{"id":"cc0aae0b-516b-4195-90ce-3a65d32815e0","arxiv_id":"2602.14719","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For Hua-Pickrell diffusions, the large-N empirical limits are independent of β, and the frozen β=∞ limits are the zeros of pseudo-Jacobi polynomials.","lead":"The paper studies families of many interacting particles described by Hua-Pickrell diffusions and shows that, in the frozen noise-free limit, these systems are governed by pseudo-Jacobi polynomials. It also proves that certain large-system limits do not depend on the noise strength, linking random matrix theory with orthogonal-polynomial asymptotics.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.1's repair of Graczyk–Malecki is asserted, not proved; the corrected identity (4.6') omits drift terms, so all later results depend on an unverified non-collision argument.","rationale":"The reader's weakest assumption identifies exactly the same load-bearing concern: Theorem 1.1's proof depends on an incomplete repair of the Graczyk–Malecki argument, and Theorem 5.1 inherits this dependence. I see no reason to move the verdict: the paper is plausible and largely computational, but the existence/uniqueness theorem is foundational and its proof is only sketched at the critical point. A line-by-line verification of the corrected (4.6'), including the drift terms, would settle whether the gap is real; if the cancellation holds, the conditional acceptance is justified. I also noticed minor issues such as a likely typo in the Carleman condition exponent and a possible scaling mismatch in Theorem 5.3, but these do not threaten the main theorem and are secondary to the GM repair.","tokens_in":32445,"tokens_out":14866,"duration_ms":147775,"concrete_test":"Take the finite-variation process D_n from Eq. (4.2) of Graczyk–Malecki for the SDE (1.3) with β≥1, a,b∈R, and write out the drift contribution from b_i(λ_i)=2(b−aλ_i). On the event {V_n=e_n(A)=0}, compute ∂_{λ_i} e_n(A) and check whether ∑_i b_i(λ_i) ∂_{λ_i} e_n(A)=0 identically. If it does not vanish, identify the surviving terms and show they combine with (4.6') to keep all terms sign-definite; alternatively, verify the entire proof of Proposition 4.3 of [GM] line-by-line with (4.6') and the drift terms included. This can be done symbolically for small N (e.g., N=2,3) and then for general N.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim Theorem 5.1 presupposes Theorem 1.1, the existence and uniqueness of strong solutions of (1.3). The appendix correctly notes that Graczyk–Malecki's Theorem 2.2 cannot be applied directly because H(x,y)=2(1+xy) fails the non-negativity hypothesis. The proposed fix is a corrected identity (4.3') and a replacement (4.6'), followed by the sentence 'It can be now easily checked that the proof of Proposition 4.3 still works'. This is not a proof. In particular, equation (4.6') asserts that 0=D_n equals a sum of non-negative terms only; however, the process V_n in [GM] has a finite-variation part D_n that also contains drift terms b_i(λ_i) ∂_{λ_i} e_n(A). No argument is given that these drift terms vanish on {V_n=0}; without this, the implication 'D_n=0 ⇒ each e-term vanishes' is unjustified. Since Sections 2–5, including the moment ODEs and the freezing CLT, rely on Theorem 1.1, an incompletely verified repair is a load-bearing gap. The issue is not that the conclusion is false; it is that the central theorem's proof is conditional on an unstated algebraic cancellation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies general β-Hua-Pickrell diffusions of N particles on R, both in the stochastic regime β∈[1,∞) and in the deterministic frozen regime β=∞. The main claims are: (i) Theorem 1.1 on existence, uniqueness, and non-collision of strong solutions of the SDE/ODE system; (ii) a characterization of stationary solutions of the frozen ODEs via zeros of pseudo-Jacobi polynomials (Lemma 2.5); (iii) stationary Hua-Pickrell measures for β<∞ (Proposition 3.2); (iv) a freezing CLT for β→∞ with explicit covariance, related to pseudo-Jacobi zeros (Theorems 4.1–4.2); and (v) Theorem 5.1, the central result, asserting that under suitable initial conditions and scaling assumptions a_N/N→â, b_N/N→b̂, the empirical measures converge weakly a.s. to β-independent limits μ_t whose moments satisfy the closed ODE system (5.3). The paper also gives an explicit free-probability description of the limit for â=-1, b̂=0 (Theorem 5.3) and connects the