly stated Assumptions 5.1 and 5.5; it does not assume the tangential cone condition, and the paper explicitly says that fully establishing the TCC and a convergence guarantee is future work. There are no fitted parameters renamed as predictions, and the self-citations (e.g., [26, 35, 45, 46, 47]) are contextual or provide general frameworks rather than carrying the proof of Theorem 3.5 or Theorem 5.7. The skeptical concern about the unquantified claim that the extended adjoint state 'halves the number of required PDE solves' is a substantive evidence gap in the computational comparison, not a circularity: it concerns whether an advertised efficiency advantage is fully established, not whether the derivation assumes its conclusion. The stated limitations, including the reliance on state-linearity in the fourth equality of Theorem 3.5's proof and the overstatement of 'any elliptic PDE' relative to Discussion 4, are discussed openly and do not make the argument circular.

Assumptions & free parameters 0 free parameters · 7 assumptions · 1 invented entities

The paper introduces no fitted numerical parameters. Its results rest on standard functional-analytic tools and on explicitly stated domain assumptions (affine bilinearity, bounded invertibility, lifted regularity, known source covariance). The only new conceptual object is the extended adjoint state, which is constructively defined and carries no independent physical postulate.

assumptions (7)
  • domain assumption D(θ) is boundedly invertible with ||D(θ)^{-1}|| ≤ π(||θ||_X) (Assumption 5.1)
    Used throughout the nonlinearity analysis to bound state differences and to establish Lipschitz properties in Lemma 5.2.
  • domain assumption The model is affine bilinear: B(θ,u) = K u + B(u)θ with B(u) linear in u (Eq. 2.2)
    Central to Theorem 3.5; the covariance adjoint formula and the definition of B^cov_1 rely on linearity in u.
  • domain assumption Data belongs to the covariance cone Ycov or the symmetric decomposable subspace Ysep,sym (Definition 3.3, Remark 6)
    The extended adjoint state solves the adjoint PDE slice-wise only for decomposable and symmetric data.
  • domain assumption Lifted regularity: D(θ)^{-1} ∈ L(L2(Ω), Û) with bound (5.8) (Assumption 5.5)
    Needed for the TCC-like linearization error bound in the L2 data norm; the paper gives one example but no general verification.
  • standard math Implicit function theorem and standard Fréchet differentiability
    Used in Lemma 3.1 to derive the linearized state equation.
  • standard math Gelfand triple embeddings and Hilbert-Schmidt composition rules
    Used throughout Sections 2 and 3 to transfer dualities and norms.
  • domain assumption Source covariance CovJ[f,f] is known and lies in HS(W,W*)
    The forward covariance representation (2.14) and Lemma 5.4 depend on this.
invented entities (1)
  • Extended adjoint state Ψ independent evidence
    purpose: A function on Ω×Ω solving D(θ)⋆ Ψ = y for each slice, used to express the covariance adjoint compactly and reduce backpropagation to one block of adjoint PDE solves.
    It is explicitly constructed in Lemma 3.4 as Σ ψ_i q_i, where each ψ_i solves a standard adjoint PDE; existence follows from bounded invertibility of D(θ)⋆.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The extended adjoint state and nonlinearity in correlation-based passive imaging." pith.science (2026). https://pith.science/paper/LQON62YE

@misc{pith2026250416797,
  author       = {Pith},
  title        = {Pith review of: The extended adjoint state and nonlinearity in correlation-based passive imaging},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LQON62YE}},
  note         = {Machine review of arXiv:2504.16797}
}
read the original abstract

This articles investigates physics-based passive imaging problem, wherein one infers an unknown medium using ambient noise and correlation of the noise signal. We develop a general backpropagation framework via the so-called extended adjoint state, suitable for any elliptic PDE; crucially, this approach reduces by half the number of required PDE solves. Applications to several different PDE models demonstrate the universality of our method. In addition, we analyze the nonlinearity of the correlated model, revealing a surprising tangential cone condition-like structure, thereby advancing the state of the art towards a convergence guarantee for regularized reconstruction in passive imaging.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Full cross-correlation inversion for quantitative passive imaging with time-harmonic acoustic waves

    math.NA 2026-07 conditional novelty 6.0 of 10

    Expected cross-correlations of ambient acoustic signals can be inverted quantitatively for medium parameters via an adjoint-state HDG scheme, recovering structures comparable to active-source FWI on synthetic data.

  2. Data assimilation via model reference adaptation for linear and nonlinear dynamical systems

    math.OC 2026-02 conditional novelty 5.0 of 10

    A model reference adaptive system recovers unknown coefficients in nonlinear parabolic PDEs from time-series data in four benchmark tests.

