{"id":"aa0e0494-e7db-43a0-980a-a82b43066039","arxiv_id":"1908.01392","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Discrete analogs of the Kontorovich-Lebedev transform are defined, with expansion theorems for sequences and functions, applied to a Dirichlet problem for the Helmholtz equation.","lead":"This paper introduces discrete versions of a classical mathematical transform that uses modified Bessel functions, and proves new expansion formulas for sequences and functions. It also solves a Dirichlet boundary value problem for the Helmholtz equation in the upper half-plane using these expansions.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 7's (2.41) is proved for f(x)=Σ a_m K_{ix}(m), but stated for f(x)=Σ a_m K_{im}(x); for f=K_i(x) the integral in (2.41) diverges, so the theorem as written is false.","rationale":"The paper's main framework, the biorthogonality Lemma 1 and the sequence expansions (2.30)-(2.33), are supported by explicit estimates and look sound. The Abel-regularized expansion (2.34) is justified by the bound in Theorem 4 and formula (3.12); the terse 'left to the reader' proof of (2.35) can probably be completed for the stated condition (3.9) by using the truncated-integral estimates from Lemma 1. The serious problem is Theorem 7. The statement says f is given by (3.3), which is f(x)=Σ a_m K_{i m}(x), but the proof uses f(τ)=Σ a_m K_{iτ}(m). The biorthogonality (2.24) only supports the latter class. For f(τ)=K_i(τ) (a_1=1) the integral in (2.41) has integrand growing like τ^{-3/2}e^{(π/2-1)τ}, so it is not even Abel-convergent; hence the theorem as stated is false. The fix is to restate the second part of Theorem 7 for the class f(x)=Σ a_m K_{i x}(m), after which the displayed proof goes through. Because this is a main expansion theorem, the paper needs this correction; the BVP application via (2.36) appears unaffected. I therefore keep a CONDITIONAL verdict, but the condition is the correction of Theorem 7, not merely the filling of Abel-limit details.","tokens_in":14756,"tokens_out":33798,"duration_ms":323322,"concrete_test":"Carry out the asymptotic check for the one-term case: with a_1=1, a_m=0 (m≥2), f(τ)=K_i(τ), and for n=1 form I(R)=∫_1^R τ sinh(πτ) K_s(1,iτ) K_i(τ)dτ. Use K_i(τ)∼(π/(2τ))^{1/2}e^{-τ} and (3.10) to get K_s(1,iτ)∼C e^{-πτ/2}/τ^2; then |I(R)| grows like C' R^{-3/2}e^{(π/2-1)R}, so the improper integral in (2.41) does not exist. Also test the Abel version with α close to π/2; for α>1 its integrand grows as e^{(α-1)τ}, so the Abel limit is undefined. This would settle that Theorem 7 as written is false and that the proof's class is the corrected one.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The most load-bearing defect is in Theorem 7, second part, not the omitted details in Theorem 4. Equation (3.3), to which the theorem refers, defines f(x)=Σ_{m=1}^∞ a_m K_{i m}(x), i.e. K_{i m} with index im and argument x. But the proof of (2.41) substitutes f(τ)=Σ_m a_m K_{iτ}(m), i.e. index iτ and argument m. These classes are not the same. The substitution is the one that makes biorthogonality (2.24) applicable. With the stated class, the theorem is actually false: take a_1=1 and all other a_m=0, so f(τ)=K_i(τ). Then as τ→∞, K_i(τ)∼(π/(2τ))^{1/2}e^{-τ}, while (3.10) gives K_s(1,iτ)∼C e^{-πτ/2}/τ^2. The integrand of (2.41) therefore behaves like C' τ^{-3/2}e^{(π/2-1)τ}, which does not decay. The quoted Abel sense (2.34) does not rescue it: for α>1 the Abel integrand behaves like C' τ^{-3/2}e^{(α-1)τ}, also divergent. Hence (2.41) cannot hold for f=K_i(τ), although this f lies in the stated class. The proof establishes (2.41) for the corrected class f(x)=Σ_m a_m K_{i x}(m), and the statement should be changed accordingly.