{"id":"e4b92a1c-8efd-451e-a4e8-d0cd6f5e9cdb","arxiv_id":"2505.07292","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"The number of exponentially small singular values of the weighted ∂-bar operator obeys a Weyl-type asymptotic law whose coefficient is computed from the solution of a double obstacle problem.","lead":"A team of mathematicians proves a precise formula for the number of exponentially small singular values of a semiclassical ∂-bar operator on a torus, expressed through a free boundary problem for an auxiliary weight. The result is a Weyl-type law with a sharp correction term when the decay rate is small.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 4 rests on the non-degeneracy hypothesis (1.10): without it, the free-boundary measure-zero property and the strict positivity of Δφ on the contact set are unproved, so the matching upper and lower bounds may break.","rationale":"I reviewed the flow from Theorems 1–4. The proofs of the upper and lower bounds are internally consistent: the Carleman estimates, Bergman projection and trace-class arguments in Section 3 are standard and the use of the C^{1,1} optimal weight is legitimate. The weakest point is the passage through the free-boundary regularity in Section 5: the equality of the upper and lower constants depends on the contact set having nonempty interior and lying strictly inside {Δφ>0}, which are obtained via Proposition 5.4 and Theorem 8, both relying essentially on the non-degeneracy (1.10). This is an explicit hypothesis, so the theorem as stated is probably correct, but it is the narrowest assumption and no argument is provided for the degenerate case. Separately, the metadata abstract claims results on a compact Riemann surface while the body only treats the torus and explicitly defers Riemann surfaces to future work (Section 1.6); this overclaim should be corrected. The reader's CONDITIONAL verdict already captures this presentation issue; my mathematical concern does not change that verdict.","tokens_in":60481,"tokens_out":32599,"duration_ms":313992,"concrete_test":"Take φ(x,y)=sin x sin y on T² (so Δφ=-2 sin x sin y and dΔφ=0 at the four crossing points of Γ). Solve the double obstacle problem (1.11) numerically, e.g., by the penalization scheme of §5.1 on a fine grid, and measure vol(∂M±(ψ)) and the Weyl constant ∫_{M+(ψ)}Δφ. Then compute N([0,e^{-τ/h}]) directly by discretizing P*P for a small h and τ=0.05. If the free-boundary volume is non-negligible or the counting function deviates from the predicted (1/2πh)∫_{M+(ψ)}Δφ, the concern lands; if both match, (1.10) is not necessary and the proof might be extendable.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Theorem 4) derives from matching upper and lower bounds in Theorems 1 and 2 via an optimal weight ψ that is both superharmonic on {ψ<φ} and subharmonic on {ψ>φ-τ}. The matching requires (i) M+(ψ)⊆{Δφ>0} and (ii) vol(∂M±(ψ))=0, so that int(M+(ψ)) is nonempty and the strict subharmonicity holds in a neighborhood of the contact set. Both facts are proved in Section 5 only under the hypothesis dΔφ≠0 along Γ=(Δφ)^{-1}(0). Proposition 5.4 uses Hopf's lemma to show M±(ψ)∩Γ=∅, and Theorem 8 then uses the positive distance so obtained to prove that the free boundaries are porous, hence of measure zero. If dΔφ vanishes at a point of Γ (e.g., Δφ behaves like xy near a crossing), the Hopf argument fails at that point; the contact set may touch Γ, and no porosity bound is available. Without a measure-zero free boundary, int(M+(ψ)) could be empty even if vol(M+(ψ))>0, so Theorem 1 cannot be applied and the lower bound in Theorem 2 may not match. Thus the stated asymptotic (1/2πh)∫_{M+(ψ)}Δφ is only as secure as (1.10); the paper gives no fallback for the degenerate case, and a different leading term may arise.