Reference graph

Works this paper leans on

60 extracted references · 44 canonical work pages · cited by 2 Pith papers

  1. [1]

    Aarset, Global optimality conditions for sensor placement, with extensions to binary A-optimal experi- mental designs, Inverse Problems, 41 (2024), p

    C. Aarset, Global optimality conditions for sensor placement, with extensions to binary A-optimal experi- mental designs, Inverse Problems, 41 (2024), p. 065013

    work page 2024

  2. [2]

    Aarset, M

    C. Aarset, M. Holler, and T. T. N. Nguyen , Learning-informed parameter identification in nonlinear time-dependent PDEs, Applied Mathematics and Optimization, 88 (2023), p. 53 pp, https://doi.org/10. 1007/s00245-023-10044-y

    work page 2023

  3. [3]

    Aerts, J

    C. Aerts, J. Christensen-Dalsgaard, and D. Kurtz , Asteroseismology, Springer Dordrecht Heidelberg London New York, 2010, https://doi.org/10.1007/978-1-4020-5803-5

    doi:10.1007/978-1-4020-5803-5 2010

  4. [4]

    H. T. Banks and K. Kunisch , Estimation Techniques for Distributed Parameter Systems , Birkhauser, Boston, Basel, Berlin, 1989

    work page 1989

  5. [5]

    Borcea, Electrical impedance tomography, Inverse Problems, 18 (2002), p

    L. Borcea, Electrical impedance tomography, Inverse Problems, 18 (2002), p. R99, https://doi.org/10.1088/ 0266-5611/18/6/201, https://dx.doi.org/10.1088/0266-5611/18/6/201

    doi:10.1088/0266-5611/18/6/201 2002

  6. [6]

    Borcea, T

    L. Borcea, T. Callaghan, and G. Papanicolaou, Synthetic aperture radar imaging with motion estimation and autofocus, Inverse Problems, 28 (2012), p. 045006, https://doi.org/10.1088/0266-5611/28/4/045006, https://dx.doi.org/10.1088/0266-5611/28/4/045006

    work page doi:10.1088/0266-5611/28/4/045006 2012

  7. [7]

    Borcea, G

    L. Borcea, G. Papanicolaou, and C. Tsogka, Theory and applications of time reversal and interferometric imaging, Inverse Problems, 19 (2003), p. S139, https://doi.org/10.1088/0266-5611/19/6/058, https://dx. doi.org/10.1088/0266-5611/19/6/058

    work page doi:10.1088/0266-5611/19/6/058 2003

  8. [9]

    Brenguier, D

    L. Brenguier, D. Clarke, Y. Aoki, N. M. Shapiro, M. Campillo, and V. rie Ferrazzini , Monitoring volcanoes using seismic noise correlations , 343 (2011), pp. 633–638, https://doi.org/10.1016/j.crte.2010. 12.010

    work page doi:10.1016/j.crte.2010 2011

Show all 60 references
  1. [10]

    Chadan, D

    K. Chadan, D. Colton, L. P ¨aiv¨arinta, and W. Lassi , An Introduction to Inverse Scattering and Inverse Spectral Problems, Society for Industrial and Applied Mathematics, 1997, https://doi.org/10.1137/1. 9780898719710, https://epubs.siam.org/doi/abs/10.1137/1.9780898719710, h...

    1997 doi

  2. [11]

    J. F. Claerbout , Synthesis of a layered medium from its acoustic transmission response , Geophysics, 33 (1968), pp. 264–269, https://doi.org/10.1190/1.1439927

    1968 doi

  3. [12]

    Divitt, D

    S. Divitt, D. F. Gardner, and A. T. Watnik , Imaging around corners in the mid-infrared using speckle correlations, 28 (2020), pp. 11051–11064, https://doi.org/10.1364/OE.388260

    2020 doi

  4. [13]

    Draganov, X

    D. Draganov, X. Campman, J. Thorbecke, A. Verdel, and K. Wapenaar , Seismic exploration-scale velocities and structure from ambient seismic noise (¿1 Hz) , 118 (2013), pp. 4345–60, https://doi.org/0. 1002/jgrb.50339

    2013

  5. [14]

    Dryden, A

    I. Dryden, A. Koloydenko, and D. Zhou , A non-euclidean statistics for covariance matrices, with appli- cations to diffusion tensor imaging , N. Annals of Applied Statistics, 3 (2009), pp. 1102–1123

    2009

  6. [16]

    Eller, R

    M. Eller, R. Griesmaier, and A. Rieder , Tangential Cone Condition for the Full Waveform Forward Operator in the Viscoelastic Regime: The Nonlocal Case , SIAM Journal on Applied Mathematics, 84 (2024), pp. 412–432, https://doi.org/10.1137/23M1551845, https://doi.org/10.1137/23...