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces discrete analogs of the classical Kontorovich-Lebedev transform, based on series involving Macdonald functions K_{in}(x) and incomplete Bessel functions J(x,in,π), Kc(x,in,π), Ks(n,iτ,sinh^{-1}(π)). The main results are expansions (2.30)-(2.41) for sequences and functions, a biorthogonality lemma (Lemma 1), and an application to solving a Dirichlet boundary value problem for the inhomogeneous Helmholtz equation in the upper half-plane. The proofs rely on integral representations, standard integral formulas, and estimates such as Lebedev's inequality.","tokens_in":15110,"tokens_out":7476,"duration_ms":69906,"significance":"If the stated results are correct, the paper provides a useful discrete counterpart of a well-known integral transform, with explicit expansion formulas and a PDE application. The author gives detailed estimates in several proofs, uses machine-checkable standard integral formulas, and proves the biorthogonality relations with nontrivial limiting arguments. However, one main theorem is false as stated, and another expansion is left essentially unproved, so the paper needs substantial revision before it can be accepted.","major_comments":[{"comment":"The second part of Theorem 7, expansion (2.41), is false as stated. The theorem says f is given by series (3.3), i.e. f(x)=∑_{m=1}^∞ a_m K_{im}(x), but the proof substitutes f(τ)=∑_{m=1}^∞ a_m K_{iτ}(m). These are different functions: in the first the index is the discrete integer m and the argument is x; in the second the index is iτ and the argument is m. The proof establishes (2.41) for the corrected class f(x)=∑_{m=1}^∞ a_m K_{i x}(m). For the stated class, take a_1=1 and all other a_m=0, so f(τ)=K_i(τ). As τ→∞, K_i(τ)∼(π/(2τ))^{1/2}e^{-τ}, while by (3.10) K_s(1,iτ,sinh^{-1}(π))=O(e^{-πτ/2}/τ^2). Thus the integrand in (2.41) behaves like C τ^{-3/2} e^{(π/2-1)τ}, which diverges; the Abel sense (2.34) does not rescue it because the factor cosh(ατ) makes the divergence worse for α>1. The statement and proof must be reconciled, and the theorem should be corrected to the class for which the proof actually works.","section":"Theorem 4, proof of (2.35)"},{"comment":"Expansion (2.35) is one of the main discrete analogues, but its proof is explicitly left to the reader: after invoking biorthogonality (2.24) and (2.28), the author states only that the order of integration and summation is justified and then writes \"The completion of the proof is left to the reader.\" This is a load-bearing gap, because (2.35) involves an improper integral over τ, an infinite series, and the interplay of the limiting procedures used in the proof of Lemma 1. A rigorous verification of these interchanges, with explicit estimates, is necessary to establish (2.35).","section":"Theorem 7, proof of (2.40)-(2.41)"},{"comment":"In the proof of the first part of Theorem 7 and in the (intended) proof of (2.41), the author asserts that \"all interchanges of the summation, integration and the passage to the limit are allowed via the absolute and uniform convergence,\" but no uniform bounds are provided. In particular, in the chain leading to (2.41) one must justify exchanging the outer sum over n, the inner series over m, the improper τ-integral, and the Abel limit α→π/2−. These are nontrivial because the kernel Ks and the Macdonald functions have oscillatory and exponential behavior; the earlier proof of a similar interchange in Lemma 1 required a detailed asymptotic argument. Without explicit verification, the expansions (2.40) and (2.41) are not fully established.","section":"Theorem 4, statement"},{"comment":"The statement of Theorem 4 says the general term can be expanded \"with respect to (2.33)\" with convergence in the Abel sense, but the proof establishes (2.34). This appears to be a typo: it should refer to (2.34). While not a technical error by itself, it confuses which expansion is being proved and should be fixed.","section":"Theorem 7 proof, summation index"},{"comment":"In the proof of Theorem 7, after Eq. (3.12), the same symbol n is used for the outer summation index and for the inner summation index in expressions such as \"∑_{n=1}^∞ a_m\"; this should be m. The repeated use of n obscures the argument and should be corrected.","section":"Abstract"}],"minor_comments":[{"comment":"The abstract contains a spacing typo: \"tra nsforms\" should be \"transforms\".","section":"Introduction, displayed formulas"},{"comment":"In Eq. (1.1), the factor 2/(π^2 τ sinh(πτ)) is written without a space; the typesetting is otherwise clear, but this should be cleaned up to avoid ambiguity between τ sinh(πτ) and τ·sinh(πτ).","section":"Section 2, Eq. (2.24)"},{"comment":"The proof of Lemma 1 uses the value of the improper integral (2.29) from reference [4], but the passage to the limit under the integral in (2.28) is justified with an asymptotic argument; the wording \"The problem now is to motivate...