{"id":"c5d54bcd-e61b-40ef-a07e-f685d38b65f3","arxiv_id":"2506.02901","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In weighted Bergman spaces, equally spaced poles minimize the norm of high-degree simple partial fractions exactly when the first N^2-2 moments vanish, and this fails in general.","lead":"This paper studies sums of fractions 1/(z-a)^N with poles on the unit circle, measured in weighted Bergman spaces of the disk. It shows that evenly spaced poles are optimal only under extra moment conditions, and that the unconstrained optimum can be asymmetric.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.4's proof hinges on an unverified finite modification of the coefficient sequence a_m; the strict convexity of the auxiliary function, and hence the equidistribution uniqueness, is not established, and the theorem is also misstated for small n.","rationale":"A good-faith reading of the paper shows that the asymptotic and density results, Theorems 1.5 through 1.7, are supported by explicit computations and appear sound, and the numerical counterexample to equidistribution is concrete and reproducible. However, the central claim is Theorem 1.4, and its proof is not self-contained at the critical convexity step. The reader identified exactly this as the weakest assumption: the phrase 'it can be easily verified' in Section 6 covers a nontrivial sequence modification that is necessary for applying Bari's theorem and Lemma 13 of [1]. This concern is real and load-bearing because the strict convexity of the auxiliary function is what forces the equidistributed configuration to be the unique minimizer. The small-n issue is also genuine but less central; it can be fixed by restricting n > N^2 - 2, and it does not undermine the large-n content of the theorem. These are fixable gaps rather than evidence that the main asymptotic results are wrong, so the appropriate verdict remains CONDITIONAL.","tokens_in":18282,"tokens_out":27517,"duration_ms":283844,"concrete_test":"For N = 2 and N = 3, compute a_m = 1 for m < N and a_m = 1 - (m!)^2/((m-N)!(m+N)!) for m >= N. Set N0 = N^2 - 1 and run the recursive construction of Proposition 4.3 to obtain tilde a_1, ..., tilde a_{N0-1}. Then evaluate psi(theta) = tilde a_0/2 + sum_{m>=1} tilde a_m cos(m theta) on a fine grid over (0,2pi). If min psi < 0, or if the recursion fails to preserve positivity, strict decrease, and convexity, the proof of Theorem 1.4 collapses. Separately, for N = 2, n = 2, solve the two moment equations to confirm that W_2 is empty.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Section 6, after defining a_m, the proof asserts without verification that the sequence satisfies the hypotheses of Proposition 4.3 and can be finitely modified so that the resulting series defines a strictly convex auxiliary function on (0,2pi). This is the essential step that allows the author to invoke Lemma 13 of [1] and conclude both minimality and uniqueness of the equidistributed configuration. Proposition 4.3 requires checking that a_n + n Delta a_n is eventually decreasing and that the recursive construction with N0 = N^2 - 1 yields a positive, strictly decreasing, convex sequence with tilde a_m = a_m for m >= N0. The text only says these conditions 'can be easily verified' and that one 'may decrease coefficients a_1, ..., a_{N^2-2}'; no verification or explicit modified sequence is given. If the modified sequence is not convex, Bari's theorem cannot be applied, so the auxiliary series may take negative values and the decomposition (14) together with strict convexity of tilde_phi is unsupported. This is load-bearing because it is the only route to the equality and uniqueness statement of Theorem 1.4. A separate, more elementary issue is that the theorem is already false or undefined for n <= N^2 - 2: for example, when N = 2 and n = 2, W_2 is empty since a + b = 0 and a^2 + b^2 = 0 force a = 0, while the right-hand side is finite. The theorem needs an explicit hypothesis such as n > N^2 - 2.