Pith. sign in

REVIEW 3 major objections 5 minor 28 references

Spectral densities from Euclidean-time lattice correlation functions

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Euclidean-time lattice correlation functions can be inverted analytically: the paper derives explicit coefficients $g_\alpha(t|\omega)$ such that the spectral density $\rho(\omega)$ is obtained as the small-regulator limit of an integral…

desk verdict A clean continuum derivation with a genuinely new but unverified discrete-lattice construction; the exactness claim on the lattice outruns the evidence in this proceedings paper. read the letter →

arxiv 2501.16527 v1 pith:RDH3KHQL submitted 2025-01-27 hep-lat

Matteo Saccardi , Mattia Bruno , Leonardo Giusti This is my paper · ORCID
classification hep-lat MSC 44A1045C0581T25 PACS 12.38.Gc11.15.Ha
keywords spectraldensityinverseLaplacetransformlatticeQCDEuclideancorrelationfunctionsCarlemanoperatorMellinbasisTikhonovregularizationHilbertmatrix
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper aims to remove the ill-posedness from a central lattice-QCD task: turning the Euclidean-time correlation function $C(t)=\int \rho(\omega)e^{-\omega t}\,d\omega$ into the spectral density $\rho(\omega)$. It proposes analytic formulae, in the continuum and on an infinite lattice, that express the regulated reconstruction $\rho_\alpha(\omega)$ as an integral (or lattice sum) of $C(t)$ against coefficients $g_\alpha(t|\omega)$ that are known in closed form. The exact $\rho$ is the limit $\alpha\to0$ taken after the coefficients are integrated; for the discrete case the paper claims that the smeared delta approaches the Dirac delta even at fixed lattice spacing, so the inversion procedure itself introduces no discretization error. If this holds, spectral densities and their smeared versions become directly computable with controlled bias, a step that would matter for observables such as scattering amplitudes, transport coefficients, and the hadronic vacuum polarization contribution to the muon anomalous moment.

What carries the argument

The machinery is the spectral decomposition of the Laplace-transform kernel: the Carleman operator $H(\omega,\omega')=1/(\omega+\omega')$ in the continuum, and the infinite Hilbert matrix with entries $a/(t+t'+2a)$ in the discrete case. The continuum eigenfunctions are the Mellin basis $u_s(x)=e^{i s\log x}/\sqrt{2\pi x}$; the discrete eigenvectors $v_s(n,a)$ come from the singly-infinite Hilbert matrix. The identity carrying the argument is $\rho(\omega)=\lim_{\alpha\to0}\int dt\, g_\alpha(t|\omega)C(t)$, where $g_\alpha$ is built from the regularized inverse $(H+\alpha I)^{-1}$ in the continuum and from $(A_a+\alpha I)^{-1}$ on the lattice. These coefficient functions convert the ill-posed inversion into a well-posed integral against data, provided the regulator is removed only after the integral is evaluated.

What would settle it

Take a synthetic spectral density with a known analytic form, generate its correlator on a finite lattice with realistic noise, apply the discrete reconstruction in Eq. (21) for decreasing values of the smoothing parameter at fixed time extent and fixed signal-to-noise ratio, and plot the reconstruction error: if the noise-induced variance grows faster than the bias falls as the parameter is reduced, the exact limit cannot be approached on finite data.

