{"id":"8ca9e45b-5625-47c6-994a-ec1d671dcc5e","arxiv_id":"2505.23332","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The Titchmarsh-Weyl m-function is expressed via the diagonal of a Volterra integral equation, yielding a new exponential bound for the A-amplitude.","lead":"This paper connects boundary control theory of wave equations to the Titchmarsh-Weyl m-function, deriving a new integral equation for Simon's A-amplitude. It proves an exponential bound that answers a conjecture by Gesztesy and Simon, offering a new route to compute a central object in spectral theory.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The exponential bound on A is plausible, but the unproved identification of the solution of (3.2) with the response kernel / A-amplitude for L1_loc potentials is load-bearing: without it, Theorem 2 does not justify Algorithm 1.","rationale":"The reader identified the missing existence/regularity theory for L1_loc potentials as the weakest assumption, and I agree that this is the load-bearing point. The main theorem's iteration estimates are a real contribution: after correcting the orientation of the inner integral (the paper writes ∫_x^v where the intended domain has v≤x, so the orientation is negative), the induction (4.9) and Stirling estimates (4.12)-(4.13) go through, and the exponential bound is credible. The issue is not the internal bound but its target: the A-amplitude in (2.15) is defined through the spectral m-function, and the bridge via the response kernel r=w_x(0,·) is asserted rather than proved for this potential class. The Laplace-domain identities (2.8)-(2.13) require enough regularity and growth of r to justify the absolute convergence of the Laplace integrals, and that regularity is exactly what the paper does not establish independently. This is a fixable gap, not a demonstrated contradiction, so a conditional verdict remains appropriate.","tokens_in":10670,"tokens_out":18753,"duration_ms":207974,"concrete_test":"Independently derive the A-amplitude from Simon's spectral definition for a potential satisfying (4.4) but q∉L1∪L∞, e.g., q(x)=Σ_{n≥1} n 1_{[n,n+n^{-2}]}, and check that this A satisfies (3.2) on 0<y≤x; equivalently, solve (2.3) by the standard Volterra iteration in characteristic coordinates and verify (3.1) for this q. If either derivation requires q∈L1 or q∈L∞, the bound in Theorem 2 does not apply to the m-function, and Algorithm 1 is unsupported under (4.4).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is Algorithm 1: solve (3.2), take the diagonal, and Laplace transform to obtain m. The unproved link is that the diagonal of (3.2) equals the A-amplitude appearing in (2.15) for the claimed generality. Theorem 1 derives (3.2) from the Goursat problem (2.3) by formal differentiation, but for q merely in L1_loc (or satisfying (4.4)) the paper neither proves existence/uniqueness of w nor that w_x(0,·) is a locally integrable function whose Laplace transform can be identified with m via (2.12)-(2.14). Theorem 2 proves an exponential bound for the formal iteration solution of (3.2), but if that object has not been shown to coincide with -2w_x(0,2·), the bound does not imply (2.15) for the spectral m-function. The domain issue in (3.2) reinforces the gap: as stated for x,y>0, the term q(x-v) is evaluated at negative arguments whenever y>x; the intended domain is 0<y≤x, and the existence theory on that domain is not supplied.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a boundary-control approach to the Titchmarsh-Weyl m-function for the half-line Schrödinger operator with locally integrable potential. It introduces the response operator R of the wave equation (2.1), identifies its convolution kernel r with w_x(0,·), connects r to Simon's A-amplitude via (2.16), derives the linear Volterra equation (3.2) for A(x,y), and proves an exponential bound (4.5) for the diagonal A(α)=A(α,α) under the condition (4.4). It then presents Algorithm 1, which evaluates m by solving (3.2) and taking the Laplace transform (2.15).","tokens_in":10885,"tokens_out":8155,"duration_ms":86364,"significance":"If the identification between the formal solution of (3.2) and the response kernel was made rigorous, the