{"id":"8c40a7e9-aed4-4f2e-8081-dcddba7b945f","arxiv_id":"2411.12378","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Using Grunsky coefficients, the authors prove new upper bounds |H2(2)| <= 1.3614... and |H3(1)| <= 1.6787... for the class S of univalent functions.","lead":"This paper improves the known upper bounds for two Hankel determinants of univalent functions, using Grunsky coefficients. The new bounds, 1.3614 and 1.6787, are much closer to the conjectured sharp values.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The proof of Theorem 2(ii) depends on an unverified numerical assertion that the partial-derivative system for F2 has only two real solutions; a missed interior critical point with a higher value would invalidate the claimed |H3(1)| bound.","rationale":"The reader's verdict identifies the same load-bearing assumption: the numerical global-maximization step for F2 is asserted without rigorous verification. My review independently confirms that the entire upper-bound proof for |H3(1)| reduces to this optimization, and that the derivation of F2 itself is valid (the earlier coefficients in (7), (10), (13), and (14) check out; the slightly mixed treatment of |ω15| in bounding (15) is conservative and does not break the argument). The paper also lacks a certificate for the boundary derivative claim ∂F2/∂x(x,0)<0, but that is a minor omission compared with the critical-point count. I agree with the condition: the numerical step should be made rigorous or reproducible. No other flaw of comparable weight emerged; the formulas appear internally consistent and the reported extrema are plausible. Thus the reader's CONDITIONAL verdict should stand, with the requested check being a certified global maximization of F2 over D1.","tokens_in":5846,"tokens_out":30255,"duration_ms":250757,"concrete_test":"Use a certified global optimization method (e.g., Mathematica's interval-arithmetic NMaximize or a rigorous branch-and-bound solver) to maximize F2 over D1. Specifically, eliminate R from ∂F2/∂x=0 and ∂F2/∂y=0 to obtain a polynomial system, then apply cylindrical algebraic decomposition to count all real solutions inside D1 and verify that the only stationary values are 1.6787... and 1.5559..., with 1.6787... the global maximum. If the certified maximum equals 1.6787..., the concern is resolved; if any point exceeds it, Theorem 2(ii) is false.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central proof step is the global maximization of F2 on D1 (Section 2, after (17)). The authors state 'Numerical calculation give that the system ... has only two real solutions' and then report maxima 1.6787... and 1.5559... at those points, but give no code, no interval-arithmetic certificate, and no algebraic elimination reducing the system to a provably finite polynomial system whose real-root count can be certified. Without this, the inequality |H3(1)| ≤ 1.6787 is not established: an additional interior critical point with F2 > 1.6787 would invalidate the bound. Spot evaluations (e.g., F2(0.58,0.21) ≈ 1.6789) are consistent with the reported maximum, so the numerical claim may be correct, but the proof as written is conditional on an unverified computation. In part (i) the analogous uniqueness claim can be verified analytically (the critical equations reduce to x=√(11/30), y=√(281/1800)), which highlights the absence of a comparable derivation in part (ii). The boundary assertions, including ∂F2/∂x(x,0)<0 for 0<x<1, are also stated without proof but are far easier to check; the main gap is the interior critical-point count.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the Hankel determinants H2(2) and H3(1) for the class S of normalized univalent functions. Using the Grunsky coefficient representation (7) and the Grunsky inequality (3), the authors express H2(2) and H3(1) in terms of four Grunsky coefficients, apply the modulus bounds (9) and the estimate |a3-a2^2|<=1, and reduce the problem to maximizing the explicit functions F1 and F2 on the domain D1. They report that the interior critical systems have one (respectively two) real solutions and find the boundary maxima, leading to Theorem 2: |H2(2)| <= 1.3614... and |H3(1)| <= 1.6787..., improving the earlier bounds 11/3 and 3.2588... from [7].","tokens_in":6147,"tokens_out":10899,"duration_ms":93451,"significance":"Assuming the numerical maximization claims