Watch

Extended reading notes

Core claim

The central discovery, stated on the paper's own terms, is that the Carleman operator $H(\omega,\omega')=1/(\omega+\omega')$, which arises from applying the Laplace transform twice, can be diagonalized exactly by the Mellin basis $u_s(x)=e^{i s\log x}/\sqrt{2\pi x}$, with eigenvalues $\lambda_s=\Gamma(1/2+i s)$ of squared modulus $\pi/\cosh(\pi s)$. Tikhonov regularization $H_\alpha=H+\alpha I$ turns the inverse problem into a one-parameter family of well-posed linear functionals, and the contraction $H H_\alpha^{-1}$ gives an explicit smeared delta $\delta_\alpha(\omega,\omega')$ that approaches $\delta(\omega-\omega')$ as $\alpha\to0$. The reconstruction formula $\rho(\omega)=\lim_{\alpha\to0}\int_0^\infty dt\, g_\alpha(t|\omega)C(t)$ follows with $g_\alpha(t|\omega)$ real and computable. In the discrete case the same construction is carried out with the infinite Hilbert matrix $aA_a(t,t')=a/(t+t'+2a)$, whose eigenvectors $v_s(n,a)$ for $s\ge0$ are complete on $\ell^2(\mathbb{Z}_+)$; the paper asserts that the induced $\delta_{a,\alpha}(\omega,\omega')$ converges to the Dirac delta as $\alpha\to0$ even at fixed lattice spacing, so the only remaining discretization error in the reconstruction is the one already present in the sampled correlator $C_a(t)$.

Load-bearing premise

The method works only if the smoothing parameter can be made tiny enough before statistical noise and the finite time extent of the lattice dominate the result, and only if the discrete eigenvectors' completeness and the fixed-lattice-spacing convergence, whose proofs are deferred, hold as stated.

Editorial extensions

If this is right

  • The inverse Laplace transform of a Euclidean correlator becomes a closed-form linear functional of $C(t)$: for each $\omega$, $\rho_\alpha(\omega)$ is a single integral with known kernel, and the exact $\rho$ is the $\alpha\to0$ limit of that integral.
  • Smeared spectral densities $\rho_\kappa$ for kernels relevant to the hadronic vacuum polarization can be computed by integrating the same coefficients against $\kappa$, so the smearing kernel can be chosen without solving a second inverse problem.
  • Subtracted spectral densities, needed when $\rho(\omega)$ grows like $\omega^k$ at large $\omega$, follow from a straightforward generalization that replaces the Mellin eigenvalue $\lambda_s$ with $\lambda_{s,b}=\Gamma(1/2+b+i s)$.
  • On an infinite lattice, the discrete reconstruction uses eigenvectors of the infinite Hilbert matrix; as $\alpha\to0$ its smeared delta becomes the Dirac delta at fixed lattice spacing, so inversion-induced discretization errors vanish and only the correlator's own discretization errors remain.
  • The known finite-$N$ expansion of the spectral density on $N$ exponentials emerges as the large-$N$ limit of the discrete coefficients, giving a principled reading of the truncation parameter in terms of the temporal extent.

Reading between the lines

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

  • The paper does not quantify noise amplification in the $\alpha\to0$ limit; if the variance of the coefficients $g_\alpha$ grows faster than their bias shrinks, real lattice data with finite statistics will force a nonzero optimal $\alpha$, and the advertised exactness becomes a bias-variance trade-off rather than a limit that can be reached.
  • The claimed fixed-$a$ convergence of $\delta_{a,\alpha}$ to a delta suggests the discrete eigenvectors carry the same spectral content as the continuum Mellin basis; a direct numerical check of whether the two reconstructions agree to $O(a^2)$ after the $\alpha$ limit would test the paper's error structure.
  • Because the same squared eigenvalues $|\lambda_s|^2=\pi/\cosh(\pi s)$ regulate both continuum and discrete problems, other regularizers beyond Tikhonov are likely to share the same $\alpha\to0$ limit but differ in finite-$\alpha$ bias and noise; comparing them on a fixed test spectral density would map the practical reach of the method.
  • A natural extension not discussed here is the finite-volume case, where the correlator is a sum over discrete momenta; the same construction with a finite-dimensional Hilbert-like matrix may give closed-form finite-volume corrections.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. This proceedings paper reviews the authors' recent proposal to invert the Laplace transform relating Euclidean-time correlation functions C(t) to spectral densities ρ(ω). In the continuum, it diagonalizes the Carleman operator 1/(ω+ω') using the Mellin basis, obtains Tikhonov-regularized inverse coefficients g_α(t|ω), and extracts ρ(ω) as α→0. For a discrete lattice with spacing a, it replaces the Carleman operator by the infinite Hilbert matrix A_a, uses its eigenvectors to define analogous coefficients g_{a,α}(t|ω), and claims that the smeared delta function δ_{a,α}(ω,ω') tends to δ(ω−ω') as α→0 even at fixed a, so that the only remaining discretization errors come from the correlator. It also treats smeared and subtracted spectral densities and integrated correlators.

