{"id":"aadc28fa-dc3c-4ca3-a193-d8b999753cb5","arxiv_id":"1908.10299","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For fractional Gaussian covariance operators satisfying (K_alpha psi)(x) = -lambda psi''(x), the eigenvalues obey lambda_n = C (pi n - c_alpha - kappa pi/(3-alpha) + O(n^{-1}))^{alpha-3}, and the same formula persists when a potential term is added.","lead":"This paper derives two-term eigenvalue asymptotics for a family of integro-differential equations linked to fractional Gaussian processes, under general self-adjoint boundary conditions. The result yields exact L2 small-ball probabilities for several fractional processes and unifies earlier work by Chigansky and Kleptsyna.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.2's p-independence depends on an unproved self-cited perturbation lemma (Prop. 3.1 = Thm 1 of [17]), and even granting it, the verification of its decay assumption (35) for non-separated boundary conditions is only asserted, not shown.","rationale":"The genuinely new content of the paper beyond [6] is the perturbation result for p∈L1 and the boundary-condition taxonomy. Section 2's derivation of (24) is a careful adaptation of [6], and the formula reduces correctly to known Sturm-Liouville and FBM cases; the small-ball table follows from the spectral formulas by known machinery. The load-bearing step is Section 3, where Theorem 3.2 is not proved but quoted from [17]. Proposition 3.1 is a nontrivial abstract perturbation statement; it is not reproduced, and the verification that its condition (35) holds in all boundary-condition regimes is compressed into a single sentence for non-separated conditions. The theorem's stated p∈L1 also lacks the positivity hypothesis needed by Proposition 3.1. Thus the conditional verdict is appropriate: the result is likely true but not self-contained. A concrete check of [17] and of (35) for periodic/anti-periodic BC would settle the main uncertainty.","tokens_in":17014,"tokens_out":13308,"duration_ms":122718,"concrete_test":"Inspect the proof of Theorem 1 in arXiv:1908.09365 and, for H = energy space of problem (1), check that its assumptions are exactly met by K and B defined in (33) with p∈L1 and I+B merely positive (not necessarily positive definite). Separately, for the periodic/anti-periodic case β=±γ in (25), compute (Bψ_n,ψ_m)_H for p=χ_E using the two-sequence eigenfunction representation from (30); verify the O((mn)^{-1}) bound follows from (26). If either check fails, Theorem 3.2 needs a corrected hypothesis or a new proof.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of Section 3 is that adding p∈L1 to the right-hand side of (1) does not change the two-term asymptotics. The proof is entirely carried by Proposition 3.1, which is quoted verbatim as Theorem 1 of the author's own preprint [17] and is not proved here. If that lemma has hidden hypotheses or a different remainder exponent, Theorem 3.2 collapses. Moreover, the paper's verification of condition (35) is incomplete. The bound |(Bψn,ψm)_H|≤c(mn)^{-1} is obtained from |ψ_n|∞=O(n^{-1}), which in turn follows from (30)-(31) plus the assertion that all eigenfunctions except the first change sign. For separated conditions this is plausible, but for the non-separated case the paper simply states 'the estimate (35) also holds' and cites Remark 2 of [17]; no derivation is given for the two interleaved subsequences in (26). There is also a generality mismatch: Proposition 3.1 requires I+B positive, while Theorem 3.2 states p∈L1 with no positivity condition; arbitrary negative p can make I+B indefinite, and the sketched 'standard way' does not cover this. A reader cannot verify the main perturbation result without access to a proof of [17] and a complete check of (35).","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","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.","tokens_in":17295,"tokens_out":6386,"duration_ms":61079,"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":[{"comment":"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":"Section 3, Proposition 3.1"},{"comment":"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":"Section 3, verification of (35) for non-separated boundary conditions"},{"comment":"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":"Section 3, positivity hypothesis in Theorem 3.2"},{"comment":"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.","section":"Section 4, Examples 4 and 7"}],"minor_comments":[{"comment":"There is a typo in the Acknowledgements: 'greatful' should be 'grateful'.","section":"Acknowledgements"},{"comment":"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":"Section 3, Theorem 3.2 statement"},{"comment":"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":"Section 2.4, Theorem 2.4"},{"comment":"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.","section":"Section 5, notation"}],"recommendation":"major_revision","confidential_remarks":"The central difficulty is the heavy reliance of Section 3 on the author's unpublished preprint [17], both for Proposition 3.1 and for Remark 2. I recommend requesting a full proof of Proposition 3.1 or a published version of [17] before the paper can be accepted. The Section 2 derivation appears sound and is a genuine contribution; if the perturbation gaps are filled, the paper would be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe 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.\n\nWhat'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.\n\nSoft spots, in order:\n\n1. 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.\n\n2. 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.\n\n3. 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.\n\nAlso, the paper's own Remark 2.2 about zero root shifting indices is careful, which I appreciate.\n\nBottom 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.\n\nFor 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.