\" is somewhat informal and should be rephrased to state directly that the justification is provided in the following lines.","section":"Section 3, Theorem 4"},{"comment":"In the proof of Theorem 4, the estimate ∫_0^∞ cosh(ατ)/√(sinh(πτ)) dτ < ∞ for 0<α<π/2 is asserted without a reference; a one-line justification would be helpful.","section":"Section 4, Eq. (4.5)"},{"comment":"In Theorem 8, the function u(r,θ) is written as u(r,θ)=2/π^2 ∑ n sinh(θn) J(r,in,π) a_n; this is consistent with the text, but in the boundary condition (4.7) the notation f(r) is used both for the boundary data and for the function in Theorem 5; this is not a mathematical error but could be clarified.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The false statement in Theorem 7 is serious, but it is localized and fixable by changing the function class in the statement to match the proof (f(x)=∑ a_m K_{ix}(m)). The omitted proof of (2.35) is also fixable with additional estimates. I do not see signs of circularity or novelty problems. The paper fits the scope of math.CA."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does something genuinely new: it builds discrete analogs of the Kontorovich–Lebedev transform from Fourier series over incomplete Bessel functions, and proves biorthogonality relations (2.22)–(2.24) that are not in the literature. The expansions (2.30)–(2.33) and (2.36)–(2.39) are derived with substantial estimates, and the Helmholtz application in Section 4 is a legitimate concrete payoff. The proof of Lemma 1 is a bit terse, but the key asymptotic argument is there.\n\nThe trouble is Theorem 7. As stated, the second part says expansion (2.41) holds for f given by series (3.3), i.e. f(x) = Σ a_m K_{im}(x). But the proof substitutes f(τ) = Σ a_m K_{iτ}(m). These are different classes. With a_1 = 1 and the rest zero, f(τ) = K_i(τ), and the integral in (2.41) diverges: the integrand grows like τ^{-3/2} e^{(π/2−1)τ}. The Abel regularization with cosh(ατ) does not help because for α < π/2 the exponent α + π/2 − 1 is still positive as α → π/2−. So the theorem as written is false. The proof does establish (2.41) for the corrected class f(τ) = Σ a_m K_{iτ}(m), which is the class matching the biorthogonality (2.24). That is a serious fix, not a typo.\n\nSeparately, the proof of (2.35) is explicitly left to the reader, and the Abel interchanges in Theorem 4 are asserted rather than shown. Those are presentation gaps, but based on the surrounding estimates they look fillable. The Theorem 7 issue is not fillable; the statement has to change.\n\nWho is this for? Harmonic analysts and anyone working on index transforms or special functions for boundary value problems. The paper deserves a serious referee, but the referee should be asked to check Theorem 7 carefully. Right now I would treat the paper as conditionally correct: the core constructions are valuable, but the claimed expansion (2.41) for the stated class is false.\n\nRecommendation: send to peer review, but require the author to correct Theorem 7 and verify the affected Abel interchanges.","headline":"New discrete Kontorovich–Lebedev expansions with real potential, but Theorem 7's second part as stated is false—the proof proves a different expansion.","tokens_in":15641,"tokens_out":2882,"would_cite":false,"duration_ms":25399,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["45A05","44A15","42A16","33C10"],"pacs":[],"model":"deepseek-v4-flash","headline":"Discrete analogs of the Kontorovich-Lebedev transform are introduced, with expansions for sequences and functions in Macdonald and incomplete Bessel functions, and an application to the Helmholtz equation.","keywords":["Kontorovich-Lebedev transform","modified Bessel function","Macdonald function","incomplete Bessel function","Fourier series","Dirichlet problem","Helmholtz equation","index transform"],"falsifier":"Take the sequence $a_m=e^{-m}$, which satisfies the paper's decay conditions, and numerically evaluate