Spectral asymptotics for a class of integro-differential equations arising in the theory of fractional Gaussian processes

Alexander I. Nazarov

REVIEW 4 major objections 4 minor 25 references

Spectral asymptotics for a class of integro-differential equations arising in the theory of fractional Gaussian processes

T0 review · 4 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read The paper proves that a single boundary-condition parameter controls the second term of the eigenvalue asymptotics for fractional Gaussian covariance operators, with applications to exact small-ball probabilities.

desk verdict Solid generalization of Chigansky–Kleptsyna with a real but containable soft spot: the perturbation theorem leans on a self-cited lemma and an overbroad positivity claim. read the letter →

arxiv 1908.10299 v2 pith:N2TLKQQK submitted 2019-08-27 math.SP math.PR

Alexander I. Nazarov This is my paper · ORCID

classification math.SPmath.PR MSC 34L2047A7545C0560G1560G22

keywords fractionalBrownianmotioneigenvalueasymptoticsintegro-differentialequationsRiemann-HilbertproblemsmallballprobabilitiesSlepianprocessOrnstein-UhlenbeckL2-small

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

This paper establishes the two-term eigenvalue asymptotics for a family of integro-differential equations, $(K_\alpha\psi)(x)=-\lambda\psi''(x)$, that arise from the covariance operators of fractional Gaussian processes. The leading eigenvalue was already known, but exact small-ball probabilities require the second term; the paper shows that for separated self-adjoint boundary conditions the second term is controlled by a single integer, the sum $\kappa$ of the orders of the derivatives appearing in the boundary conditions, and writes the formula explicitly. It then proves that adding a potential $p\in L_1(0,1)$ on the right-hand side does not alter this two-term asymptotics. These spectral facts are converted into exact $L_2$-small-ball probabilities, meaning the probability that the $L_2$ norm of the process is at most $\varepsilon$, for fractional Brownian bridges, centered fractional processes, fractional Slepian processes, and fractional Ornstein–Uhlenbeck processes. A reader should care because knowing the second spectral term is exactly what turns logarithmic small-ball estimates into estimates with correct constants.

What carries the argument

The object that carries the argument is the operator $K_\alpha$ given by $(K_\alpha\psi)(x)=(1-\alpha/2)\frac{d}{dx}\int_0^1 \mathrm{sign}(x-y)|x-y|^{1-\alpha}\psi(y)\,dy$; after an integration by parts it is the covariance kernel of fractional Brownian motion written as an integro-differential operator. The proof machinery is the Laplace-transform reduction of the eigenproblem to a Riemann–Hilbert problem on the real axis. The key derived objects are the angle function $\theta_0(t)=\arctan\left(\frac{\sin(\pi(1-\alpha)/2)}{\cos(\pi(1-\alpha)/2)+t^{3-\alpha}}\right)$, its integral $b_\alpha=\cot(\pi/(3-\alpha))$, and the Sokhotski–Plemelj solution $X_0(z)$ of the jump problem on the positive semiaxis. Together these produce a $4\times4$ linear system whose solvability condition yields the eigenvalue equation; the integer $\kappa$ enters as the shift in that condition and is the only boundary-condition data that survives in the second term.

What would settle it

Take $\alpha=1/2$, $p\equiv 1$, Dirichlet boundary conditions, and compute the first several dozen eigenvalues of $(K_\alpha\psi)(x)=\lambda(-\psi''(x)+\psi(x))$. Writing $\lambda_n=\sin(\pi\alpha/2)\Gamma(3-\alpha)\nu_n^{\alpha-3}$, formula (24) predicts $\nu_n=\pi n-\pi(1-\alpha)/4+O(n^{-1})$ with no additional constant; a numerical intercept different from $-\pi(1-\alpha)/4$ outside the stated remainder, or a difference between $p\equiv 0$ and $p\equiv 1$ at that order, would falsify Theorem 3.2.