    2024 doi

  7. [17]

    H. Engl, M. Hanke, and A. Neubauer , Regularization of Inverse Problems , Kluwer, Dordrecht, 1996

    1996

  8. [18]

    L. C. Evans , Partial Differential Equations , Graduate Studies in Mathematics 19. AMS, Providence, RI, 1998

    1998

  9. [19]

    Fournier, L

    D. Fournier, L. Gizon, and L. Hyest , Viscous inertial modes on a differentially rotating sphere: Compar- ison with solar observations , Astronomy & Astrophysics, 664 (2022), A6, p. 16 pp, https://doi.org/10. 1051/0004-6361/202243473

    2022

  10. [20]

    Garnier and G

    J. Garnier and G. Papanicolaou, Passive Imaging with Ambient Noise , Cambridge University Press, 2016

    2016

  11. [21]

    Gizon and A

    L. Gizon and A. C. Birch , Local helioseismology, Living Reviews in Solar Physics, 2 (2005), 6, p. 131 pp, https://doi.org/10.12942/lrsp-2005-6

    2005 doi

  12. [22]

    Haddar and P

    H. Haddar and P. Monk , The linear sampling method for solving the electromagnetic inverse medium problem, Inverse Problems, 18 (2002), p. 891, https://doi.org/10.1088/0266-5611/18/3/323, https://dx. doi.org/10.1088/0266-5611/18/3/323

    2002 doi

  13. [23]

    Halko, P

    N. Halko, P. G. Martinsson, and J. A. Tropp, Finding structure with randomness: Probabilistic algorithms for constructing approximate matrix decompositions , SIAM Review, 53 (2011), pp. 217–288

    2011

  14. [24]

    Harrach, An Introduction to Finite Element Methods for Inverse Coefficient Problems in Elliptic PDEs , Jahresbericht der Deutschen Mathematiker-Vereinigung, 123 (2021), pp

    B. Harrach, An Introduction to Finite Element Methods for Inverse Coefficient Problems in Elliptic PDEs , Jahresbericht der Deutschen Mathematiker-Vereinigung, 123 (2021), pp. 183–210, https://doi.org/0.1365/ s13291-021-00236-2

    2021

  15. [25]

    Helin, M

    T. Helin, M. Lassas, L. Oksanen, and T. Saksala , Correlation based passive imaging with a white noise source , Journal de Math´ ematiques Pures et Appliqu´ ees, 116 (2018), pp. 132–160, https://doi.org/https://doi.org/10.1016/j.matpur.2018.05.001, https://www.sciencedirect.co...

    2018 doi

  16. [26]

    Hoffmann, A

    H. Hoffmann, A. Wald, and T. T. N. Nguyen , Parameter identification for elliptic boundary value prob- lems: an abstract framework and application , Inverse Problems, 38 (2022), p. 44 pp, https://doi.org/10. 1088/1361-6420/ac6d02. This manuscript is for review purposes only. 22

    2022

  17. [27]

    Hubmer, E

    S. Hubmer, E. Sherina, A. Neubauer, and O. Scherzer, Lam´ e Parameter Estimation from Static Displace- ment Field Measurements in the Framework of Nonlinear Inverse Problems , SIAM Journal on Imaging Sciences, 11 (2018), https://doi.org/10.1137/17M1154461

    2018 doi

  18. [28]

    Isakov, Inverse Problems for Partial Differential Equations , Springer Cham, 2017, https://doi.org/10

    V. Isakov, Inverse Problems for Partial Differential Equations , Springer Cham, 2017, https://doi.org/10. 1007/978-3-319-51658-5

    2017

  19. [29]

    Jiang, G

    Y. Jiang, G. Nakamura, and K. Shirota, Levenberg–Marquardt method for solving inverse problem of MRE based on the modified stationary Stokes system , Inverse Problems, 37 (2021), 125013

    2021

  20. [30]

    S. Jung, A. Schwartzman, and D. Groisser , Scaling-rotation distance and interpolation of symmetric positive-definite matrices, SIAM Journal on Matrix Analysis and Applications, 36 (2015), pp. 1180–1201

    2015

  21. [31]

    Kaltenbacher, Some Newton-type methods for the regularization of nonlinear ill-posed problems , Inverse Problems, 13 (1997), pp

    B. Kaltenbacher, Some Newton-type methods for the regularization of nonlinear ill-posed problems , Inverse Problems, 13 (1997), pp. 729–753, https://doi.org/10.1088/0266-5611/13/3/012