paper would provide a new interpretation of the A-amplitude and a potentially efficient numerical procedure for the m-function. The explicit bound (4.5) with concrete constants is a substantive contribution and would answer a conjecture of Gesztesy-Simon for potentials with bounded L1 norm on unit intervals. The combination of the Goursat problem and boundary-control ideas is elegant and likely to be useful. However, the load-bearing analytic steps are currently asserted rather than proved, so the central claims are not yet established at the stated level of generality.","major_comments":[{"comment":"The representation (2.2) of the weak solution to (2.1) in terms of a solution w of the Goursat problem (2.3), and the identification r(·)=w_x(0,·), are asserted without proof. For q merely in L1_loc, w is not shown to be differentiable up to the boundary x=0, nor is w_x(0,·) shown to be a locally integrable function whose Laplace transform can be used in (2.14). This is not a technicality: Theorem 1 and representation (2.15) both rely on this identification.","section":"§2, Eq. (2.2)–(2.3), (2.6)"},{"comment":"Equation (3.2) is stated for x,y>0, but the term q(x-v) is evaluated at negative arguments whenever v>x, so the equation is not defined as written. The derivation from (3.5) is valid only on the triangular domain 0≤y≤x. The theorem should state that domain and prove existence and uniqueness of the solution A(x,y) there, for example by the Volterra iteration. Without this, the use of A(α,α) in Theorem 2 and Algorithm 1 is not justified.","section":"Theorem 1, Eq. (3.2)"},{"comment":"The proof estimates the formal Neumann series for (3.2) and concludes the bound for the A-amplitude. This conclusion requires that the actual A-amplitude defined by (2.16) satisfies (3.2) and equals the sum of the Neumann series. Under the stated assumption (4.4), neither the existence of a solution to (3.2) nor its coincidence with -2r(2α) is proved. The bound therefore applies to the formal series, not yet to the spectral object appearing in (2.15).","section":"Theorem 2, proof of (4.5) and of (2.15)"},{"comment":"Algorithm 1 instructs to evaluate m(z) by (2.15) after computing A(α). The paper proves absolute convergence of the integral in (2.15) only for Re k > 2 max{√(2||q||), e||q||}, and it does not prove the identity m(-k^2) = -k - ∫_0∞ A(α)e^{-2αk}dα for all such k; the derivation via (2.11)–(2.14) inherits the gaps described above. Moreover, for a given finite z∈C+, the k with -k^2=z may have Re k below that threshold, so the algorithm's range of applicability to finite z is not demonstrated.","section":"§5, Algorithm 1 and Remark 4"}],"minor_comments":[{"comment":"Equation (2.10) appears to contain a typographical error: the intended statement is likely \\(\\widehat{Rf}(k)=\\widehat{u_x}(0,k)\\).","section":"§2, Eq. (2.10)"},{"comment":"The integration-by-parts proof uses the symbol b^n; for n=0 and b=0 the expression 0^0 should be avoided by treating n=0 separately or by taking n≥1.","section":"Lemma 1, Eq. (4.3)"},{"comment":"In reference [12], the page range \"491–436\" appears to be printed in reverse order.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the journal scope and the main idea is attractive, but the current version contains unproved identifications that are load-bearing for the algorithm. The gaps are fixable by adding existence and uniqueness results for the Goursat problem and for (3.2) on the triangular domain, and by proving that the solution of (3.2) equals the response kernel. The authors should also clarify the range of z for which Algorithm 1 actually evaluates m(z)."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe short version: this paper has a real new result — the exponential bound (4.5) for the A-amplitude under the mild decay condition (4.4) — and the proof strategy via iteration of a Volterra equation is clean. But the paper as written has a misstated central equation and a gap in the identification step that the claims paper over. It deserves a serious referee, not a desk reject, but it needs a substantive revision.