both sides of expansion (2.35) for $n=1$ and a fixed $x>0$: if the improper integral over $\\tau$ of $\\tau\\sinh(\\pi\\tau)K_{i\\tau}(1)\\sum_m a_m K_s(m,i\\tau,\\sinh^{-1}\\pi)$ does not converge to $a_1$, the claimed inversion fails; a similar check of the limit $\\alpha\\to\\pi/2^-$ in (2.34) would settle the Abel-regularized case.","tokens_in":14530,"feed_emoji":"","tokens_out":12662,"duration_ms":98070,"temperature":0.7,"pith_summary":"The paper introduces discrete analogs of the classical Kontorovich-Lebedev transform, replacing the continuous index integral by series over Macdonald functions $K_{in}(x)$ paired with incomplete Bessel functions $J(x,in,\\pi)$, $K_c(x,in,\\pi)$, and $K_s(n,i\\tau,\\sinh^{-1}(\\pi))$. It establishes biorthogonality of these kernel sequences under suitable measures and proves a family of expansions, formulas (2.30)-(2.41), recovering a sequence term $a_n$ or a function $f(x)$ from such series and integrals. A sympathetic reader would care because these formulas give a genuinely discrete, Fourier-style calculus for the index transform, and the paper demonstrates the payoff by solving a Dirichlet boundary-value problem for the inhomogeneous Helmholtz equation in the upper half-plane explicitly.","feed_headline":"Discrete Kontorovich-Lebedev transforms yield new inversion formulas","feed_subtitle":"The new expansions apply directly to solving a Dirichlet problem in the upper half-plane.","key_machinery":"The load-bearing objects are the incomplete modified Bessel function $J(z,\\nu,w)=\\int_0^w e^{-z\\cosh u}\\cosh(\\nu u)\\,du$ and its trigonometric siblings $K_c(x,\\nu,w)$ and $K_s(x,\\nu,w)$, defined by cutting the standard integral representations (1.3)-(1.4) at $w$. These sit between the full Macdonald function $K_\\nu(z)$ and Fourier series: cutting the integrals makes $u\\in[0,\\pi]$ a natural domain on which the Fourier-series identities (2.9)-(2.13) hold pointwise. The biorthogonality relations (2.22)-(2.24), obtained by combining those Fourier expansions with classical integral evaluations (2.26), (2.27), and (2.29), are what turn the series-and-integral expressions into inversion formulas.","core_discovery":"The central claim is that the classical Kontorovich-Lebedev inversion can be discretized: for suitable sequences $\\{a_n\\}$ and functions $f$, formulas (2.30)-(2.35) recover $a_n$ from integrals against $\\sum_m a_m K_{im}(x)$ or $\\sum_m a_m J(x,im,\\pi)$, and formulas (2.36)-(2.41) recover $f(x)$ from series over $J(x,in,\\pi)$, $K_c(x,in,\\pi)$, $K_{ix}(n)$, or $K_s(n,ix,\\sinh^{-1}\\pi)$. The proof rests on biorthogonality of the kernel sequences $\\{K_{in}(x)\\}$ with $\\{J(x,in,\\pi)\\}$ and $\\{K_c(x,in,\\pi)\\}$, and $\\{K_{i\\tau}(n)\\}$ with $\\{K_s(n,i\\tau,\\sinh^{-1}\\pi)\\}$, with respect to $dx/x$ and $\\tau\\sinh(\\pi\\tau)\\,d\\tau$. As an application, the paper shows that a specific series in $J(r,in,\\pi)$ solves the inhomogeneous Helmholtz equation $\\Delta u-u=h$ in the upper half-plane, vanishing at infinity and taking prescribed boundary values $u(r,0)=0$, $u(r,\\pi)=f(r)$.","pith_inferences":["Extending the paper's method, one could try to construct discrete analogs of other index transforms by replacing continuous index integration with a sum over discrete indices and using finite-window integral representations.","The same biorthogonality machinery could support spectral or quadrature methods for Helmholtz-type boundary problems in half-plane and wedge geometries, using expansions (2.36)-(2.41) as approximation bases.","A numerical check of the Abel-regularized formulas (2.34) and (2.35) on simple sequences such as $a_m=e^{-m}$ would clarify the exact convergence conditions the paper leaves implicit."],"forward_implications":["A sequence $(a_n)$ satisfying $\\sum |a_n|e^{-\\pi n/2}<\\infty$ can be recovered from the function $f(x)=\\sum_m a_m K_{im}(x)$ through the integral formula (2.30), and dually, under $\\sum |a_n|/n<\\infty$, through (2.31).","Functions represented as $f(x)=\\int_{-\\pi}^{\\pi} e^{-x\\cosh u}\\phi(u)\\,du$ with Lipschitz $\\psi$ admit the expansion (2.36) in the incomplete-Bessel kernels $J(x,in,\\pi)$.","Functions represented as $f(x)=\\int_{-\\pi}^{\\pi}\\sin(x\\sinh