Watch

Extended reading notes

Core claim

The central claim is Theorem 2.1: for the problem $(K_\alpha\psi)(x)=-\lambda\psi''(x)$ with separated self-adjoint boundary conditions and $\alpha=2-2H\in(0,2)\setminus\{1\}$, the eigenvalues obey $$\lambda_n = \sin\left(\frac{\pi\$\alpha$}{2}\right)\Gamma(3-\$\alpha$)\left(\pi n - \frac{\pi(1-\$\alpha$)}{4} - \frac{\kappa\pi}{3-\$\alpha$} + O($n^{{-1}}$)\right)^{\$\alpha$-3},$$ where $\kappa\in\{0,1,2\}$ is the sum of the orders of the derivatives in the boundary conditions. Theorem 2.4 gives the analogous result for non-separated self-adjoint boundary conditions, with the eigenvalues splitting into two subsequences whose shifts differ by an arcsine term. Theorem 3.2 then shows that in the generalized problem $(K_\alpha\psi)(x)=\lambda(-\psi''(x)+p(x)\psi(x))$ with $p\in L_1(0,1)$, the same two-term formulas hold unchanged. The final section feeds these formulas into the standard transfer principle to produce exact $L_2$-small-ball asymptotics for the fractional Gaussian processes considered in Section 4.

Load-bearing premise

The load-bearing premise is an abstract perturbation lemma quoted verbatim from the author's companion preprint, together with the assumption that the free operator is positive definite; if either fails, the p-independence of the two-term asymptotics is not established.

Editorial extensions

If this is right

For every separated self-adjoint boundary condition, the two-term eigenvalue formula holds, with only three possible second-term shifts corresponding to $\kappa=0,1,2$.
Adding an $L_1$ potential to the right-hand side of the eigenproblem does not change the two-term spectrum, so the spectral formulas are universal across a whole family of fractional processes.
The fractional Brownian bridge and the centered fractional Brownian motion receive explicit eigenvalue asymptotics, with the bridge result reproducing an earlier formula with a sharper remainder estimate.
The fractional Slepian process and both fractional Ornstein–Uhlenbeck cases fit the same scheme, including variants where the spectral parameter appears in the boundary conditions.
The table in Section 5 yields exact constants in $L_2$-small-ball probabilities for all processes considered, not merely logarithmic rates.

Reading between the lines

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

Beyond the paper, the conjecture stated in the introduction suggests that the same $\kappa$-driven shift may control two-term asymptotics for higher-order analogues of the equation; the boundary-condition mechanism here is geometric enough that this would be a natural testbed.
Because the perturbation lemma is stated abstractly, it likely applies to potentials that are finite signed measures rather than only $L_1$ functions; if so, Theorem 3.2 would extend to delta-type and boundary-contact perturbations without new ideas.
The small-ball table implies that the centered fractional Brownian motion and the fractional Brownian bridge, which coincide for $H=1/2$, have power-level small-ball exponents differing by $1$ for every $H\ne 1/2$; direct simulation of small-ball probabilities at small $\varepsilon$ could test this separation empirically.

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.

Desk Editor's Note

read the letter

Colleague,

The thing to know: Section 2 is the real work and it holds up. The two-term eigenvalue asymptotics for (K_α ψ) = −λψ'' with general self-adjoint boundary conditions, with the shift κ depending on the sum of derivative orders, is a genuine extension of Chigansky–Kleptsyna. I checked the reductions to known cases (α = 1, FBM [6], fractional bridge [7]) and the formulas match. The non-separated case producing two interleaved subsequences with opposite shifts also looks right and parallels Nazarov's earlier ODO results.

What's new: the κ-dependence, the potential perturbation p ∈ L1 invariance, and the new processes (fractional Slepian, centered FBM/bridge, fractional OU with nonzero initial variance). The small-ball table is a direct application of known theorems; the new exponents are credible.

Soft spots, in order:

1. Theorem 3.2's proof is carried by Proposition 3.1, quoted verbatim from the author's own preprint [17] and not proved here. That is acceptable in a preprint culture, but for a journal referee it is a real gap: the lemma has hypotheses (I+B positive, decay condition (35)) and the paper does not verify them fully for all stated cases. If [17] has hidden hypotheses or a weaker remainder, Theorem 3.2 collapses. I would ask the author to include the proof of Prop. 3.1 in an appendix or cite a peer-reviewed version.

2. More concretely, the stated generality of Theorem 3.2 (p ∈ L1, no positivity) is not supported. Prop. 3.1 requires I+B positive, and the 'standard way' for non-positive-definite second-order operators is not spelled out. For the examples in Section 4 the potential is β² ≥ 0, so I+B is positive and the applications are safe. But the theorem as stated overreaches. This is fixable by adding a positivity hypothesis or a proper approximation argument.