\n\nWhat is actually new: Theorem 2 gives the affirmative answer to the Gesztesy-Simon conjecture on exponential bounds for potentials with uniformly bounded L1 mass on unit intervals. The iteration argument with Lemma 1 is elementary and the bounds are plausible. The boundary-control interpretation of the A-amplitude as the response kernel is nice and likely useful. The algorithm (solve (3.2) once, Laplace transform for all z) is appealing.\n\nNow the soft spots, in proportion. First, equation (3.2) is not correctly stated. As printed, for x,y>0 the potential q(x−v) is evaluated at negative arguments whenever y>x, so the domain must be 0<y≤x. More importantly, the inner integral is written ∫_x^v, but the derivation from the Goursat problem gives ∫_v^x. The sign matters: for constant q, the printed equation yields A(α) ≈ c + c^2 α^2/2, while the correct equation gives c − c^2 α^2/2, which matches the actual m-function expansion. So Theorem 1's statement has a sign error, even if the intended equation is recoverable from the proof.\n\nSecond, and this is the load-bearing issue the stress-test flags: Theorem 2 proves a bound for the formal iteration solution of (3.2), but the paper does not prove that this object is the A-amplitude in the claimed generality. The derivation of (3.1) requires existence and differentiability of the solution w to the Goursat problem; for potentials only in L1_loc this is asserted by 'direct computation' but not supplied. The identification of the diagonal of the integral equation with the response kernel is formal. Without that identification, the bound does not yet apply to the spectral m-function.\n\nThird, the Laplace representation (2.15) is shown absolutely convergent only for large Re k, so Algorithm 1's claim to evaluate m at finite z is stronger than what is proved.\n\nNone of this kills the core idea. The exponential bound is probably correct after the typos are fixed, and the intended equation is clear. For a reader in spectral/inverse problems, the paper is worth engaging with.\n\nMy recommendation: send it to peer review. A referee should ask for a corrected (3.2), a proper well-posedness and identification argument for the integral equation under (4.4), and a more modest statement of the algorithm's scope.","headline":"A genuinely new bound for the A-amplitude under weak decay, buried under a misstated integral equation and an unproved identification step; worth refereeing after revision.","tokens_in":11470,"tokens_out":14695,"would_cite":true,"duration_ms":136949,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["34B20","34E05","34L25","34E40","47B20","81Q10"],"pacs":[],"model":"deepseek-v4-flash","headline":"One integral equation computes the Titchmarsh-Weyl m-function","keywords":["Schrödinger operator","Titchmarsh-Weyl m-function","boundary control","A-amplitude","response operator","Dirichlet-to-Neumann map","Volterra integral equation","inverse spectral theory"],"falsifier":"For a real potential satisfying (4.4) but with $\\int|q|=\\infty$ (for instance $q(x)=\\sin(x^2)$), solve (3.2) by iteration to obtain $A(\\alpha)=A(\\alpha,\\alpha)$, insert it into (2.15), and compare $m(-k^2)$ with a direct high-precision solution of (1.10)–(1.4) at a fixed $k$ with $\\operatorname{Re}k>2\\max\\{\\sqrt{2\\|q\\|},e\\|q\\|\\}$; a disagreement beyond the bound (4.5) would show the integral equation does not yield the true A-amplitude.","tokens_in":10452,"feed_emoji":"📐","tokens_out":12605,"duration_ms":110686,"temperature":0.7,"pith_summary":"This paper shows that the Titchmarsh-Weyl m-function of a Schrödinger operator on the half-line, which encodes the spectral measure and is difficult to compute at finite spectral values, can be evaluated through the controlled wave equation associated with the operator. The m-function is the Laplace transform of the kernel of the Dirichlet-to-Neumann map (the response operator) of that wave equation, and that kernel is, up to scaling, the A-amplitude of the Gesztesy–Simon representation. The paper proves that the A-amplitude is the diagonal value of the solution of a linear Volterra integral equation, yielding an algorithm: solve once for the amplitude, then take Laplace transforms for each desired point. It also proves an explicit exponential bound for the A-amplitude under