Significance. If the discrete spectral-theoretic claims hold, the paper offers a closed-form, parameter-free (apart from the regulator α) inversion of the Laplace transform, a potential major advance for spectral-density extraction in lattice QCD. The continuum derivation is explicit and the algebraic consistency checks (Eqs. (5)-(13)) are convincing. The paper is commendably explicit about the assumptions requiring proof and about the limitations of truncation and finite α, pointing to a comprehensive study as essential. However, the central lattice claim rests on deferred proofs and is supported only by noiseless synthetic tests, leaving a gap between the advertised exactness and demonstrated performance.

major comments (3)
  1. [Section 3, paragraphs following Eq. (19) and Eq. (22)] The discrete inversion (Eq. (21)) depends on the completeness and orthonormality of the eigenvectors v_s(n,a) of the infinite Hilbert matrix A_a in ℓ^2(Z+) and of the eigenfunctions v_s(ω,a) in L^2(R+), and on the fixed-a limit δ_{a,α}(ω,ω')→δ(ω−ω') as α→0. These statements are asserted in the text without proof, and the proof of completeness is deferred to Ref. [1]. Since the abstract's claim of an exact lattice solution relies on these assertions, the manuscript should either state them with precise hypotheses and proofs (or an appendix) or explicitly frame them as conjectures with pointers to the exact theorems in Ref. [1].
  2. [Section 4, first paragraph] The paper concedes that the formalism "necessarily break[s] down for α>0 and when the sums in Eqs. (21) and (29) are truncated to values of t<tmax" and calls for a comprehensive study of these effects. This concession is in tension with the abstract's unqualified promise to "exactly solve" the inversion problem on the lattice. The central claims of Sec. 3 should be restated to make explicit that exactness holds only in the limits α→0 and tmax→∞, and the practical reach of the method for noisy, finite lattice data is not yet established.
  3. [Equations (21)-(23) and Figure 1b] The only numerical support for the fixed-a α→0 limit is a noiseless synthetic reconstruction at two α values. Given the exponential ill-posedness, with eigenvalues |λ_s|^2 = π/cosh(πs), a demonstration of controlled convergence as α→0 at fixed a, together with an assessment of how the coefficients g_{a,α} amplify statistical noise, would be needed to substantiate the claim that "the only remaining source of discretization errors in the α→0 limit would be the correlator." As written, this claim is plausible but not demonstrated.
minor comments (5)
  1. [Section 2, Eq. (10) and following text] The coefficients g_α(t|ω) are said to be "real"; since the Mellin basis functions u_s(t) are complex, it would help to give the explicit real form or note the pairing of s and -s.
  2. [Figure 2b caption] The caption "Relative errors on the computation of discrete coefficients at finite N, see the discussion below Eq. (25), with respect to Eq. (23)" is ambiguous; specify the norm used for the relative error and the values of N and s plotted.
  3. [Section 3, Eq. (24)] The least-square problem is written for an infinite sum over n; the finite-N version used in the numerical tests appears only later in the text. Clarify the truncation and the dependence of the solution on N.
  4. [References] Reference [1] is cited as an arXiv preprint; if a journal or PoS proceedings version is available, citing it would allow the reader to verify the deferred proofs.
  5. [Section 2, Eq. (15)] In the condition k > 5/2 for the light isovector vector current, specify that k is the power of ω in the subtracted representation and that the bound refers to the behavior at large ω; as written it is slightly ambiguous.

Circularity Check

2 steps flagged · score 4.0 of 10

Continuum inversion is independent, but the lattice exact-inversion claim leans on the authors' own Ref. [1] for completeness, fixed-a convergence, and truncation-error suppression.