3. The verification of (35) for non-separated BCs is terse: the paper says it 'also holds' and cites Remark 2 of [17]. In my reading, (30)–(31) are BC-independent, so |ψ_n|_∞ = O(n^{-1}) does follow in the non-separated case too, and then (35) is immediate. The stress-test note is harsher than I would be here—I think it is a missing sentence, not a hidden flaw.

Also, the paper's own Remark 2.2 about zero root shifting indices is careful, which I appreciate.

Bottom line: this deserves peer review. The central Section 2 result is solid and useful, and the examples give exact small-ball asymptotics for several processes that lacked them. I would send it with a request to address the Prop. 3.1 dependency and the positivity mismatch; neither looks fatal for the applications, but the theorem statement should match the proof.

For you: if you work on fractional Gaussian processes or spectral asymptotics of integral operators, it is worth a read. I would not cite it directly in my own work, but I would put it on the reading list.

Recommendation: accept with revision—the core is sound.

Referee Report

4 major / 4 minor

Summary. This paper studies the spectral problem (K_alpha psi)(x) = -lambda psi''(x) on (0,1) with general self-adjoint boundary conditions, where K_alpha is the integro-differential operator associated with fractional Brownian motion. Following the Laplace-transform and Riemann-Hilbert approach of Chigansky and Kleptsyna, the author derives two-term eigenvalue asymptotics: Theorem 2.1 for separated boundary conditions, with the shift parameter kappa equal to the sum of the orders of the derivatives in the boundary conditions, and Theorem 2.4 for non-separated boundary conditions. Section 3 states Theorem 3.2, asserting that adding a potential p in L1(0,1) on the right-hand side does not affect the two-term asymptotics. Section 4 applies the results to the fractional Brownian bridge, centered FBM, centered Brownian bridge, fractional Slepian processes, and fractional Ornstein-Uhlenbeck processes. Section 5 converts these spectral results into exact L2-small-ball asymptotics, summarized in Theorem 5.1 and a table.

Significance. If the results are correct, the paper provides a unified extension of the two-term spectral asymptotics and exact L2-small-ball asymptotics to a family of fractional Gaussian processes, going substantially beyond the FBM case treated in [6] and covering several new processes. The consistency checks are persuasive: the formulas reduce properly to the Sturm-Liouville case alpha=1 and to the FBM asymptotics of [6]. The perturbation principle in Section 3, if established, would be a useful general tool. However, the proof of Theorem 3.2 depends on an unproved, self-cited perturbation lemma and on an incomplete verification of the key decay estimate in the non-separated case. These gaps make the significance conditional on substantial revision.

major comments (4)

[Section 3, Proposition 3.1] The perturbation mechanism is Proposition 3.1, which is quoted verbatim as Theorem 1 of the author's preprint [17] and is not proved in this manuscript. This is load-bearing for the central claim: the entire proof of Theorem 3.2 reduces to applying Proposition 3.1 after checking (34) and (35). Without a proof, or at least a complete statement of all hypotheses and the exact remainder exponent, a reader cannot verify the main perturbation result. The paper should either include a proof of Proposition 3.1 or supply the full statement with the actual conditions and a pointer to a published, refereed proof.
[Section 3, verification of (35) for non-separated boundary conditions] After equation (36), the paper asserts that for non-separated boundary conditions the estimate (35) also holds and cites Remark 2 of [17], but no derivation is given. In the non-separated case the eigenvalues form two interleaved subsequences, as shown in (26), and the bound |psi_n| = O(n^{-1}) used to derive (36) is not justified for these subsequences. The separate-case argument relies on the assertion that all eigenfunctions except the first change sign, which is not automatically true for periodic or anti-periodic boundary conditions. Therefore Theorem 3.2 is not proven for non-separated boundary conditions as the text stands.
[Section 3, positivity hypothesis in Theorem 3.2] Theorem 3.2 states p in L1(0,1) with no positivity condition, while Proposition 3.1 requires I+B to be positive. The opening paragraph of Section 3 only assumes that -psi'' with the given boundary conditions is positive definite and defers the remaining cases to 'a standard way'. If p is negative on a set of positive measure, the operator I+B can be indefinite or have negative eigenvalues; in that situation the decreasing-enumeration and spectral parameterization used in Proposition 3.1 are not well defined. The stated generality of Theorem 3.2 therefore exceeds what is proved. The paper should either restrict Theorem 3.2 to p for which I+B is positive, or provide a complete argument for the indefinite case.
[Section 4, Examples 4 and 7] The fractional Slepian process (Example 4) and the fractional Ornstein-Uhlenbeck process with sigma != 0 (Example 7) have boundary conditions containing the spectral parameter lambda, as seen in (41) and (48). Consequently Theorem 2.1 and Theorem 3.2 do not apply directly to these problems. The text states that 'the basic scheme runs without essential changes' and then derives the eigenvalue asymptotics from (19), but the required analogue of the Rouché theorem step and the validity of the eigenfunction estimates (30)-(31) under lambda-dependent boundary conditions are not shown. Since these examples feed into the small-ball table in Theorem 5.1, those entries are not fully justified by the arguments presented.