u)\\phi(u)\\,du$ admit the expansion (2.38) in the kernels $K_c(x,in,\\pi)$.","The series $u(r,\\theta)=\\frac{2}{\\pi^2}\\sum_{n=1}^\\infty n\\sinh(\\theta n)J(r,in,\\pi)a_n$ solves the inhomogeneous Helmholtz equation (4.3) in the upper half-plane, vanishes at infinity, and, when $a_n$ is the discrete Kontorovich-Lebedev transform of $f$, satisfies boundary conditions $u(r,0)=0$, $u(r,\\pi)=f(r)$.","Under stronger decay conditions, expansions (2.34), (2.35), (2.40), and (2.41) provide Abel-regularized and improper-sense inversions involving the $K_s$ kernels."],"supporting_citations":[{"why":"Defines the incomplete modified Bessel function $J(z,\\nu,w)$ and supplies the differential equation (2.6) used in the Helmholtz application.","marker":"[1]"},{"why":"Lebedev's original inversion formula (1.1), the classical continuous transform whose discrete analogs this paper constructs.","marker":"[2]"},{"why":"Provides the integral evaluations (2.26), (2.27), (2.29), and (3.12) and the formulas used in Examples 4-7 that carry the Fourier-series calculations.","marker":"[3]"},{"why":"Supplies the Lebedev inequality, the index-transform theory, and the asymptotics of $K_{i\\tau}$ used for convergence and limit interchanges.","marker":"[4]"}],"fun_headline_variants":["Discrete KL transforms yield new inversion formulas","Discrete Kontorovich-Lebedev inversion for Helmholtz problem","Discrete KL transforms solve Dirichlet Helmholtz boundary value problem","New discrete transforms invert Bessel series for Helmholtz","Discrete KL inversion solves Helmholtz boundary value problem"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The expansions proved via Abel regularization depend on being allowed to exchange the limit with infinite sums and integrals; the paper leaves some of those exchanges to the reader or asserts them by uniform convergence without a fully detailed verification.","fun_headline_variants_meta":{"raw":{"variants":["Discrete KL transforms yield new inversion formulas","Discrete Kontorovich-Lebedev inversion for Helmholtz problem","Discrete KL transforms solve Dirichlet Helmholtz boundary value problem","New discrete transforms invert Bessel series for Helmholtz","Discrete KL inversion solves Helmholtz boundary value problem"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000863,"raw_usage":{"total_tokens":3719,"prompt_tokens":894,"completion_tokens":2825,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":510,"completion_tokens_details":{"reasoning_tokens":2745}},"tokens_in":510,"tokens_out":2825,"duration_ms":20742,"temperature":1.0,"reasoning_tokens":2745,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:13:58.785198+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the sequence $a_m=e^{-m}$, which satisfies the paper's decay conditions, and numerically evaluate both sides of expansion (2.35) for $n=1$ and a fixed $x>0$: if the improper integral over $\\tau$ of $\\tau\\sinh(\\pi\\tau)K_{i\\tau}(1)\\sum_m a_m K_s(m,i\\tau,\\sinh^{-1}\\pi)$ does not converge to $a_1$, the claimed inversion fails; a similar check of the limit $\\alpha\\to\\pi/2^-$ in (2.34) would settle the Abel-regularized case.","supporting_citations":[{"cited_title":"Jones, Incomplete Bessel functions","cited_arxiv_id":null,"evidence_quote":"Defines the incomplete modified Bessel function $J(z,\\nu,w)$ and supplies the differential equation (2.6) used in the Helmholtz application."},{"cited_title":"Lebedev, Sur un formule d’inversion, C.R","cited_arxiv_id":null,"evidence_quote":"Lebedev's original inversion formula (1.1), the classical continuous transform whose discrete analogs this paper constructs."},{"cited_title":"Prudnikov, Yu.A","cited_arxiv_id":null,"evidence_quote":"Provides the integral evaluations (2.26), (2.27), (2.29), and (3.12) and the formulas used in Examples 4-7 that carry the Fourier-series calculations."},{"cited_title":"Yakubovich, Index Transforms, World Scientific Publishing Company, Singapore, New Jersey, London and Hong Kong, 1996","cited_arxiv_id":null,"evidence_quote":"Supplies the Lebedev inequality, the index-transform theory, and the asymptotics of $K_{i\\tau}$ used for convergence and limit interchanges."}],"review_version":1}