  1. self citation load bearing [Section 3, after Eq. (20) and before Eq. (21)]
    "Given the eigenvectors 𝑣𝑠(𝑛,𝑎) (𝑠∈ R+) of A 𝑎 [27], one readily finds the eigenfunctions 𝑣𝑠(𝜔,𝑎) of H 𝑎. The sets of these eigenvectors and eigenfunctions are separately complete and orthonormal, acting respectively on 𝐿2(R+) and ℓ2(Z+), the latter being the set of square-summable sequences. It is possible to prove [1] that, in the continuum limit, the eigenfunctions 𝑣𝑠(𝜔,𝑎) approach the correct linear combination of 𝑢𝑠(𝜔) and 𝑢∗𝑠(𝜔)."

    The discrete inversion in Eq. (21) is constructed from these eigenvectors/eigenfunctions. The completeness and orthonormality assertions, which are the mathematical foundation for the discrete exact-inversion claim, are not proved here but deferred to the authors' own Ref. [1] ('It is possible to prove [1]'). The review itself provides no independent derivation or external check (e.g., a proof in the text or a machine-checked theorem), so the central lattice result is load-bearing on a self-citation. This does not make the continuum derivation circular, but it does make the lattice extension's core premise rest on the authors' prior work.

  2. self citation load bearing [Section 3, after Eq. (22), and Section 4]
    "Notice that 𝛿𝑎,𝛼(𝜔,𝜔′) tends to 𝛿(𝜔−𝜔′) in the limit 𝛼→0, even at fixed 𝑎 (for an infinite lattice), implying that the limit 𝜌(𝜔)=lim𝛼→0 𝜌𝑎,𝛼(𝜔) exists. ... While the above statements are valid in the 𝛼→0 limit and for an infinite lattice, they necessarily break down for 𝛼>0 and when the sums in Eqs. (21) and (29) are truncated to values of 𝑡<𝑡max. Although truncation errors are expected to be exponentially suppressed in 𝑡max [1]..."

    The key lattice claim that inversion-induced discretization errors vanish as 𝛼→0 at fixed lattice spacing is asserted without proof in this text and is directly supported only by the authors' Ref. [1]. Likewise, the expectation that finite-time truncation errors are exponentially suppressed is cited to [1] rather than demonstrated here. These statements are load-bearing because they are exactly what turns Eq. (21) from a formal intermediate expression into the advertised exact lattice inversion. The paper does honestly flag the breakdown for 𝛼>0 and finite 𝑡max, so this is not a hidden fit; nevertheless, the practical and theoretical support for the central lattice result is imported from the authors' own companion work.

full rationale

The continuum solution is a genuine analytic inversion: it diagonalizes the Carleman operator with the Mellin basis, uses Tikhonov regularization, and cites external classical mathematics (Carleman, McWhirter and Pike, Epstein and Schotland). There are no fitted parameters and no prediction that reduces to an input by construction; the numerical example uses a known input spectral function to demonstrate convergence, which is a legitimate consistency check rather than circular reasoning. The identified circularity is confined to the discrete/lattice generalization. There, the completeness and orthonormality of the infinite Hilbert matrix eigenvectors, the fixed-a limit 𝛿𝑎,𝛼→𝛿, and the exponential suppression of truncation errors are all deferred to the authors' own Ref. [1]. Because the review provides no independent proof or external verification of these load-bearing statements, the lattice exact-inversion claim is partially dependent on a self-citation chain. However, the paper openly states that the ideal limits break down for 𝛼>0 and truncated sums, and it calls for a comprehensive study of the interplay; it does not disguise the conditionality. On balance, the central continuum logic is independent and the discrete result has substantial independent content in the reduction to classical Hilbert-matrix theory, so the appropriate score is 4: some self-citation is load-bearing, but the core derivation is not equivalent to its inputs.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

No free parameters are fitted to physical data; the numerical part is a synthetic check. The listed parameters are regularization, discretization, truncation, and test-function inputs. No physical entities are invented; the coefficients g_alpha, the smeared deltas delta_alpha, the operators H_a and A_a, and the adapted kernels K_{a,alpha} are defined mathematical objects.