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies simple partial fractions of degree N with poles on the unit circle in the standard weighted Bergman spaces A^2_alpha. The main results are: an extension of Korevaar's theorem to degree-N fractions (Theorem 1.2) and a corresponding density statement (Corollary 1.3); a constrained minimization theorem (Theorem 1.4) asserting that, under vanishing of the first N^2-2 moments, the minimal Bergman norm is attained up to rotation by the equidistributed configuration; sharp asymptotics for the equidistributed norms (Theorem 1.5); a matching order-of-magnitude lower bound for arbitrary configurations (Theorem 1.6); and a density/nowhere-density dichotomy for SF^N(T) in A^2_alpha (Theorem 1.7). The paper also gives explicit counterexamples showing that without the moment constraints equidistribution is not optimal, including the N=2, alpha=3 two-point case computed in closed form.","tokens_in":18684,"tokens_out":18675,"duration_ms":169999,"significance":"If fully established, the results extend the Chui-type norm-minimization theory from first-order to higher-order simple partial fractions, identify a genuinely new non-equidistribution phenomenon, and provide a sharp asymptotic constant. Theorems 1.5 and 1.6 are carefully derived and appear sound; Theorem 1.5 has a detailed elementary proof, and Theorem 1.6 gives an explicit constant. The counterexamples in Section 5 are computed exactly and corroborated numerically. No free parameters are fitted: all constants come from the Bergman norm itself. The paper's main bottleneck is the proof of Theorem 1.4, which contains an unverified convexification step and a small-n statement problem. The density results rely only on Theorems 1.5 and 1.6 and on the cited work of Borodin, so they are not affected by the gap in Theorem 1.4 once Propositions 8.1 and 8.3 are checked; I found those arguments sound.","major_comments":[{"comment":"As stated, the theorem is not well-posed for n <= N^2 - 2. For instance, when N = 2 and n = 2, the constraints a0 + a1 = 0 and a0^2 + a1^2 = 0 force a0 = a1 = 0, so no distinct points on the unit circle satisfy them; W2 is empty, the minimum over W2 is undefined (or +infinity), while the right-hand side ||Psi_2^N||_{alpha*} is finite. The statement should be restricted to n > N^2 - 2 or otherwise made conditional on Wn being nonempty.","section":"Theorem 1.4 (statement)"},{"comment":"The proof hinges on the unverified assertion, after the definition of a_m, that the sequence satisfies the hypotheses of Proposition 4.3 and that the inequality (m+1)a_m - m a_{m+1} <= 1 holds for m >= N^2-1. These facts are load-bearing: they justify the finite modification of a_1,...,a_{N^2-2} and the invocation of Bari's theorem. The text says only that these conditions 'can be easily verified' and that one 'may decrease' the early coefficients; no verification and no explicit modified sequence is supplied. Without a proof that the modified sequence is positive, strictly decreasing, and convex, the strict convexity of tilde_phi, and with it the equality case in Lemma 13 of [1], is not established.","section":"Section 6, proof of Theorem 1.4"},{"comment":"Even if the modified coefficients are convex, the paper needs an explicit argument that tilde_phi is strictly convex on (0,2pi) rather than merely convex. Remark 4.2 states that a merely convex sequence can produce zeros of the cosine series, so the positivity of psi alone does not immediately give strict convexity of its second antiderivative; the fact that the zeros are countable and hence no interval of constancy occurs should be stated and proved. This point is necessary because Lemma 13 of [1] uses strict convexity for the uniqueness conclusion in Theorem 1.4.","section":"Equation (14) and Remark 4.2"}],"minor_comments":[{"comment":"The phrase 'minimize the quantity in (5)' should refer to the energy in (13) or the equivalent expression in Section 3.2; equation (5) in the text is the series expansion of 1/(z-e^{it})^N.","section":"Section 5.1"},{"comment":"The cross-reference 'Remark 2.2' should be 'Remark 4.2'.","section":"Section 6"},{"comment":"In the displayed lower-bound statement the indexing '0 <= k <= N-1' should be '0 <= k < n'; n is the number of poles.","section":"Section 7, proof of Theorem 1.6"},{"comment":"Reference [11] is listed but not cited in the text, and the spelling 'Bari' used in the text should be harmonized with 'Bary' in the bibliography.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal. My main uncertainty is Theorem 1.4: the statement needs a small-n correction, and the proof requires a detailed verification of the convexification step that is currently only asserted. The asymptotic and density results appear solid. If the author supplies the missing verification, I would be inclined to accept."