    1997 doi

  22. [32]

    Kaltenbacher, All-at-once versus reduced iterative methods for time dependent inverse problems , Inverse Problems, 33 (2017)

    B. Kaltenbacher, All-at-once versus reduced iterative methods for time dependent inverse problems , Inverse Problems, 33 (2017)

    2017

  23. [33]

    Kaltenbacher , Convergence rates under a range invariance condition with application to electrical impedance tomography, IMA Journal of Numerical Analysis, (2024), p

    B. Kaltenbacher , Convergence rates under a range invariance condition with application to electrical impedance tomography, IMA Journal of Numerical Analysis, (2024), p. drae063

    2024

  24. [34]

    Kaltenbacher, A

    B. Kaltenbacher, A. Neubauer, and O. Scherzer , Iterative Regularization Methods for Nonlinear Prob- lems, de Gruyter, Berlin, New York, 2008. Radon Series on Computational and Applied Mathematics

    2008

  25. [35]

    Kaltenbacher, T

    B. Kaltenbacher, T. T. N. Nguyen, and O. Scherzer , The tangential cone condition for some coefficient identification model problems in parabolic PDEs , Springer, Switzerland, 2021, https://doi.org/10.1007/ 10.1007/978-3-030-57784-1 5

    2021 doi

  26. [36]

    Kindermann, Convergence of the gradient method for ill-posed problems , Inverse Problems & Imaging, 11 (2017), pp

    S. Kindermann, Convergence of the gradient method for ill-posed problems , Inverse Problems & Imaging, 11 (2017), pp. 703–720

    2017

  27. [37]

    Kindermann, On the tangential cone condition for electrical impedance tomography, Electronic Transaction on Numerical Analysis, 57 (2022), pp

    S. Kindermann, On the tangential cone condition for electrical impedance tomography, Electronic Transaction on Numerical Analysis, 57 (2022), pp. 17–34

    2022

  28. [38]

    Kirsch, An Introduction to the Mathematical Theory of Inverse Problems , Springer, New York, NY, 2011, https://doi.org/10.1007/978-1-4419-8474-6

    A. Kirsch, An Introduction to the Mathematical Theory of Inverse Problems , Springer, New York, NY, 2011, https://doi.org/10.1007/978-1-4419-8474-6

    2011 doi

  29. [39]

    O. A. Ladyzhenskaya and N. N. Ural’tseva , Linear and Quasilinear Equations of Elliptic Type , Izd. Nauka, Moscow, 1964. (Engl. Transl.: Acad. Press, New York, 1968.)

    1964

  30. [40]

    Lazar, G

    M. Lazar, G. A. Maugin, and E. C. Aifantis , On a theory of nonlocal elasticity of bi-helmholtz type and some applications , International Journal of Solids and Structures, 43 (2006), pp. 1404–1421, https: //doi.org/https://doi.org/10.1016/j.ijsolstr.2005.04.027, https://www.s...

    2006 doi

  31. [41]

    Lotfinia, N

    H. Lotfinia, N. Chegini, and R. Mokhtari, The bi-helmholtz equation with cauchy conditions: ill-posedness and regularization methods, Inverse Problems in Science and Engineering, 29 (2021), pp. 17–39, https:// doi.org/10.1080/17415977.2020.1764950, https://doi.org/10.1080/1741...

    2021

  32. [42]

    Malthus, An Essay on the Principles of Population , J

    T. Malthus, An Essay on the Principles of Population , J. Johnson, London, 1798

  33. [43]

    R. J. Muirhead , Aspects of multivariate statistical theory , Wiley series in probability and mathematical statistics, Wiley, New York, N.Y. [u.a.], 1982. Literaturverz. S. 650-661

    1982

  34. [44]

    M¨uller, T

    B. M¨uller, T. Hohage, D. Fournier, and L. Gizon , Quantitative passive imaging by iterative holography: The example of helioseismic holography , 40(4) (2024), https://doi.org/10.1088/1361-6420/ad2b9a

    2024 doi

  35. [45]

    T. T. N. Nguyen , Landweber–Kaczmarz for parameter identification in time-dependent inverse problems: all-at-once versus reduced version , Inverse Problems, 35 (2019), p. 035009, https://doi.org/10.1088/ 1361-6420/aaf9ba

    2019

  36. [46]

    T. T. N. Nguyen, Bi-level iterative regularization for inverse problems in nonlinear PDEs , Inverse Problems, 40 (2024), p. 045020, https://doi.org/10.1088/1361-6420/ad2905

    2024 doi

  37. [47]