limit measures to free Hua-Pickrell processes (Proposition 5.6). The proof of Theorem 1.1 in Section 6 attempts to repair the non-negativity hypothesis in Graczyk–Malecki [GM] that fails for H(x,y)=2(1+xy).","tokens_in":32795,"tokens_out":14063,"duration_ms":124588,"significance":"If correct, the paper gives a substantial unification of dynamic random matrix limits and orthogonal-polynomial structures: the moment ODEs are explicit, the freezing CLT is concrete, and the free convolution formulas are new. The paper has real strengths: Lemma 2.5 and the moment ODE derivation in Theorem 5.1 are explicit; the Laplace-method CLT and the comparison with Jacobi ensembles are checkable; and the free Itô construction in Theorem 5.7 is a useful contribution. However, the central existence theorem depends on a repair of [GM] that is asserted rather than proved, and Theorem 5.1 delegates a key analytic passage to a previous paper. These are load-bearing gaps in the current version, even though the overall strategy appears credible and the conclusions may be true.","major_comments":[{"comment":"The proof of Theorem 1.1 is incomplete. The text states that 'It can be now easily checked that the proof of Proposition 4.3 still works with (4.6') instead of (4.6)', but the derivation of (4.6') itself is not fully given. In the notation of [GM], the finite-variation part D_n of V_n also contains drift terms involving b_i(λ_i) ∂_{λ_i} e_n(A). The manuscript does not show that these terms vanish on {V_n=0}; it only asserts that some products (λ_i-λ_j)e(...) and (λ_i-λ_j)(λ_i-λ_k)e(...) are zero. Without a proof that the full D_n identity holds as displayed, the implication D_n=0 ⇒ each non-negative term in (4.6') vanishes is unjustified. Since Theorem 1.1 supplies the unique strong solutions used in Sections 2–5, including Theorem 5.1, this is a load-bearing gap, not a local omission.","section":"Section 6, after Eq. (4.6')"},{"comment":"The passage from moment convergence to the Cauchy-transform PDE (5.4) is delegated to 'the same argument as in the proof of Proposition 2.9 in [VW1]'. That reference concerns Bessel and Dunkl processes, while the present setting has a different drift structure and a (1+z^2) coefficient. The authors should provide the missing tightness/compactness argument or a precise reduction showing that the Cauchy transforms of μ_{N,t} converge to a solution of (5.4). As written, the proof of a central claim in the paper is conditional on an unstated adaptation.","section":"Section 5, proof of Theorem 5.1, 4th step"}],"minor_comments":[{"comment":"The displayed Carleman condition reads Σ (m̂_{2n}(t))^{-2n} = ∞. This is not the Carleman condition; the correct exponent is -1/(2n). The conclusion is still recoverable from the bound (5.11), so this appears to be a typographical error, but it should be corrected.","section":"Section 5, 3rd step"},{"comment":"In case (2), the text says 'then ̃x_{(N+1)/2,t}=0'. Since ̃x_{j,t}=x_{j,t}^2+1 and the middle particle is x_{(N+1)/2,t}=0 by symmetry, the value should be 1, not 0.","section":"Lemma 2.9, odd case"},{"comment":"The summation index in the frozen ODE is written 'k:j≠j'; it should be 'k:k≠j'. There is also a typo in the first paragraph: 'X_{1,t} < X_{1,t} < ... < X_{N,t}' should start with X_{2,t}.","section":"Introduction, Eq. (1.4)"},{"comment":"In the proof, the sentence 'by Theorem 3.1, the corresponding empirical measures satisfy...' refers to no Theorem 3.1 in this paper; it should presumably be Theorem 5.1. Please fix the cross-reference.","section":"Proof of Theorem 5.3"}],"recommendation":"major_revision","confidential_remarks":"The paper is potentially valuable and the technical core is largely explicit, but I cannot recommend acceptance until the Section 6 repair of Graczyk–Malecki is proved in detail. The authors may well be able to supply the missing verification; if so, the revised version could be acceptable. The delegation of part of Theorem 5.1's proof to [VW1] also needs to be replaced with a self-contained argument or a clearly justified reduction."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a solid, useful paper. The authors extend Assiotis's β=2 Hua-Pickrell diffusion theory to general β, and the headline results — β-independent empirical-measure limits, the freezing CLT, and the explicit free-convolution description at a=-N, b=0 — are new and largely correct. The pseudo-Jacobi connection is well worked out; Lemma 2.5's proof via elementary symmetric polynomials is clean.