can be made rigorous, the result is a genuine and welcome improvement for the class S, where sharp estimates are not known. The algebraic reduction from Grunsky relations to F1 and F2 is mostly correct, and the use of (9) and |a3-a2^2|<=1 is standard. The paper is not circular: the bounds are consequences of the Grunsky inequalities, not fitted to the target constants. However, the proof as written is conditional on unverified numerical statements about global maxima; without a rigorous certification the theorem is not established. The paper also contains no code or machine-checkable artifacts, so the numerical claims cannot be independently reproduced from the text.","major_comments":[{"comment":"The assertion that the system dF2/dx=0, dF2/dy=0 has only two real solutions in the interior of D1 is the load-bearing step for the bound |H3(1)| <= 1.6787. The text supports this only by 'Numerical calculation give...' and reports the values at the two points. A missed interior critical point with F2 larger than the reported 1.6787... would invalidate Theorem 2(ii). Please replace this with a rigorous argument: for example, reduce the two equations to a univariate polynomial and use Sturm sequences or resultant computations, or provide an interval-arithmetic certified computation with the code or scripts used, or give an analytic proof. The boundary analysis alone does not remove this gap.","section":"§2, after (17)"},{"comment":"The analogous statement in part (i), that dF1/dx=0, dF1/dy=0 has only one real solution in the interior of D1, is also made on the basis of numerical verification. This case is much easier: the equations reduce analytically to x^2=11/30 and y^2=281/1800, so the uniqueness and the value F1=1.19889... can be proven in a few lines. Since this point is part of the justification for the H2 bound, the derivation should be included rather than left to numerical assertion.","section":"§2, after (11)"},{"comment":"The boundary-maximum arguments for F2 also contain unproved monotonicity claims, in particular dF2/dx(x,0)<0 for 0<x<1 and dF2/dy(0,y)<=0 for 0<=y<=1/sqrt(3). These are elementary but load-bearing: they are how the text concludes F2(x,0)<=F2(0,0) and F2(0,y)<=F2(0,0). Please supply the derivative calculations or a certified interval check. This is a smaller gap than the interior critical-point count, but it should be closed as well.","section":"§2, boundary of D1 in part (ii)"}],"minor_comments":[{"comment":"In the derivation of (8)-(9), the inequality '5|omega15|^2+7|omega15|^2 <= 1' should read '5|omega15|^2+7|omega17|^2 <= 1'.","section":"§1, equations (8)-(9)"},{"comment":"Theorem 2 labels both parts as '(i)'; the second should be '(ii)'.","section":"Theorem 2"},{"comment":"There are several typos: 'NExt' before (10), 'satrt' after (15), 'seams' in Section 1, and 'univalent function' in reference [1]. These should be corrected.","section":"Throughout"},{"comment":"In the reduction to F2, the terms (4xy+2x^3)*sqrt(...) are obtained by bounding |omega15| by sqrt(1-x^2-3y^2), which is weaker than the estimate in (9) and drops the factor 1/sqrt(5). The resulting F2 is a valid upper bound, but the weaker estimate should be stated explicitly to avoid appearing as a factor error.","section":"§2, reduction to F2 after (16)"},{"comment":"The maximum of 2/(3*sqrt(3))*(1-x^2)^{3/2}+x^2(1-x^2) on [0,1] is quoted as 7/16 at x=1/2 without derivation; a one-line derivative check would help the reader.","section":"§2, boundary analysis, part (ii), item 3"}],"recommendation":"major_revision","confidential_remarks":"To the editor: The main mathematical idea is sound and the numerical claims are likely correct, but the proof as written is not fully rigorous because the global-maximization steps are verified only numerically. A revision that supplies certified computations or analytic root counts would make the paper acceptable. I would not reject on the basis of the missing 1/sqrt(5) factor; it makes the H3 bound coarser but does not break the argument. The journal may also wish to ask for the code or scripts behind the numerical checks."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the good news: the paper delivers what it promises. The bounds for |H2(2)| and |H3(1)| over the full class S are improved substantially (3.666 to 1.3614, and 3.2588 to 1.6787), and the proof is a clean Grunsky-coefficient argument. The reduction to the two-variable functions F1 and F2 is transparent, the use of the Grunsky inequalities is legitimate, and there is no circularity. I checked the algebra in Section 2; apart from a small coefficient slip discussed below, it works.