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the semiclassical Cauchy-Riemann operator P = h∂_z + ∂_z φ on exponentially weighted L^2 spaces over the two-dimensional torus, focusing on the number N([0,e^{-τ/h}]) of singular values in an exponentially small interval. Theorems 1 and 2 provide upper and lower bounds on this counting function in terms of auxiliary upper and lower bound weights satisfying suitable (super/sub)harmonicity conditions. Under the non-degeneracy hypothesis dΔφ≠0 on (Δφ)^{-1}(0), Theorem 3 constructs a unique C^{1,1} solution of a double obstacle problem, whose contact sets have positive measure, lie in the regions ±Δφ>0, and have measure-zero free boundaries. Theorem 4 derives a Weyl law with leading term (2πh)^{-1}∫_{M+(ψ)}Δφ dL. Theorem 5 gives a refined expansion when τ=ε^3 bτ is small: the leading term is expressed through the positive region of Δφ minus a thin band around the zero curve, with an explicit τ^{2/3} correction. The proofs combine Hörmander estimates, asymptotic Bergman projections, trace computations, a penalization argument for the obstacle problem, and a two-scale pseudodifferential analysis of the Dirichlet problem in a thin band.","tokens_in":60671,"tokens_out":23619,"duration_ms":245629,"significance":"If the results are correct, Theorem 4 is a genuinely parameter-free Weyl law for exponentially small singular values of a non-self-adjoint operator: the leading constant is determined by the obstacle problem and contains no fitted parameter. The paper is also valuable for its self-contained treatment of double obstacle problems on compact manifolds without boundary, including existence, uniqueness, C^{1,1} regularity, and porosity of the free boundary. The thin-band analysis in Section 4 is a technically substantial construction of near-optimal weights. The upper and lower bound framework of Theorems 1 and 2 is clean and likely to be reusable. The main concern is a gap in the final step of the proof of Theorem 5, discussed below; Theorem 4 itself appears sound under its stated hypotheses.","major_comments":[{"comment":"The final step of the proof of Theorem 5 is not justified as written. Proposition 4.6 and (4.117) are estimates for N([0,e^{-eτ/h}]) with eτ=τ−Cε^N<τ. The text then shows that Vol(Ω_f^+(eτ)) can be replaced by Vol(Ω_f^+(τ)) up to O(ε^{N-1}), and immediately writes 'we get N([0,e^{-τ/h}]) = ...' in (4.118). This replaces the decay rate in the counting interval without argument. Since e^{-eτ/h}=e^{-τ/h}e^{Cε^N/h}, the larger interval is exponentially wider, and a lower bound for the larger interval does not imply the same lower bound for the smaller interval. Moreover, on Ω− the function ψlb defined in (4.92) satisfies ψlb=φ−eτ>φ−τ and Δψlb=Δφ<0, so ψlb is not a lower bound weight for the original τ, and the Lipschitz clause of Theorem 2 cannot be invoked to transfer the lower bound to the interval [0,e^{-τ/h}]. The authors must either prove that the shell (e^{-τ/h}, e^{-eτ/h}] contains only o(1/h) singular values, construct lower bound weights directly for τ, or explicitly restate Theorem 5 with the decay rate eτ and verify that the leading term is unchanged up to the stated error.","section":"Section 4, equations (4.117)–(4.118)"}],"minor_comments":[{"comment":"The abstract speaks of a compact Riemann surface, while the main body works on the two-dimensional torus and Section 1.6 explicitly lists the extension to general compact Riemann surfaces as future work. The abstract should be aligned with the actual scope.","section":"Abstract and Section 1.1"},{"comment":"There is a missing closing parenthesis in the displayed statement: 'N([0,e−τ/h]' should read 'N([0,e^{-τ/h}])'.","section":"Theorem 1 statement"},{"comment":"In the displayed formula (1.13), the notation 'O(ε^{(N+1)/2})' and the subsequent 'o(1)' are not clarified regarding uniformity in ε; the proof later states that the O-term is uniform and the o(1) is not. This distinction should be stated in the theorem itself.","section":"Section 1.4, Theorem 5"},{"comment":"The phrase '∂y∆φ> 0, when y = 0' is slightly ambiguous because Δφ is a function on the torus; it would be clearer to write ∂_y(Δφ)>0 on γ, where y is the normal coordinate in the almost holomorphic extension.","section":"Section 4, around (4.3)"},{"comment":"The constants C0 and C in the statement are said to depend only on M and g, but the proof also uses ∥Δφ−∥_{L∞} and the C² norm of φ−; the statement should mention this dependence explicitly.","section":"Section 5, Proposition 5.6"},{"comment":"The sentence 'writing τ instead of eτ' is ambiguous: if the theorem's τ is redefined to be eτ at the end of the proof, the hypotheses τ=ε^3 bτ with bτ≍1 and the volume formula (1.14) still hold with