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper extends Abakumov-Borichev-Fedorovskiy's Chui-conjecture work from degree 1 to degree N in weighted Bergman spaces. What is actually new: the density dichotomy (Theorem 1.7, now with the critical exponent), the sharp norm asymptotics (Theorem 1.5), the two-sided comparison (Theorem 1.6), and the genuinely surprising counterexamples in Section 5 showing that for N=2, alpha=3, two poles minimize the norm at angle ~0.919 rad rather than opposite poles, and three points are not equispaced. Those counterexamples are exact integral evaluations plus numerical minimization; they are convincing and are the most valuable part. The Korevaar extension (Theorem 1.2) is routine differentiation, as the reader noted.\n\nThe main asymptotic and density arguments appear sound. I checked the series manipulations in Theorem 1.5 and they are careful; the use of Borodin's theorem in Proposition 8.3 is appropriate. The paper is honestly written and gives credit where due.\n\nSoft spots. Theorem 1.4 as stated is false for n <= N^2-2: W_n is empty because the first moment conditions force all points to coincide, while the right-hand side is finite. The theorem needs a hypothesis n > N^2-2 (or n large enough in terms of N), and this is probably harmless because the intended statement is for large n.\n\nMore important is the proof gap in Section 6. After defining the coefficient sequence a_m, the proof asserts without verification that the sequence satisfies Proposition 4.3's hypotheses and that a finite modification of a_1,...,a_{N^2-2} gives a strictly convex auxiliary function. Proposition 4.3 requires checking that a_n + n Delta a_n is eventually decreasing and that the recursive construction yields a positive strictly decreasing convex sequence with tilde_a_m = a_m for m >= N0. The text simply says these \"can be easily verified\" and that one \"may decrease\" the early coefficients. No explicit modified sequence is given. This is load-bearing because strict convexity of tilde_phi is what lets the author apply Lemma 13 of [1] and get uniqueness of the equidistributed minimizer. If the modification cannot be made with the required properties, the equality and uniqueness statement of Theorem 1.4 is unsupported. The small-n false statement is easy to fix; the convex-sequence step is a genuine gap that needs a real proof, not a remark.\n\nAlso minor: the text has a few typos, and the numerical section has an incomplete sentence; nothing that affects the mathematics.\n\nWho this is for: anyone working on Chui-type problems, simple partial fractions, or Bergman space approximation. The counterexamples and asymptotics are worth having even if Theorem 1.4 needs revision.\n\nMy recommendation: send it to peer review. The main results are significant enough, and the gap is identifiable and likely fixable. A serious referee will want the Section 6 gap closed and the small-n hypothesis added before publication.","headline":"Solid new results on degree-N simple partial fractions, but Theorem 1.4 has a real proof gap and a small-n false statement; fixable, worth refereeing.","tokens_in":19160,"tokens_out":2188,"would_cite":true,"duration_ms":20171,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["30H20","30E10","41A20"],"pacs":[],"model":"deepseek-v4-flash","headline":"For degree-N simple partial fractions, equally spaced poles are the unique Bergman minimizer exactly under vanishing moments.","keywords":["simple partial fractions","weighted Bergman spaces","equidistribution","moment conditions","interaction function","convex sequences","sharp asymptotics","rational approximation"],"falsifier":"Search numerically for $n$ distinct points on the unit circle satisfying $\\sum_k a_k^m=0$ for $1\\le m\\le N^2-2$ whose norm at $\\alpha^*=2N-1$ is strictly smaller than $\\|\\Psi_n^N\\|_{\\alpha^*}$; Theorem 1.4 predicts none exist, so a single such configuration would refute it. Alternatively, compute the second-difference sequence of the coefficients in Section 6 and check directly whether finitely many modifications can make it strictly convex while preserving the required endpoint conditions; failure of that step would break the uniqueness conclusion.","tokens_in":18097,"feed_emoji":"📐","tokens_out":10854,"duration_ms":97644,"temperature":0.7,"pith_summary":"This paper asks where points $a_0,\\dots,a_{n-1}$ on the unit circle should be placed so that the high-degree simple partial fraction $f(z)=\\sum_{k=0}^{n-1}(z-a_k)^{-N}$ has the smallest possible norm in the weighted Bergman space $A^2_{\\alpha}(\\mathbb{D})$. The main result is that at the critical weight $\\alpha^*=2N-1$, the evenly spaced configuration is the unique minimizer up to rotation, provided the first $N^2-2$ moments $\\sum_k a_k^m$ vanish. Without that moment condition, equidistribution is not optimal: for $N=2$ and $\\alpha=3$, two poles separated by about $0.919$ radians beat opposite poles, so the symmetric configuration is generically wrong. The paper also establishes sharp asymptotics for the equidistributed norms, proves the unconstrained minimum is always comparable to that value, and locates precisely the exponents at which $SF^N(\\mathbb{T})$ stops being dense. A sympathetic reader should take the core insight to be that symmetry in this extremal problem is restored exactly when low-order moments are forced to zero.","feed_headline":"At one critical weight, evenly spaced poles win only if moments vanish","feed_subtitle":"For higher-order pole sums, symmetry is optimal only under vanishing moments, with sharp asymptotics and a density threshold.","key_machinery":"The load-bearing object is the interaction function $\\phi_{\\alpha,N}(\\vartheta)$, defined by $\\phi_{\\alpha,N}(\\vartheta)=\\frac{k_g}{((N-1)!)^2}\\sum_{m=N}^{\\infty} c_m(\\alpha,N)\\cos(m\\vartheta)$, which represents the real part of the Bergman inner product between two poles separated by angle $\\vartheta$. The minimization problem becomes the purely angular problem of minimizing $\\sum_{j\\ne k}\\phi_{\\alpha,N}(\\vartheta_j-\\vartheta_k)$. The proof at $\\alpha^*=2N-1$ rewrites the second derivative of this function as a cosine series whose coefficients are asymptotically convex, applies Bari's theorem on positive cosine series, and uses Proposition 4.3 to modify finitely many coefficients so that the resulting function $\\tilde{\\varphi}$ is strictly convex on $(0,2\\pi)$. The strict convexity then triggers Lemma 13 of $[1]$, which says that a strictly convex pairwise interaction is minimized uniquely by equidistributed angles, proving Theorem 1.4.","core_discovery":"The central claim is Theorem 1.4: for $N>1$ and $\\alpha^*=2N-1$, among all configurations of $n$ distinct poles on the unit circle whose first $N^2-2$ moments vanish, the minimal Bergman norm of $\\sum_{k=0}^{n-1}(z-a_k)^{-N}$ is attained exactly, up to rotation, by the equally spaced points $e^{2\\pi i k/n}$. The paper shows this is a genuine phenomenon rather than a triviality, because without the moment restrictions the minimizer is different; the explicit two-pole example with $N=2$, $\\alpha=3$ has optimal separation $\\vartheta_{\\min}\\approx 0.919$ radians, not $\\pi$. Around this, the paper proves sharp asymptotic formulas, shows that the unconstrained minimum is comparable to the equidistributed norm, derives the density dichotomy for $SF^N(\\mathbb{T})$ in $A^2_{\\alpha}$, and extends Korevaar's approximation theorem to degree-$N$ simple fractions.","pith_inferences":["One extension the paper leaves implicit is that the index $N^2-2$ is not arbitrary: it is the first point where the coefficient inequality $(m+1)a_m-ma_{m+1}\\le a_0$ holds, suggesting the number of moment constraints needed to restore symmetry exactly matches the number of non-convex early coefficients.","A testable extension would be to determine the limiting shape of the optimal unconstrained configuration for $N>1$; Theorem 1.6 guarantees it stays within a constant of equidistribution, but the exact angular distribution as $n\\to\\infty$ is left open and could be found numerically.","The phenomenon may connect to Riesz-energy and logarithmic-potential problems on the circle: the interaction function is a Coulomb-type kernel, and the failure of equidistribution when the kernel is not convex resembles known behavior in energy minimization, so methods for convex kernels could be adapted to classify when symmetry holds.","A straightforward check for $N=3$, $\\alpha=5$ would test whether the two-pole optimal angle again deviates from $\\pi$ and whether the deviation follows a pattern in $N$."],"forward_implications":["At $\\alpha^*=2N-1$, the moment constraints single out equidistribution: any minimizer in $W_n$ is, after a rotation, the regular $n$-gon, so the family $W_n$ is the right constrained setting in which the symmetric conjecture survives.","For $N>1$, the unconstrained problem behaves differently: the two-pole $N=2$, $\\alpha=3$ minimizer is separated by about $0.919$ radians, so balanced placement is not optimal in general.","The unconstrained minimum remains comparable to the equidistributed norm, so asymmetric optima can improve the energy only by a constant factor as $n\\to\\infty$.","The sharp asymptotic $n^{\\alpha+1-2N}\\|\\Psi^N_n\\|_\\alpha^2 \\to \\frac{\\Gamma(\\alpha+2)\\zeta(\\alpha+1-2(N-1))}{((N-1)!)