\n\nThe soft spot is the global maximization of F1 and F2. For F1, the interior critical-point claim is actually easy to verify analytically, so that part is fine. For F2, the paper says the system has only two real interior solutions, but gives no code, no interval-arithmetic certificate, and no algebraic elimination. The boundary estimates are asserted without proof, though those are simple to check. The interior count is load-bearing: if there is an unobserved interior critical point with F2 above 1.6787, the claimed |H3(1)| bound does not follow. Spot evaluations near the reported point are consistent with the claimed maximum, so I suspect the result is true, but the proof as written is conditional on an unverified computation.\n\nMinor point: in the reduction to F2, the terms 4xy and 2x^3 in the square-root factor should carry a 1/√5 from the |ω15| estimate. The stated F2 is coarser but still a valid upper bound, so it does not hurt the result; it just wastes a little slack.\n\nBottom line: this is a real improvement by a standard method, with one addressable gap. It deserves a serious referee who can either demand a rigorous certification of the critical-point count or accept a reproducible computational certificate. I would not cite the H3 bound as a proven theorem until that gap is closed, but the paper should definitely go to review.","headline":"A genuine improvement in Hankel bounds via Grunsky coefficients, held back by an unverified numerical maximization step that should be certified before the H3 bound is taken as proven.","tokens_in":6619,"tokens_out":4022,"would_cite":false,"duration_ms":36268,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["30C45","30C50","30C55"],"pacs":[],"model":"deepseek-v4-flash","headline":"For univalent functions, this paper establishes the improved Hankel bounds $|H_2(2)| \\leq 1.3614$ and $|H_3(1)| \\leq 1.6787$.","keywords":["univalent functions","Hankel determinants","second Hankel determinant","third Hankel determinant","Grunsky coefficients","Grunsky inequality","upper bounds","coefficient estimates"],"falsifier":"Take the domain $D_1$ and compute the global maxima of $F_1$ and $F_2$ with rigorous interval arithmetic; if either maximum exceeds $1.3614\\ldots$ or $1.6787\\ldots$, Theorem 2 is false. Equivalently, a concrete univalent function with $|H_2(2)|>1.3614\\ldots$ or $|H_3(1)|>1.6787\\ldots$ would refute the bounds.","tokens_in":5669,"feed_emoji":"📐","tokens_out":9540,"duration_ms":88263,"temperature":0.7,"pith_summary":"This paper establishes improved upper bounds on the magnitudes of the second- and third-order Hankel determinants for the class of univalent (one-to-one) analytic functions on the unit disk: $|H_2(2)| \\leq 1.3614\\ldots$ and $|H_3(1)| \\leq 1.6787\\ldots$. These numbers replace the earlier bounds $3.666\\ldots$ and $3.2588\\ldots$ from [7], so the improvement is substantial. The proof rewrites the Taylor coefficients $a_2,\\ldots,a_5$ using Grunsky coefficients and Grunsky’s inequality, then reduces each determinant to the maximum of an explicit two-variable function over a triangular domain. The final step, finding those maxima, is done by solving the critical-point systems numerically and checking the boundary.","feed_headline":"Univalent function bounds fall to 1.36 and 1.68","feed_subtitle":"Using Grunsky coefficient constraints, the new upper limits nearly halve the old ones.","key_machinery":"The machinery is the Grunsky coefficient system: the logarithmic transform $\\log\\frac{f(t)-f(z)}{t-z}$ for a univalent $f$ expands as $\\sum_{p,q=0}^{\\infty}\\omega_{p,q}t^p z^q$, with symmetry $\\omega_{p,q}=\\omega_{q,p}$ and a constraint called Grunsky’s inequality. Working with the auxiliary function $f_2(z)=\\sqrt{f(z^2)}$ makes only odd-indexed Grunsky coefficients appear, and relations from [3] express $a_2,\\ldots,a_5$ as polynomials in these coefficients. Specializing Grunsky’s inequality gives the estimates (9) on $|\\omega_{11}|,|\\omega_{13}|,|\\omega_{15}|,|\\omega_{17}|$. These estimates convert the Hankel determinants into the explicit two-variable functions $F_1$ and $F_2$, whose maxima