free parameters (4)
  • Tikhonov regulator alpha
    Regularization scale. The method depends on removing it via the alpha to 0 limit, but with noisy finite data the limit is unattainable, and the paper provides no selection criterion.
  • Lattice spacing a
    Discretization scale in the discrete construction. It is a physical input, not fitted, but the fixed-a alpha to 0 limit carries the main discrete-case claim.
  • Time truncation tmax
    Real lattice data are limited to t < tmax. The paper assumes exponentially suppressed truncation errors (Section 4) but does not quantify them.
  • Test spectral function parameters (m_pi, M_rho, Gamma_rho) = 135, 776, 140 MeV
    Chosen by hand for the synthetic demonstration in Eq. (11) and Figs. 1-2. They are inputs used to generate C(t), not fitted by the method.
assumptions (6)
  • standard math Mellin basis diagonalizes the Carleman operator: the Laplace transform of u_s(x) gives lambda_s u*_s(omega), and H(omega,omega') = 1/(omega+omega') has eigenvalues |lambda_s|^2 = pi/cosh(pi s), Eqs. (5)-(6).
    Classical singular-value decomposition of the Laplace transform kernel (Carleman 1923; McWhirter-Pike 1978; Epstein-Schotland 2008), cited and used without proof.
  • standard math Tikhonov regularization: H_alpha = H + alpha I is invertible and the smeared delta delta_alpha tends to delta as alpha to 0, Eqs. (7)-(9).
    Standard regularization theory for compact operators. Strong convergence of regularized inverses on exact data is assumed.
  • domain assumption The spectral density lies in L^2(R+), or a subtracted rho_s lies in L^2(R+) with C(t) = integral rho_s omega^k e^{-omega t}, Eqs. (14)-(15).
    The paper states the formulas are rigorous only in this class. QCD spectral densities need subtractions with k > 5/2 for the vector current.
  • domain assumption Discrete spectral theory: completeness and orthonormality of the eigenvectors v_s(n,a) of the infinite Hilbert matrix, and existence of the fixed-a alpha to 0 limit, Eqs. (20)-(23).
    The load-bearing discrete-case claims are asserted with proofs deferred to the authors' Ref. [1]. Hill (1960) supplies the Hilbert matrix eigenvectors.
  • ad hoc to paper Exponential suppression of tmax truncation errors, Section 4.
    Stated as 'expected to be exponentially suppressed in tmax [1]'. No proof or bound appears in this paper, and residual O(a^2) errors are admitted.
  • domain assumption Interchange of time, Mellin-parameter, and frequency integrals in Eqs. (8), (10), (21), and (23).
    The coefficient formulas require Fubini-type justifications that are not discussed. Regularity of C(t) and decay of g_alpha are implicitly assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Spectral densities from Euclidean-time lattice correlation functions." pith.science (2026). https://pith.science/paper/RDH3KHQL

@misc{pith2026250116527,
  author       = {Pith},
  title        = {Pith review of: Spectral densities from Euclidean-time lattice correlation functions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RDH3KHQL}},
  note         = {Machine review of arXiv:2501.16527}
}
read the original abstract

In quantum field theories, spectral densities are directly related to relevant physical observables. In Lattice QCD, their non-perturbative extraction from first principles requires the Inverse Laplace transform of Euclidean-time correlation functions, a notorious ill-posed problem. Here we review our recent proposal [1,2] for a new strategy to perform this inversion both in the continuum and on the lattice, also suitable for smeared spectral densities, both in the continuum and in the discrete cases.

Figures

Figures reproduced from arXiv: 2501.16527 by the authors.