minor comments (4)

[Acknowledgements] There is a typo in the Acknowledgements: 'greatful' should be 'grateful'.
[Section 3, Theorem 3.2 statement] The phrase 'self-adjoint boundary conditions' in Theorem 3.2 is not fully specified. The paper uses the parametric forms (20) and (25), but it does not state the conditions on beta, gamma, delta under which (25) are self-adjoint and the underlying operator -d^2/dx^2 is positive definite.
[Section 2.4, Theorem 2.4] In formula (27), the two subsequences are combined using the factor (-1)^n. The paper should specify the exact ordering convention for the eigenvalues when the two subsequences interleave, so that the indexing of lambda_n is unambiguous.
[Section 5, notation] The symbol B used for the quadratic-form constant B(H) in Theorem 5.1 conflicts with the operator B introduced in Section 3. This is not mathematically wrong but makes the reading unnecessarily confusing.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the main two-term eigenvalue asymptotics are derived from the Riemann–Hilbert reduction, and the cited perturbation lemma is a generic transfer theorem whose hypotheses do not contain its conclusion.

full rationale

The central derivation of formula (24) is not circular. The paper recomputes the Riemann–Hilbert reduction for general boundary conditions, solves the factorization problem via Sokhotski–Plemelj, obtains the characteristic system (19), and applies Rouché's theorem; the eigenvalue asymptotics are consequences of the equation and the boundary conditions, not assumed inputs. The numbering constant k is fixed by comparison with the classical Sturm–Liouville case α = 1, an external benchmark, and Section 4 checks several resulting formulas against known results [7], [8], [22], [18]. Section 3's p-independence is a transfer theorem: Proposition 3.1 is quoted from the author's own preprint [17], but its assumptions (two-term asymptotics for the unperturbed operator plus the matrix estimate (35)) do not include the perturbed conclusion, and the lemma is parameter-free with a stated general hypothesis; under the review rules this is independent support rather than circularity. The same applies to Remark 2 of [17] for the two-subsequence case. The genuine weaknesses are proof-gap and limitation issues, not circularity: Proposition 3.1 is not proved in this text; the verification of (35) for non-separated boundary conditions is only asserted ('the estimate (35) also holds'); and the perturbation proof assumes -ψ'' is positive definite with other cases deferred 'in a standard way'. These affect verifiability, completeness, and scope, but they do not make any prediction equivalent to its input by construction or by self-referential definition. Therefore the circularity score is 0.

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

The paper introduces no fitted constants and no new physical or mathematical entities. The functions Phi, Psi, X0, S and D are technical constructs of the proof, not invented entities with independent physical handles. The only external assumptions are standard analytic tools, process identities, and the self-cited perturbation lemma from [17].

assumptions (6)