over the domain $D_1$ are then evaluated.","core_discovery":"The central claim is Theorem 2: for every normalized univalent function $f(z)=z+a_2z^2+a_3z^3+\\cdots$, the quantities $|a_2 a_4 - a_3^2|$ and $|a_3(a_2 a_4 - a_3^2) - a_4(a_4 - a_2 a_3) + a_5(a_3 - a_2^2)|$ are bounded by $1.3614\\ldots$ and $1.6787\\ldots$, respectively. The proof obtains these bounds by expressing $a_2,\\ldots,a_5$ in terms of the Grunsky coefficients of the odd function $\\sqrt{f(z^2)}$, applying Grunsky’s inequality to control those coefficients, and maximizing the resulting functions $F_1$ and $F_2$ over the domain $D_1$. The reported bounds improve Theorem 1 of [7], which only gave $3.666\\ldots$ and $3.2588\\ldots$.","pith_inferences":["The numerical verification of the two critical-point systems could be upgraded to a rigorous computer-assisted proof using interval arithmetic, removing the one non-analytic step.","The same Grunsky-coefficient reduction may apply to other subclasses of univalent functions or to higher-order Hankel determinants, where the coefficient count grows but the structure persists.","The constants $1.3614$ and $1.6787$ are probably not sharp; the method does not identify extremal functions, and sharper bounds may exist."],"forward_implications":["Every univalent function has $|a_2 a_4 - a_3^2| \\leq 1.3614\\ldots$.","Every univalent function has $|H_3(1)| \\leq 1.6787\\ldots$.","These values replace the previous best bounds $3.666\\ldots$ and $3.2588\\ldots$, more than halving the old estimates.","The reduction to the explicit functions $F_1$ and $F_2$ turns each Hankel estimate into a two-variable maximization problem."],"supporting_citations":[{"why":"Supplies the Grunsky inequality (3) and the bound $|a_3 - a_2^2| \\leq 1$ used in part (ii).","marker":"[1]"},{"why":"Provides the relations (7) expressing the coefficients $a_2,\\ldots,a_5$ through Grunsky coefficients.","marker":"[3]"},{"why":"Gives the previous bounds on $|H_2(2)|$ and $|H_3(1)|$ that Theorem 2 improves.","marker":"[7]"}],"fun_headline_variants":["Hankel bounds for univalent functions drop to 1.36, 1.68","Sharpened Hankel determinant bounds: 1.36 and 1.68","Grunsky coefficients improve univalent Hankel bounds","Univalent functions: new Hankel limits 1.36 and 1.68"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that numerically solving the systems $\\partial F_1/\\partial x=\\partial F_1/\\partial y=0$ and $\\partial F_2/\\partial x=\\partial F_2/\\partial y=0$ really uncovers all interior maxima of $F_1$ and $F_2$ on $D_1$, and that the boundary checks are complete; the paper states this can be verified numerically but gives no proof or code.","fun_headline_variants_meta":{"raw":{"variants":["Hankel bounds for univalent functions drop to 1.36, 1.68","Sharpened Hankel determinant bounds: 1.36 and 1.68","Grunsky coefficients improve univalent Hankel bounds","Univalent functions: new Hankel limits 1.36 and 1.68"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000187,"raw_usage":{"total_tokens":1251,"prompt_tokens":791,"completion_tokens":460,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":407,"completion_tokens_details":{"reasoning_tokens":375}},"tokens_in":407,"tokens_out":460,"duration_ms":5084,"temperature":1.0,"reasoning_tokens":375,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T17:36:08.519897+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the domain $D_1$ and compute the global maxima of $F_1$ and $F_2$ with rigorous interval arithmetic; if either maximum exceeds $1.3614\\ldots$ or $1.6787\\ldots$, Theorem 2 is false. Equivalently, a concrete univalent function with $|H_2(2)|>1.3614\\ldots$ or $|H_3(1)|>1.6787\\ldots$ would refute the bounds.","supporting_citations":[{"cited_title":"Duren, Univalent function, Springer-Verlag, New Yo rk, 1983","cited_arxiv_id":null,"evidence_quote":"Supplies the Grunsky inequality (3) and the bound $|a_3 - a_2^2| \\leq 1$ used in part (ii)."},{"cited_title":"Lebedev, Area principle in the theory of univalent fu nctions, Published by”Nauka”, Moscow, 1975 (in Russian)","cited_arxiv_id":null,"evidence_quote":"Provides the relations (7) expressing the coefficients $a_2,\\ldots,a_5$ through Grunsky coefficients."},{"cited_title":"Obradovi´ c and N","cited_arxiv_id":null,"evidence_quote":"Gives the previous bounds on $|H_2(2)|$ and $|H_3(1)|$ that Theorem 2 improves."}],"review_version":1}