Figure 1
Figure 1. Spectral reconstructions, with 𝜌(𝜔) defined as in Eq. (11). 2. In order to invert Eq. (3), we first diagonalize H [22–24] with the “Mellin basis” 𝑢𝑠 (𝑥) = 𝑒 𝑖𝑠 log 𝑥 √ 2𝜋𝑥 , 𝑠 ∈ R s.t. ∫ 𝑦 𝑒 −𝑥 𝑦𝑢𝑠 (𝑦) = 𝜆𝑠𝑢 ∗ 𝑠 (𝑥) , 𝜆𝑠 ≡ Γ  1 2 + 𝑖𝑠 (5) satisfying ∫ 𝜔′ H (𝜔, 𝜔′ )𝑢𝑠 (𝜔 ′ ) = |𝜆𝑠 | 2 𝑢𝑠 (𝜔) , |𝜆𝑠 | 2 = 𝜋 cosh 𝜋𝑠 . (6) 3. Due to its exponentially decreasing eigenvalues |𝜆𝑠 | 2 , the operator H must be regulated pri… view at source ↗
Figure 2
Figure 2. Comparison of discrete coefficients to their continuum (a) and finite-𝑁 counterparts (b). from which we can retrieve the subtracted spectral density as [1] 𝜌s (𝜔) = lim 𝛼→0 ∫ 𝑡 𝑔𝛼,𝑘 (𝑡|𝜔)𝐶(𝑡) , 𝑔𝛼,𝑘 (𝑡|𝜔) = 𝑡 𝑘 ∫ 𝑠 𝑢 ∗ 𝑠 (𝜔)𝜆𝑠,𝑎𝑢 ∗ 𝑠 (𝑡) 𝜆𝑠,𝑎𝜆 ∗ 𝑠,𝑘 + 𝛼 ∀ 𝑎 > − 1 2 . (18) 3. The discrete case We now address the case where the correlation function 𝐶(𝑡) is sampled on an infinite but discrete set of points, 𝑡/𝑎 ∈ N. In… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 11 canonical work pages

  1. [1]

    Bruno, L

    M. Bruno, L. Giusti and M. Saccardi,Spectral densities from Euclidean lattice correlators via the Mellin transform, 2407.04141

    arXiv

  2. [2]

    Bruno, L

    M. Bruno, L. Giusti and M. SaccardiPublication in preparation

    work page

  3. [3]

    Hansen, H.B

    M.T. Hansen, H.B. Meyer and D. Robaina,From deep inelastic scattering to heavy-flavor semileptonic decays: Total rates into multihadron final states from lattice QCD,Phys. Rev. D 96 (2017) 094513 [1704.08993]

    arXiv 2017

  4. [4]

    Bulava and M.T

    J. Bulava and M.T. Hansen,Scattering amplitudes from finite-volume spectral functions, Phys. Rev. D100(2019) 034521 [1903.11735]

    arXiv 2019

  5. [5]

    Barata and K

    J.C.A. Barata and K. Fredenhagen,Particle scattering in Euclidean lattice field theories, Commun. Math. Phys.138 (1991) 507

    work page 1991

  6. [6]

    Patella and N

    A. Patella and N. Tantalo,Scattering Amplitudes from Euclidean Correlators: Haag-Ruelle theory and approximation formulae, 2407.02069

    arXiv

  7. [7]

    Bailas, S

    G. Bailas, S. Hashimoto and T. Ishikawa,Reconstruction of smeared spectral function from Euclidean correlation functions, PTEP 2020(2020) 043B07 [2001.11779]

    arXiv 2020

  8. [8]

    Blum,Lattice calculation of the lowest order hadronic contribution to the muon anomalous magnetic moment,Phys

    T. Blum,Lattice calculation of the lowest order hadronic contribution to the muon anomalous magnetic moment,Phys. Rev. Lett.91 (2003) 052001 [hep-lat/0212018]

    arXiv 2003

Show all 28 references
  1. [9]

    Bernecker and H.B

    D. Bernecker and H.B. Meyer,Vector Correlators in Lattice QCD: Methods and applications,Eur. Phys. J. A47(2011) 148 [1107.4388]

    2011 arXiv

  2. [10]

    Gambino and S

    P. Gambino and S. Hashimoto,Inclusive Semileptonic Decays from Lattice QCD,Phys. Rev. Lett. 125(2020) 032001 [2005.13730]

    2020 arXiv

  3. [11]

    Jeon and L.G

    S. Jeon and L.G. Yaffe,From quantum field theory to hydrodynamics: Transport coefficients and effective kinetic theory, Phys. Rev. D53(1996) 5799 [hep-ph/9512263]

    1996 arXiv

  4. [12]

    Hansen, A

    M. Hansen, A. Lupo and N. Tantalo,Extraction of spectral densities from lattice correlators, Phys. Rev. D99 (2019) 094508 [1903.06476]