domain assumption Theorem 1 of [17], quoted as Proposition 3.1: for compact self-adjoint K and B with K and I+B positive, a two-term spectral asymptotics with eigenvector bounds transfers to the generalized problem.
Imported from the author's own preprint and used as the mechanism for showing that the potential p in (2) does not change the asymptotics; not proven in this text.
domain assumption The Riemann-Hilbert factorization lemmas of [6] (Lemma 5.5, 5.6, 5.7), including the asymptotic behavior of X0, p0 plus/minus, p1 plus/minus and the contraction property of the integral operator A.
These lemmas are used throughout Section 2 to control the functions Phi0, Psi0 and the integral equations; the paper refers to them without reproving them.
domain assumption The small-ball theorem of [18, Theorem 6.2] and the comparison results of [11] that convert two-term spectral asymptotics with remainder into exact L2 small-ball asymptotics.
Section 5 applies this conversion; the paper does not derive it.
domain assumption The process identities and factorizations used for the examples, such as integral-zero properties of centered fractional Brownian motion and the doubling relation for fractional Slepian processes, plus the Lifshits lemma for critical perturbations.
These are stated as known or directly verified process facts; they are needed to reduce the covariance eigenproblems to the spectral problem (1).
domain assumption Main theorems assume self-adjoint boundary conditions that do not contain the spectral parameter; Section 3 also assumes positivity of the operator -psi'' with the given boundary conditions, with other cases deferred to standard modifications.
Stated at the beginning of Section 3; examples with spectral parameter or non-positive operators are treated by sketches rather than full proofs.
standard math Sokhotski-Plemelj formula, analytic continuation, Rouche's theorem, and standard spectral theory of ordinary differential operators are used without proof.
Background tools invoked in Sections 2 and 3.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Spectral asymptotics for a class of integro-differential equations arising in the theory of fractional Gaussian processes." pith.science (2026). https://pith.science/paper/N2TLKQQK

@misc{pith2026190810299,
  author       = {Pith},
  title        = {Pith review of: Spectral asymptotics for a class of integro-differential equations arising in the theory of fractional Gaussian processes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N2TLKQQK}},
  note         = {Machine review of arXiv:1908.10299}
}

abstract

We study spectral problems for integro-differential equations arising in the theory of Gaussian processes similar to the fractional Brownian motion. We generalize the method of Chigansky--Kleptsyna and obtain the two-term eigenvalue asymptotics for such equations. Application to the small ball probabilities in $L_2$-norm is given.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 25 canonical work pages

[6]

Exact Asymptotics in Ei genproblems for Fractional Brownian Covariance Operators

Chigansky, P., and Kleptsyna, M. Exact Asymptotics in Ei genproblems for Fractional Brownian Covariance Operators. Stoch. Proc. Ap pl., 128 (2018), N6, 2007–2059

work page 2018
[7]

On the eigenproblem for Gaussian bridges

Chigansky, P., Kleptsyna, M., and Marushkevych, D. On th e eigenprob- lem for Gaussian bridges. Preprint available at https://ar xiv.org/pdf/ 1706.09298.pdf. 18 pp

work page Pith review arXiv
[17]

Some lemmata on the perturbation of the spectrum

Nazarov, A.I. Some lemmata on the perturbation of the sp ectrum. Preprint available at https://arxiv.org/abs/1908.09365. 6pp

work page Pith review arXiv 1908
[1]

Asymptotic behavior of the spectrum of weakly polar integral operators

Birman, M.S., and Solomyak, M.Z. Asymptotic behavior of the spectrum of weakly polar integral operators. Izv. AN SSSR. Ser. matem ., 34 (1970), N6, 1143–1158 (Russian); English transl.: Math. of the USSR -Izvestiya, 4 (1970), N5, 1151-1168

work page 1970
[2]

Quantitative analysis in Sobolev imbed- ding theorems and applications to spectral theory

Birman, M.S., and Solomyak, M.Z. Quantitative analysis in Sobolev imbed- ding theorems and applications to spectral theory. In: Proc eed. of X Summer Mathematical School. Yu.A. Mitropol’skiy and A.F. Shestop al (Eds), 1974, 5–189 (Russian); English transl. in: AMS Translations, Ser ies 2, 114. AMS, Providence, R.I. 1980

work page 1974
[3]

Spectral theory of self -adjoint operators in Hilbert space, 2nd ed., revised and extended

Birman, M.S., and Solomyak, M.Z. Spectral theory of self -adjoint operators in Hilbert space, 2nd ed., revised and extended. Lan’, St.Pe tersburg, 2010 (Russian); English transl. of the 1st ed.: Mathematics and I ts Applications. Soviet Series, 5, Kluwer, Dordrecht etc. 1987

work page 2010
[4]

Small ball constants and tight eigenvalue asymptotics for frac- tional Brownian motions

Bronski, J.C. Small ball constants and tight eigenvalue asymptotics for frac- tional Brownian motions. J. Theor. Probab., 16 (2003), 87–100

work page 2003
[5]