    T. T. N. Nguyen, Sequential bi-level regularized inversion with application to hidden reaction law discovery , Inverse Problems, 41 (2025), p. 065015, https://doi.org/10.1088/1361-6420/addf73

    2025 doi

  38. [48]

    T. T. N. Nguyen, D. Fournier, L. Gizon, and T. Hohage, Linear toroidal-inertial waves on a differentially rotating sphere with application to helioseismology: Modeling, forward and inverse problems , ArXiv: 2507.20488 [math.AP], (2025), p. 34 pp

    2025 arXiv

  39. [49]

    T. T. N. Nguyen, T. Hohage, D. Fournier, and L. Gizon, Inferring solar differential rotation and viscosity via passive imaging with inertial waves , IProc. of Math. and Num. Aspects of Wave Propagation, (2024), pp. 215–216, https://doi.org/10.17617/3.MBE4AA

    2024 doi

  40. [50]

    M. R and V. D , Einf¨ uhrung in die Funktionalanalysis, Vieweg+Teubner Verlag, 1992

    1992

  41. [51]

    Raumer, C

    H.-G. Raumer, C. Spehr, T. Hohage, and D. Erns , Weighted data spaces for correlation-based array imaging in experimental aeroacoustics, JJournal of Sound and Vibration, 494 (2020), https://doi.org/10. 1016/j.jsv.2020.115878

    2020

  42. [52]

    Roub´ıˇcek, Nonlinear Partial Differential Equations with Applications , Springer, Basel, 2013

    T. Roub´ıˇcek, Nonlinear Partial Differential Equations with Applications , Springer, Basel, 2013

    2013

  43. [53]

    K. G. Sabra, P. Gerstoft, P. Roux, W. A. Kuperman, and M. C. Fehler , Extracting time-domain Green’s function estimates from ambient seismic noise , Geophysical Research Letters, 32 (2005), https: //doi.org/10.1029/2004gl021862, https://hal.science/hal-04004281

    2005 doi

  44. [54]

    K. G. Sabra, A. Srivastava, F. L. di Scalea, I. Bartoli, P. Rizzo, and S. Conti , Structural health monitoring by extraction of coherent guided waves from diffuse fields , 123 (2008), https://doi.org/10. 1121/1.2820800

    2008

  45. [55]

    Sarnighausen, T

    G. Sarnighausen, T. T. N. Nguyen, T. Hohage, M. Sinha, S. K ¨oster, T. Betz, U. S. Schwarz, and A. Wald, Traction force microscopy for linear and nonlinear elastic materials as a parameter identifi- This manuscript is for review purposes only. CORRELATION-BASED PASSIVE IMAGING...

    2024 arXiv

  46. [56]

    Scherzer, On Convergence Criteria of Iterative Methods Based on Landweber Iteration for Solving Non- linear Problems, Journal of Mathematical Analysis and Applications, 194 (1995)

    O. Scherzer, On Convergence Criteria of Iterative Methods Based on Landweber Iteration for Solving Non- linear Problems, Journal of Mathematical Analysis and Applications, 194 (1995)

    1995

  47. [57]

    Scherzer, B

    O. Scherzer, B. Hofmann, and Z. Nashed , Newton’s methods for solving linear inverse problems with neural network coders, Sampling Theory, Signal Processing, and Data Analysis, 21 (2023)

    2023

  48. [58]

    Shapiro and M

    N. Shapiro and M. Campillo , Emergence of broadband rayleigh waves from correlations of the ambient seismic noise, Geophys. Res. Lett., 31 (2004), https://doi.org/10.1038/362430a0

    2004 doi

  49. [59]

    S. A. Socolofsky and G. H. Jirka , Advective Diffusion Equation, Lecture notes, 2010

    2010

  50. [60]

    Tikhonov and V

    A. Tikhonov and V. Arsenin. , Solutions of Ill-posed Problems , V. H. Winston & Sons, Washington, D.C., New York, 1977

    1977

  51. [61]

    Tr¨oltzsch, Optimal Control of Partial Differential Equations: Theory, Methods and Applications , Amer- ican Mathematical Society, Providence, Rhode Island, 2010

    F. Tr¨oltzsch, Optimal Control of Partial Differential Equations: Theory, Methods and Applications , Amer- ican Mathematical Society, Providence, Rhode Island, 2010

    2010

  52. [62]

    D. Yang, L. Gizon, and H. Barucq, Imaging individual active regions on the sun’s far side with improved he- lioseismic holography, Astronomy & Astrophysics, 669 (2023), p. 14 pp, https://doi.org/10.1086/324323, https://dx.doi.org/10.1086/324323. This manuscript is for review p...

    2023 doi

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.