    2019 arXiv

  5. [13]

    Backus and F

    G. Backus and F. Gilbert,The Resolving Power of Gross Earth Data,Geophysical Journal International 16 (1968) 169

    1968

  6. [14]

    Valentine and M

    A.P. Valentine and M. Sambridge,Gaussian process models—I. A framework for probabilistic continuous inverse theory,Geophysical Journal International220(2019) 1632

    2019

  7. [15]

    Valentine and M

    A.P. Valentine and M. Sambridge,Gaussian process models—II. Lessons for discrete inversion,Geophysical Journal International220 (2019) 1648

    2019

  8. [16]

    Del Debbio, T

    L. Del Debbio, T. Giani and M. Wilson,Bayesian approach to inverse problems: an application to NNPDF closure testing,Eur. Phys. J. C82(2022) 330 [2111.05787]. 9 Spectral densities from Euclidean-time lattice correlation functions Matteo Saccardi

    2022 arXiv

  9. [17]

    Horak, J.M

    J. Horak, J.M. Pawlowski, J. Rodríguez-Quintero, J. Turnwald, J.M. Urban, N. Wink et al., Reconstructing QCD spectral functions with Gaussian processes,Phys. Rev. D105 (2022) 036014 [2107.13464]

    2022 arXiv

  10. [18]

    Del Debbio, A

    L. Del Debbio, A. Lupo, M. Panero and N. Tantalo,Bayesian solution to the inverse problem and its relation to Backus-Gilbert methods, 2409.04413

    arXiv

  11. [19]

    Buzzicotti, A

    M. Buzzicotti, A. De Santis and N. Tantalo,Teaching to extract spectral densities from lattice correlators to a broad audience of learning-machines,Eur. Phys. J. C84(2024) 32 [2307.00808]

    2024 arXiv

  12. [20]

    Bergamaschi, W.I

    T. Bergamaschi, W.I. Jay and P.R. Oare,Hadronic structure, conformal maps, and analytic continuation, Phys. Rev. D108 (2023) 074516 [2305.16190]

    2023 arXiv

  13. [21]

    Carleman,Sur les équations intégrales singulières à noyau réel et symétrique, inUppsala universitets årsskrift, Matematik och naturvetenskap, 1923

    T. Carleman,Sur les équations intégrales singulières à noyau réel et symétrique, inUppsala universitets årsskrift, Matematik och naturvetenskap, 1923

    1923

  14. [22]

    McWhirter and E.R

    J.G. McWhirter and E.R. Pike,On the numerical inversion of the Laplace transform and similar Fredholm integral equations of the first kind,J. Phys. A: Math. Gen.11(1978) 1729

    1978

  15. [23]

    R. Pike, J. Mcwhirter, M. Bertero and C. Mol,Generalized information theory for inverse problems in signal processing,Communications, Radar and Signal Processing, IEE Proceedings F131 (1984) 660

    1984

  16. [24]

    Epstein and J

    C.L. Epstein and J. Schotland,The Bad Truth about Laplace’s Transform, SIAM Review50 (2008) 504

    2008

  17. [25]

    Tikhonov,Solution of incorrectly formulated problems and the regularization method, Soviet Math

    A. Tikhonov,Solution of incorrectly formulated problems and the regularization method, Soviet Math. Dokl.4(1963) 1035

    1963

  18. [26]

    Tikhonov and V.Y

    A.N. Tikhonov and V.Y. Arsenin,Solutions of ill-posed problems, Scripta Series in Mathematics, V. H. Winston & Sons (1977)

    1977

  19. [27]

    Hill,On the Singly-Infinite Hilbert Matrix,Journal London Math

    C.K. Hill,On the Singly-Infinite Hilbert Matrix,Journal London Math. Soc.s1-35(1960) 17

    1960

  20. [28]

    Bruno, L

    M. Bruno, L. Giusti and M. Saccardi,On the prediction of spectral densities from Lattice QCD, Nuovo Cim. C47 (2024) 197 [2310.16243]. 10

    2024 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.