Fractiona l Ornstein– Uhlenbeck processes

Cheridito, P., Kawaguchi, H., and Maejima, M. Fractiona l Ornstein– Uhlenbeck processes. Electron. J. Probab., 8 (2003), N3, 1–14

work page 2003

Show all 25 references

[8]

Exac t spectral asymp- totics of fractional processes

Chigansky, P., Kleptsyna, M., and Marushkevych, D. Exac t spectral asymp- totics of fractional processes. Preprint available at http s://arxiv.org/pdf/ 1802.09045.pdf. 52 pp. 29

arXiv
[9]

Comparison theorems for small devia- tions of random series

Gao, F., Hannig, J., and Torcaso, T. Comparison theorems for small devia- tions of random series. Electron. J. Probab., 8 (2003), paper N21, 1–17

2003
[10]

Tables of integrals , sums, series and products, 5th ed

Gradshteyn, I.S., and Ryzhik, I.M. Tables of integrals , sums, series and products, 5th ed. Moscow, Nauka, 1971 (Russian); English tr ansl.: Table of integrals, series, and products. Corr. and enl. ed. by Alan J eﬀrey. New York – London – Toronto: Academic Press, 1980

1971
[11]

Comparison results for the lower tail of Gaussi an seminorms

Li, W.V. Comparison results for the lower tail of Gaussi an seminorms. J. Theor. Probab., 5 (1992), N1, 1–31

1992
[12]

Survival Exponents for Some Gaussian Proce sses

Molchan, G. Survival Exponents for Some Gaussian Proce sses. Int. J. Stoch. Analysis, 2012 (2012), ID 137271, 1–20

2012
[13]

Linear Diﬀerential Operators

Naimark, M.A. Linear Diﬀerential Operators. 2nd ed. Mos cow, Nauka, 1969 (Russian); English transl. of the 1st ed.: Naimark M.A. Line ar Diﬀerential Operators. Part I (1967): Elementary Theory of Linear Diﬀere ntial Opera- tors. With add. material by the author. N.Y.: F. Ungar...

1967
[14]

Log-level comparison principle for smal l ball probabilities

Nazarov, A.I. Log-level comparison principle for smal l ball probabilities. Stat. & Prob. Letters, 79 (2009), N4, 481–486

2009
[15]

Exact L2-Small Ball Asymptotics of Gaussian Processes and the Spectrum of Boundary-Value Problems

Nazarov, A.I. Exact L2-Small Ball Asymptotics of Gaussian Processes and the Spectrum of Boundary-Value Problems. J. Theor. Probab. , 22 (2009), N3, 640–665

2009
[16]

On a set of transformations of Gaussian ra ndom functions

Nazarov, A.I. On a set of transformations of Gaussian ra ndom functions. Teor. Ver. Primen., 54 (2009), N2, 209–225 (Russian); English transl.: Theor. Probab. Appl., 54 (2010), N2, 203–216

2009
[18]

Exact small ball beh avior of integrated Gaussian processes under L2-norm and spectral asymptotics of boundary value problems

Nazarov, A.I., and Nikitin, Ya.Yu. Exact small ball beh avior of integrated Gaussian processes under L2-norm and spectral asymptotics of boundary value problems. Probab. Theory and Rel. Fields, 129 (2004), N4, 469–494

2004
[19]

Logarithmic L2-small ball asymptotics for some fractional Gaussian processes

Nazarov, A.I., and Nikitin, Ya.Yu. Logarithmic L2-small ball asymptotics for some fractional Gaussian processes. Teor. Ver. Primen. , 49 (2004), N4, 695–711 (Russian); English transl.: Theor. Probab. Appl., 49 (2005), N4, 645–658. 30

2004
[20]

On Small Deviation A symptotics in L2 of Some Mixed Gaussian Processes

Nazarov, A.I., and Nikitin, Ya.Yu. On Small Deviation A symptotics in L2 of Some Mixed Gaussian Processes. Mathematics, 6 (2018), N4, paper N55, 1–9

2018
[21]

The small ball asympt otics in Hilbertian norm for the Kac–Kiefer–Wolfowitz processes

Nazarov, A.I., and Petrova, Yu.P. The small ball asympt otics in Hilbertian norm for the Kac–Kiefer–Wolfowitz processes. Teor. Ver. Primen., 60 (2015), N3, 482–505 (Russian); English transl.: Theor. Probab. App l., 60 (2016), N3, 460–480