bτ changed by O(ε^{N-3}), but this change of notation should be stated explicitly rather than left implicit.","section":"Section 4, final paragraph"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this paper. First, it is a serious, technically substantial piece of work that proves the first general Weyl asymptotics for exponentially small singular values of the semiclassical ∂-bar operator on the torus, going well beyond the earlier sin(y) separation-of-variables example. Second, the metadata abstract says \"compact Riemann surface\" but the body proves everything only on the two-dimensional torus; the authors themselves say the Riemann surface case is coming. That is a real overclaim, but it is cosmetic and easily fixed.\n\nThe main result, Theorem 4, holds up on my reading. The upper and lower bounds are matched through Carleman estimates and the double obstacle problem, and the optimal weight is constructed under the explicit non-degeneracy assumption d∆φ ≠ 0 along the zero set of ∆φ. The sharp small-τ correction in Theorem 5, with its τ^{2/3} term, is a genuine new result. The proof structure is coherent: Section 3 does the trace computations, Section 4 does the thin-band analysis, Section 5 gives a self-contained treatment of the double obstacle problem on compact manifolds. The C^{1,1} regularity, the measure-zero free boundaries, and the positivity of the contact sets are all proved in a way that uses the non-degeneracy assumption exactly where it is needed. I found no circularity and no fitted parameters.\n\nThe stress-test concern about hypothesis (1.10) does not land as a flaw on my reading. Yes, the theorem relies on that assumption, and without it the free-boundary measure-zero property and the strict positivity on the contact set are not proved. But the paper states (1.10) as an explicit hypothesis and never pretends to cover the degenerate case. That is a limitation, not a hidden error. A more serious soft spot is the abstract overclaim: the arXiv metadata promises compact Riemann surfaces, while the body only proves the torus case. The authors mention in the introduction that the extension is forthcoming, so the correct fix is to align the abstract with the content.\n\nThe proofs are long and intricate; I did not independently verify every estimate, but the structure is sound and the cited tools (Bergman kernel asymptotics, two-scale pseudodifferential calculus, penalization for obstacle problems) are appropriate. The paper deserves a serious referee. For a spectral theorist or semiclassical analyst, this is worth reading and citing. My recommendation: send it to peer review, with the abstract corrected.\n\nWho is this for? People working on non-self-adjoint operators, semiclassical analysis, Bergman kernels, or tunneling. If you read it, focus on Sections 3 and 5; Section 4 is heavier but self-contained.","headline":"Solid semiclassical Weyl law for exponentially small singular values of the ∂-bar operator on the torus, with an abstract that overclaims to compact Riemann surfaces.","tokens_in":61304,"tokens_out":1727,"would_cite":true,"duration_ms":18899,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35P20","47B06","35R35","58J50"],"pacs":[],"model":"deepseek-v4-flash","headline":"The number of exponentially small singular values of the semiclassical $\\bar{\\partial}$ operator is governed by a Weyl law whose constant is the integral of $\\Delta\\varphi$ over the optimal weight's contact set.","keywords":["singular values","Weyl law","semiclassical analysis","bar-partial operator","exponential weights","double obstacle problem","free boundary","quantum tunneling"],"falsifier":"Take a smooth weight on the torus with $\\Delta\\varphi(y)=y^2$ near $y=0$ so that the non-degeneracy condition fails, compute $N([0,e^{-\\tau/h}])$ numerically for small $h$, and compare with $\\frac{1}{2\\pi h}\\int_{M_+(\\psi)}\\Delta\\varphi$; a persistent discrepancy of order $1/h$ would show the condition is load-bearing, while agreement would show it is unnecessary.","tokens_in":60174,"feed_emoji":"","tokens_out":13698,"duration_ms":126202,"temperature":0.7,"pith_summary":"This paper establishes an asymptotic Weyl law for the exponentially small singular values of the semiclassical $\\bar{\\partial}$ operator conjugated by an exponential weight on a two-dimensional torus. Under the condition that the Laplacian $\\Delta\\varphi$ changes sign transversally across its zero set, the number of singular values in $[0,e^{-\\tau/h}]$ equals $\\frac{1}{2\\pi h}\\int_{M_+(\\psi)}\\Delta\\varphi\\,L(dz)+o(1/h)$, where $\\psi$ is the unique solution of a double obstacle problem that squeezes an optimal weight between $\\varphi$ and $\\varphi-\\tau$. The leading constant is the symplectic area of the region where the optimal weight touches $\\varphi$, which is also the region where the singular states concentrate. For small $\\tau$, the paper refines the constant further, writing the correction as $\\tau^{2/3}$ times an integral along the curve where $\\Delta\\varphi$ vanishes. The result turns a fragile non-self-adjoint spectral count into a deterministic free-boundary computation.","feed_headline":"Weyl law counts exponentially small singular values","feed_subtitle":"On a torus, the count is set by the Laplacian of the weight over the optimal contact region.","key_machinery":"The load-bearing device is the optimal weight $\\psi$, obtained as the unique solution of the double obstacle problem (1.11): $\\varphi-\\tau\\le\\psi\\le\\varphi$ almost everywhere, with $\\Delta\\psi\\ge 0$ above the lower obstacle and $\\Delta\\psi\\le 0$ below the upper obstacle. Because $\\psi$ is simultaneously an upper and a lower bound weight, the singular-value count is squeezed between two trace-class computations: one uses weighted $L^2$ estimates to show singular states decay away from $M_+(\\psi)$, and the other uses coherent states localized in the interior of $M_+(\\psi)$ whose spectral projection reproduces them. The trace of the resulting asymptotic projection onto holomorphic sections gives the factor $\\frac{1}{2\\pi h}\\int\\Delta\\varphi$. In the thin-band regime the band where $\\psi$ is harmonic is analyzed in coordinates around the curve $\\gamma$, using an $\\varepsilon$-pseudodifferential calculus with operator-valued symbols in the normal variable, which constructs the boundary functions $f_\\pm$ and produces the $\\tau^{2/3}$ correction. The free-boundary part, proved on general compact manifolds, establishes existence, $C^{1,1}$ regularity, positive measure of both contact sets, and zero volume of the free boundaries, the last fact being what lets the upper and lower bounds meet.","core_discovery":"The central claim is Theorem 4: for a nonconstant smooth weight $\\varphi$ on the torus with $0<\\tau<\\max\\varphi-\\min\\varphi$ and $d\\Delta\\varphi\\neq 0$ on $(\\Delta\\varphi)^{-1}(0)$, the counting function satisfies $$N([0,$e^{{-\\tau/h}}$])=\\frac{1}{2\\pi h}\\int_{M_+(\\psi)}\\$\\Delta$\\varphi(z)\\,L(dz)+o(1/h),\\quad h\\to 0^+,$$ where $\\psi$ is the unique $C^{1,1}$ solution of the double obstacle problem (1.11) and $M_+(\\psi)=\\{\\psi=\\varphi\\}$ is its upper contact set. The same $\\psi$ is both an upper and a lower bound weight, so the count is trapped between two estimates with the same constant. The paper also proves Theorem 5, which for $\\tau=\\varepsilon^3\\hat\\tau$ and connected zero set $\\gamma$ gives the sharper expansion $\\int_{M_+(\\psi)}\\Delta\\varphi=\\int_{M_+}\\Delta\\varphi-\\tau^{2/3}\\frac{1}{2}\\left(\\frac{3}{2}\\right)^{2/3}\\int_\\gamma(\\partial_n\\Delta\\varphi)^{1/3}dx+O(\\varepsilon^3)$, making explicit how the sign-change curve controls the subleading term.","pith_inferences":["The torus is used for convenience rather than necessity; a natural testable extension is to a general compact Riemann surface, where the same double obstacle problem should yield the same Weyl constant once a suitable strictly subharmonic exhaust function is available.","The $\\tau^{2/3}$ exponent suggests a broader tunneling rule: when the Laplacian of a weight changes sign transversally, the subleading tunneling count is organized by $\\int_\\gamma |\\nabla\\Delta\\varphi|^{1/3}$, so one could look for the same exponent in magnetic Pauli or Schrödinger tunneling problems.","The theorem suggests a numerical pipeline: solve the double obstacle problem once, evaluate the contact integral, and predict the exponentially small spectrum; deviations at finite $h$ would directly measure the $o(1/h)$ remainder."],"forward_implications":["The count of exponentially small singular values is of order $1/h$, and its leading constant depends only on the Laplacian of the weight on the contact set, not on the full symbol of the operator.","Because the same optimal weight serves as both upper and lower bound weight, the asymptotic is two-sided: the true count is trapped between two trace computations with the same leading constant.","For small exponential decay rate $\\tau$, the subleading correction to the Weyl constant scales like $\\tau^{2/3}$ and is determined by the integral of $(\\partial_n\\Delta\\varphi)^{1/3}$ along the curve where $\\Delta\\varphi$ changes sign.","For the model weight $\\varphi(y)=\\sin y$, the general theorem recovers the separation-of-variables result, confirming that the free-boundary mechanism is consistent with the known case."],"supporting_citations":[{"why":"supplies the weighted $L^2$ estimates for the $\\bar{\\partial}$ equation used to localize singular states near the contact set in the upper-bound argument.","marker":"[Hö90]"},{"why":"provides the $L^2$ solvability estimates for $\\bar{\\partial}$ with a subharmonic weight used to construct quasimodes in the lower-bound argument.","marker":"[Hö94]"},{"why":"gives the asymptotic Bergman projection and its leading trace, the tool that converts the spectral count into an integral of $\\Delta\\varphi$.","marker":"[BeBrSj08]"},{"why":"supplies the direct semiclassical description of Bergman projections with smooth weights needed for the trace-class estimate.","marker":"[HiSt22]"},{"why":"provides the analytic semiclassical Bergman operator asymptotics behind the trace computations in the upper and lower bounds.","marker":"[RoSjVu20]"},{"why":"introduces the two-scale pseudodifferential calculus with operator-valued symbols used in the thin-band analysis around $\\gamma$.","marker":"[GeMaSj91]"},{"why":"supplies the Weyl law and trace-class background, as well as the operator-valued symbol framework used in the thin-band proof.","marker":"[DiSj99]"},{"why":"supplies the penalization term adapted to construct solutions of the double obstacle problem on a compact manifold.","marker":"[LePa23]"},{"why":"gives the porosity strategy that proves the free boundaries have zero volume, a step needed to apply the upper and lower bound theorems.","marker":"[PeShUr12]"},{"why":"establishes the model case $\\varphi=\\sin y$ by separation of variables, the baseline that the general theorem extends.","marker":"[SjVo24]"}],"fun_headline_variants":["Free boundary problem yields Weyl law for tiny singular values","Optimal weights count exponentially small dbar singular values","Torus weight curvature sets singular value counting asymptotics","Sharp bounds on exponentially small singular values via obstacle problem"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything depends on the Laplacian of the weight crossing zero cleanly wherever it is zero: only then do the free boundaries of the optimal weight have zero volume and the small-$\\tau$ analysis reduces to a single curve.","fun_headline_variants_meta":{"raw":{"variants":["Free boundary problem yields Weyl law for tiny singular values","Optimal weights count exponentially small dbar singular values","Torus weight curvature sets singular value counting asymptotics","Sharp bounds on exponentially small singular values via obstacle problem"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000588,"raw_usage":{"total_tokens":2771,"prompt_tokens":964,"completion_tokens":1807,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":580,"completion_tokens_details":{"reasoning_tokens":1743}},"tokens_in":580,"tokens_out":1807,"duration_ms":15510,"temperature":1.0,"reasoning_tokens":1743,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T22:19:54.373361+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a smooth weight on the torus with $\\Delta\\varphi(y)=y^2$ near $y=0$ so that the non-degeneracy condition fails, compute $N([0,e^{-\\tau/h}])$ numerically for small $h$, and compare with $\\frac{1}{2\\pi h}\\int_{M_+(\\psi)}\\Delta\\varphi$; a persistent discrepancy of order $1/h$ would show the condition is load-bearing, while agreement would show it is unnecessary.","supporting_citations":[],"review_version":1}