the condition that the L1-norm of the potential on unit intervals is uniformly bounded, answering a conjecture of Gesztesy and Simon.","feed_headline":"One integral equation computes the Titchmarsh-Weyl m-function","feed_subtitle":"Solve once for the A-amplitude, then take Laplace transforms; the method works for potentials that never decay.","key_machinery":"The central object is the response operator $R$ (the dynamic Dirichlet-to-Neumann map) of the wave equation associated with the Schrödinger operator, together with its kernel $r(s)=w_x(0,s)$, defined through the solution $w$ of the Goursat problem $w_{ss}-w_{xx}+q(x)w=0$, $w(x,0)=0$, $w(x,x)=-\\frac12\\int_0^x q$. The A-amplitude is the rescaled kernel $A(\\alpha)=-2r(2\\alpha)$. The argument's engine is the linear Volterra integral equation (3.2) for $A(x,y)$, obtained by a chain of variable changes in the Goursat problem; its diagonal value $A(\\alpha,\\alpha)$ is the A-amplitude. Iteration of the Volterra operator $K$, estimated with Lemma 1 and Stirling's bound, produces the explicit exponential estimate (4.5).","core_discovery":"The paper establishes that the Dirichlet Titchmarsh–Weyl m-function associated with $H=-\\partial_x^2+q(x)$ on $L^2(0,\\infty)$ is the Laplace transform of the response function of the wave equation $u_{tt}-u_{xx}+q(x)u=0$ with zero initial data and boundary control $u(0,t)=f(t)$. Concretely, $m(-k^2)=-k+\\int_0^\\infty e^{-k\\alpha}r(\\alpha)\\,d\\alpha$, and equivalently $m(-k^2)=-k-\\int_0^\\infty A(\\alpha)e^{-2\\alpha k}\\,d\\alpha$, where $A(\\alpha)=-2r(2\\alpha)$ is the A-amplitude. The main new results are: first, $A(\\alpha)$ equals the diagonal value $A(\\alpha,\\alpha)$ of the solution of the linear Volterra integral equation $A(x,y)=q(x)-\\int_0^y\\big(\\int_x^v A(u,v)\\,du\\big)q(x-v)\\,dv$, giving a direct algorithm to evaluate $m$ by solving once and then computing a Laplace transform; and second, under the condition $\\|q\\|:=\\sup_x\\int_x^{x+1}|q(s)|\\,ds<\\infty$, the error bound $|A(\\alpha)-q(\\alpha)|\\le \\frac12\\big(\\int_0^\\alpha|q|\\big)^2\\big[e^{2\\sqrt{2\\|q\\|\\alpha}}+\\tfrac1{\\sqrt{2\\pi}}e^{2e\\|q\\|\\alpha}\\big]$ holds, yielding absolute convergence of the Laplace integral for $\\operatorname{Re} k>2\\max\\{\\sqrt{2\\|q\\|},e\\|q\\|\\}$ and settling the conjecture that the A-amplitude has an exponential bound for such potentials.","pith_inferences":["Inverting the integral equation (3.2) would recover the potential $q$ from the boundary response kernel, suggesting a new inverse spectral algorithm for the half-line that bypasses Gelfand–Levitan–Marchenko machinery.","The Dirichlet-to-Neumann identification of the m-function is naturally multidimensional, so the same response-operator route could define operator-valued m-functions for $-\\Delta+q$ in higher dimensions; the paper notes this possibility in Remark 2 but does not develop it.","The exponential estimate (4.5) doubles as an error bound for truncated Neumann series, giving a stopping criterion for numerical evaluation of the m-function when $\\|q\\|$ is moderate.","If the iteration of (3.2) converges for potentials outside the class (4.4), the Laplace representation may hold in a larger region, so the convergence condition rather than the Volterra equation itself is the current bottleneck."],"forward_implications":["The m-function can be evaluated by solving one linear Volterra equation (3.2) and taking Laplace transforms; once $A(\\alpha)$ is known, every new spectral point costs one transform rather than a separate ODE solve.","The A-amplitude acquires a concrete physical meaning: it is (minus twice) the impulse response kernel of the Dirichlet-to-Neumann map for the wave equation.","The representation (2.15) and its absolute convergence are extended to all potentials with uniformly bounded local $L^1$-norm, including potentials with no decay at infinity.","For nonnegative potentials the terms of the alternating series (5.1) are nonnegative, so the algorithm's series converges faster and truncation is easier to control."],"supporting_citations":[{"why":"introduced the A-amplitude formalism that representation (2.15) continues and extends.","marker":"[22]"},{"why":"proved (2.15) for L1 and L∞ potentials, defined the A-amplitude, and conjectured the exponential bound that Theorem 2 settles.","marker":"[9]"},{"why":"supplied the boundary control method and response operator framework used to derive representation (2.2)-(2.6).","marker":"[3]"},{"why":"extended the boundary control approach to non-selfadjoint Sturm-Liouville operators, grounding the response-operator techniques the paper uses.","marker":"[2]"},{"why":"established equivalence of time-domain inverse problems and boundary spectral problems, supporting the spectral/time-domain correspondence of Section 2.","marker":"[12]"},{"why":"identified the Titchmarsh-Weyl m-function as a one-dimensional spectral Dirichlet-to-Neumann map, the key motivation for the response-operator interpretation.","marker":"[17]"}],"fun_headline_variants":["Volterra equation yields Titchmarsh-Weyl m-function directly","Boundary control connects wave amplitudes to m-functions","No decay needed: new route to compute m-functions","Solve once for A-amplitude, then Laplace gives m","From boundary control to m-function: a single solve"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole construction assumes that for every $q\\in L^1_{\\rm loc}$ the wave-equation solution has the representation (2.2) with $w$ solving the Goursat problem (2.3) and that $w_x(0,\\cdot)$ is the response kernel; the paper does not prove existence or uniqueness of such $w$ for general $L^1_{\\rm loc}$ potentials, and the convergence of the Volterra iteration is established only under the stronger uniform one-step $L^1$ bound (4.4).","fun_headline_variants_meta":{"raw":{"variants":["Volterra equation yields Titchmarsh-Weyl m-function directly","Boundary control connects wave amplitudes to m-functions","No decay needed: new route to compute m-functions","Solve once for A-amplitude, then Laplace gives m","From boundary control to m-function: a single solve"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000423,"raw_usage":{"total_tokens":2206,"prompt_tokens":1015,"completion_tokens":1191,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":631,"completion_tokens_details":{"reasoning_tokens":1108}},"tokens_in":631,"tokens_out":1191,"duration_ms":11317,"temperature":1.0,"reasoning_tokens":1108,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T12:48:06.063782+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a real potential satisfying (4.4) but with $\\int|q|=\\infty$ (for instance $q(x)=\\sin(x^2)$), solve (3.2) by iteration to obtain $A(\\alpha)=A(\\alpha,\\alpha)$, insert it into (2.15), and compare $m(-k^2)$ with a direct high-precision solution of (1.10)–(1.4) at a fixed $k$ with $\\operatorname{Re}k>2\\max\\{\\sqrt{2\\|q\\|},e\\|q\\|\\}$; a disagreement beyond the bound (4.5) would show the integral equation does not yield the true A-amplitude.","supporting_citations":[{"cited_title":"Simon,A new approach to inverse spectral theory","cited_arxiv_id":null,"evidence_quote":"introduced the A-amplitude formalism that representation (2.15) continues and extends."},{"cited_title":"Gesztesy and B","cited_arxiv_id":null,"evidence_quote":"proved (2.15) for L1 and L∞ potentials, defined the A-amplitude, and conjectured the exponential bound that Theorem 2 settles."},{"cited_title":"Avdonin, M.I","cited_arxiv_id":null,"evidence_quote":"supplied the boundary control method and response operator framework used to derive representation (2.2)-(2.6)."},{"cited_title":"Avdonin and M.I","cited_arxiv_id":null,"evidence_quote":"extended the boundary control approach to non-selfadjoint Sturm-Liouville operators, grounding the response-operator techniques the paper uses."},{"cited_title":"Kachalov, Y","cited_arxiv_id":null,"evidence_quote":"established equivalence of time-domain inverse problems and boundary spectral problems, supporting the spectral/time-domain correspondence of Section 2."},{"cited_title":"Pavlov,S-Matrix and Dirichlet-to-Neumann Operators, In: Encyclopedia of Scattering, ed","cited_arxiv_id":null,"evidence_quote":"identified the Titchmarsh-Weyl m-function as a one-dimensional spectral Dirichlet-to-Neumann map, the key motivation for the response-operator interpretation."}],"review_version":1}