2015
[22]

Sharp small ball asy mptotics for Slepian and Watson processes in Hilbert norm

Nikitin, Ya.Yu., and Orsingher, E. Sharp small ball asy mptotics for Slepian and Watson processes in Hilbert norm. ZNS POMI, 320 (2004), 120–128 (Russian); English transl.: J. Math. Sci. (N.Y.), 137 (2006), N1. 4555–4560

2004
[23]

Asymptotic behavior of the spectrum and eigenfunctions of convolution operators on a ﬁnite interval with the kernel ha ving a homoge- neous Fourier transform

Pal’tsev, B.V. Asymptotic behavior of the spectrum and eigenfunctions of convolution operators on a ﬁnite interval with the kernel ha ving a homoge- neous Fourier transform. Dokl. Akad. Nauk SSSR, 218 (1974), N1, 28–31 (Russian)

1974
[24]

L2-small ball asymptotics for a family of ﬁnite-dimensional perturbations of Gaussian functions

Petrova, Yu.P. L2-small ball asymptotics for a family of ﬁnite-dimensional perturbations of Gaussian functions. Preprint available a t https://arxiv.org/ pdf/1905.07804.pdf. 17 pp

1905 arXiv
[25]

Asymptotic distribution of eigenvalues of the kernel in the Kirkwood–Riseman integral equation

Ukai, S. Asymptotic distribution of eigenvalues of the kernel in the Kirkwood–Riseman integral equation. J. Math. Phys., 12 (1971), N1, 83–92. 31

1971

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [6]

Exact Asymptotics in Ei genproblems for Fractional Brownian Covariance Operators

Chigansky, P., and Kleptsyna, M. Exact Asymptotics in Ei genproblems for Fractional Brownian Covariance Operators. Stoch. Proc. Ap pl., 128 (2018), N6, 2007–2059

work page 2018

[2] [7]

On the eigenproblem for Gaussian bridges

Chigansky, P., Kleptsyna, M., and Marushkevych, D. On th e eigenprob- lem for Gaussian bridges. Preprint available at https://ar xiv.org/pdf/ 1706.09298.pdf. 18 pp

work page Pith review arXiv

[3] [17]

Some lemmata on the perturbation of the spectrum

Nazarov, A.I. Some lemmata on the perturbation of the sp ectrum. Preprint available at https://arxiv.org/abs/1908.09365. 6pp

work page Pith review arXiv 1908

[4] [1]

Asymptotic behavior of the spectrum of weakly polar integral operators

Birman, M.S., and Solomyak, M.Z. Asymptotic behavior of the spectrum of weakly polar integral operators. Izv. AN SSSR. Ser. matem ., 34 (1970), N6, 1143–1158 (Russian); English transl.: Math. of the USSR -Izvestiya, 4 (1970), N5, 1151-1168

work page 1970

[5] [2]

Quantitative analysis in Sobolev imbed- ding theorems and applications to spectral theory

Birman, M.S., and Solomyak, M.Z. Quantitative analysis in Sobolev imbed- ding theorems and applications to spectral theory. In: Proc eed. of X Summer Mathematical School. Yu.A. Mitropol’skiy and A.F. Shestop al (Eds), 1974, 5–189 (Russian); English transl. in: AMS Translations, Ser ies 2, 114. AMS, Providence, R.I. 1980

work page 1974

[6] [3]

Spectral theory of self -adjoint operators in Hilbert space, 2nd ed., revised and extended

Birman, M.S., and Solomyak, M.Z. Spectral theory of self -adjoint operators in Hilbert space, 2nd ed., revised and extended. Lan’, St.Pe tersburg, 2010 (Russian); English transl. of the 1st ed.: Mathematics and I ts Applications. Soviet Series, 5, Kluwer, Dordrecht etc. 1987

work page 2010

[7] [4]

Small ball constants and tight eigenvalue asymptotics for frac- tional Brownian motions

Bronski, J.C. Small ball constants and tight eigenvalue asymptotics for frac- tional Brownian motions. J. Theor. Probab., 16 (2003), 87–100

work page 2003

[8] [5]

Fractiona l Ornstein– Uhlenbeck processes

Cheridito, P., Kawaguchi, H., and Maejima, M. Fractiona l Ornstein– Uhlenbeck processes. Electron. J. Probab., 8 (2003), N3, 1–14

work page 2003