\n\nRecommendation: accept with revision—the core is sound.","headline":"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.","tokens_in":17804,"tokens_out":4391,"would_cite":false,"duration_ms":39248,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["34L20","47A75","45C05","60G15","60G22"],"pacs":[],"model":"deepseek-v4-flash","headline":"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.","keywords":["fractional Brownian motion","eigenvalue asymptotics","integro-differential equations","Riemann-Hilbert problem","small ball probabilities","fractional Slepian process","fractional Ornstein-Uhlenbeck process","L2-small ball asymptotics"],"falsifier":"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.","tokens_in":16804,"feed_emoji":"🎲","tokens_out":10056,"duration_ms":98342,"temperature":0.7,"pith_summary":"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.","feed_headline":"Two-term eigenvalue law for fractional Gaussian processes","feed_subtitle":"Boundary conditions alone set the second eigenvalue term, yielding exact small-ball asymptotics for several fractional Gaussian processes.","key_machinery":"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.","core_discovery":"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.","pith_inferences":["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."],"forward_implications":["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."],"supporting_citations":[{"why":"Supplies the baseline one-term eigenvalue asymptotics for fractional Brownian motion that the paper's two-term result refines.","marker":"[4]"},{"why":"Provides the Laplace-transform and Riemann–Hilbert reduction that this paper generalizes to arbitrary self-adjoint boundary conditions.","marker":"[6]"},{"why":"Gives the earlier two-term result for Gaussian bridges that the paper rederives as a special case of Theorem 2.1.","marker":"[7]"},{"why":"Contains exact spectral formulas for fractional Ornstein–Uhlenbeck processes that the paper recovers through Theorem 3.2.","marker":"[8]"},{"why":"Supplies the transfer principle from two-term spectra to exact $L_2$-small-ball asymptotics and the analysis of non-separated boundary conditions.","marker":"[15]"},{"why":"Classifies critical versus non-critical finite-dimensional perturbations, used for the modified fractional Slepian process.","marker":"[16]"},{"why":"Contains the abstract perturbation lemma quoted as Proposition 3.1, which carries the p-independence result.","marker":"[17]"},{"why":"Provides the general exact-small-ball theorem used in Section 5 and the boundary-order parameter idea that motivates $\\kappa$.","marker":"[18]"}],"fun_headline_variants":["Boundary conditions fix second eigenvalue term","Two-term spectrum for fractional Gaussian processes","Eigenvalue asymptotics sharpen small-ball laws","Boundary data governs eigenvalue corrections","Exact small-ball asymptotics from spectral two-term"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"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.","fun_headline_variants_meta":{"raw":{"variants":["Boundary conditions fix second eigenvalue term","Two-term spectrum for fractional Gaussian processes","Eigenvalue asymptotics sharpen small-ball laws","Boundary data governs eigenvalue corrections","Exact small-ball asymptotics from spectral two-term"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000376,"raw_usage":{"total_tokens":1959,"prompt_tokens":854,"completion_tokens":1105,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":470,"completion_tokens_details":{"reasoning_tokens":1040}},"tokens_in":470,"tokens_out":1105,"duration_ms":8972,"temperature":1.0,"reasoning_tokens":1040,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:48:08.941438+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"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.","supporting_citations":[{"cited_title":"Small ball constants and tight eigenvalue asymptotics for frac- tional Brownian motions","cited_arxiv_id":null,"evidence_quote":"Supplies the baseline one-term eigenvalue asymptotics for fractional Brownian motion that the paper's two-term result refines."},{"cited_title":"Exact Asymptotics in Ei genproblems for Fractional Brownian Covariance Operators","cited_arxiv_id":null,"evidence_quote":"Provides the Laplace-transform and Riemann–Hilbert reduction that this paper generalizes to arbitrary self-adjoint boundary conditions."},{"cited_title":"On the eigenproblem for Gaussian bridges","cited_arxiv_id":"1706.09298","evidence_quote":"Gives the earlier two-term result for Gaussian bridges that the paper rederives as a special case of Theorem 2.1."},{"cited_title":"Exact spectral asymptotics of fractional processes","cited_arxiv_id":"1802.09045","evidence_quote":"Contains exact spectral formulas for fractional Ornstein–Uhlenbeck processes that the paper recovers through Theorem 3.2."},{"cited_title":"Exact L2-Small Ball Asymptotics of Gaussian Processes and the Spectrum of Boundary-Value Problems","cited_arxiv_id":null,"evidence_quote":"Supplies the transfer principle from two-term spectra to exact $L_2$-small-ball asymptotics and the analysis of non-separated boundary conditions."},{"cited_title":"On a set of transformations of Gaussian ra ndom functions","cited_arxiv_id":null,"evidence_quote":"Classifies critical versus non-critical finite-dimensional perturbations, used for the modified fractional Slepian process."},{"cited_title":"Some lemmata on the perturbation of the spectrum","cited_arxiv_id":"1908.09365","evidence_quote":"Contains the abstract perturbation lemma quoted as Proposition 3.1, which carries the p-independence result."},{"cited_title":"Exact small ball beh avior of integrated Gaussian processes under L2-norm and spectral asymptotics of boundary value problems","cited_arxiv_id":null,"evidence_quote":"Provides the general exact-small-ball theorem used in Section 5 and the boundary-order parameter idea that motivates $\\kappa$."}],"review_version":1}