^2}$ follows from the power-series representation of $\\Psi^N_n$ as a logarithmic derivative of $z^n-1$.","Density of $SF^N(\\mathbb{T})$ in $A^2_\\alpha$ is controlled by sharp exponents: nowhere dense for $2(N-1)<\\alpha<\\alpha^*$, not dense at $\\alpha^*$, and dense for $\\alpha>\\alpha^*+1$."],"supporting_citations":[{"why":"Supplies Lemma 13, the strict-convexity criterion that forces equidistribution as the unique minimizer of the pairwise interaction sum, and the overall method of proof for Theorem 1.4.","marker":"[1]"},{"why":"Provides Bari's theorem and the Fejer-kernel series representation used to turn convexity of the modified cosine coefficients into positivity of the cosine series.","marker":"[2]"},{"why":"Supplies Theorem 5 on density of semigroups in Banach spaces, used in Proposition 8.3 to show density of $SF^N(\\mathbb{T})$ for $\\alpha>\\alpha^*+1$.","marker":"[4]"},{"why":"Establishes the equivalence between polynomial approximation sets and asymptotically neutral families, which the paper extends to degree-$N$ simple fractions in Theorem 1.2.","marker":"[12]"},{"why":"Poses the original lower-bound conjecture for degree-one simple fractions that motivates the higher-degree minimization problem addressed here.","marker":"[6]"}],"fun_headline_variants":["Evenly spaced poles optimal only under vanishing moments","Moment condition decides when equidistribution is optimal","Critical weight and the equidistribution threshold","Pole symmetry wins if and only if moments vanish"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of the main uniqueness theorem rests on an assertion, labeled 'it can be easily verified', that a certain infinite sequence of cosine coefficients can be changed in finitely many places to become strictly convex, so that a classical positivity theorem applies; if that assertion fails, the uniqueness conclusion is not justified.","fun_headline_variants_meta":{"raw":{"variants":["Evenly spaced poles optimal only under vanishing moments","Moment condition decides when equidistribution is optimal","Critical weight and the equidistribution threshold","Pole symmetry wins if and only if moments vanish"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000493,"raw_usage":{"total_tokens":2387,"prompt_tokens":873,"completion_tokens":1514,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":489,"completion_tokens_details":{"reasoning_tokens":1456}},"tokens_in":489,"tokens_out":1514,"duration_ms":11041,"temperature":1.0,"reasoning_tokens":1456,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:15:02.023420+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search numerically for $n$ distinct points on the unit circle satisfying $\\sum_k a_k^m=0$ for $1\\le m\\le N^2-2$ whose norm at $\\alpha^*=2N-1$ is strictly smaller than $\\|\\Psi_n^N\\|_{\\alpha^*}$; Theorem 1.4 predicts none exist, so a single such configuration would refute it. Alternatively, compute the second-difference sequence of the coefficients in Section 6 and check directly whether finitely many modifications can make it strictly convex while preserving the required endpoint conditions; failure of that step would break the uniqueness conclusion.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides Bari's theorem and the Fejer-kernel series representation used to turn convexity of the modified cosine coefficients into positivity of the cosine series."},{"cited_title":"Density of a semigroup in a Banach space","cited_arxiv_id":null,"evidence_quote":"Supplies Theorem 5 on density of semigroups in Banach spaces, used in Proposition 8.3 to show density of $SF^N(\\mathbb{T})$ for $\\alpha>\\alpha^*+1$."},{"cited_title":"Asymptotically Neutral Distributions of Electrons and Polynomial Approxima- tion","cited_arxiv_id":null,"evidence_quote":"Establishes the equivalence between polynomial approximation sets and asymptotically neutral families, which the paper extends to degree-$N$ simple fractions in Theorem 1.2."}],"review_version":1}