\n\nI checked the one spot that worried me: the appendix's repair of Graczyk–Malecki. The stress-test says (4.6') drops drift terms b_i ∂_{λ_i} e_n(A). That objection does not survive contact with the algebra: on the event V_n=e_n(A)=0, all n-subsets of squared differences have at least one zero factor, so any derivative term contains either a zero (λ_i-λ_j) or a zero squared difference. Hence the drift terms vanish on the boundary. The appendix is too terse — 'it can be easily checked' is carrying a real observation — but it is not a load-bearing gap. It should be expanded for publication, not rejected.\n\nWhat the paper does less well: it leans heavily on the same group's earlier results. Theorem 5.1's PDE step is delegated to [VW1], and the free process existence in Lemma 5.5 is a quick appeal to [CD]. That is legitimate, but it lowers self-containedness. The moment ODEs in Theorem 5.1 are derived explicitly, and the Carleman-condition step is spelled out — good to see.\n\nThe freezing CLT is the standard Laplace-method argument, with the positivity of the Hessian established via comparison with [HV]. The determinant formula in Corollary 4.5 is a nice bonus.\n\nWho should read it: people working on β-Pearson diffusions, Hua-Pickrell measures, or freezing limits of random matrix models. It deserves a serious referee; the main things to check are the corrected appendix identities and the positivity argument in the CLT. I would accept it for review.","headline":"Solid extension of Hua-Pickrell diffusions to general β; the Appendix's Graczyk–Malecki repair is terse but the flagged drift-term objection doesn't hold.","tokens_in":33222,"tokens_out":7074,"would_cite":true,"duration_ms":65329,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60B20","60F05","60F15","33C45","60K35","70F10","82C22"],"pacs":[],"model":"deepseek-v4-flash","headline":"The Hua-Pickrell diffusions have a β-independent large-N limit whose moments solve a closed ODE system anchored at pseudo-Jacobi zeros.","keywords":["Hua-Pickrell diffusions","pseudo-Jacobi polynomials","Pearson diffusions","empirical measures","moment ODEs","free convolution","freezing CLT","Cauchy ensembles"],"falsifier":"Simulate the SDE for β=1 and solve the frozen ODE for β=∞ from the same large-N initial empirical measure (e.g., N=200, a_N=2N, b_N=N) and compare empirical moments m_2(t) and m_4(t) at t=1 and t=2; if the β=1 and β=∞ curves do not approach the same limit within finite-N and Monte Carlo error, the β-independence claim of Theorem 5.1 is false.","tokens_in":32351,"feed_emoji":"🎲","tokens_out":11102,"duration_ms":113241,"temperature":0.7,"pith_summary":"This paper studies Hua-Pickrell diffusions, N-particle systems on the line with diffusion coefficient √(2(1+x²)) and pairwise drift (1+x_j x_k)/(x_j−x_k). After rescaling time by β, the family extends from stochastic dynamics (finite β) to a deterministic 'frozen' ODE (β=∞). The paper's central claim is that as N→∞, the empirical particle distributions converge almost surely to limits μ_t whose moments satisfy one closed system of ordinary differential equations, the same for every β∈[1,∞]; the frozen case is governed by the ordered zeros of pseudo-Jacobi polynomials. If correct, this gives a β-independent, explicitly computable description of the bulk spectrum of Hua-Pickrell/Cauchy ensembles: their relaxation to an equilibrium Hua-Pickrell measure for a>0, an explicit free-convolution formula in the symmetric a=-1,b=0 case, and a freezing central limit theorem with explicit eigenvalues. The proof mechanism is coordinate reduction: in elementary symmetric polynomial coordinates the frozen ODE becomes triangular and decays mode by mode, and the same reduction on moments produces the N→∞ hierarchy.","feed_headline":"One ODE system governs Hua-Pickrell spectra at every β","feed_subtitle":"Pseudo-Jacobi zeros anchor the frozen limit; one closed moment system covers every β.","key_machinery":"The load-bearing object is the change of coordinates from the ordered particle positions to the elementary symmetric polynomials y_m=e_N^m(x_1,…,x_N). In these coordinates the frozen ODE (2.1) becomes the triangular linear system d y_m/dt = −m(2a+m−1)y_m + lower-order terms, so the dynamics decays mode by mode and the terminal configuration is forced to be the ordered pseudo-Jacobi zeros. The same idea applied to empirical moments m_n(t)=∫x^n dμ_t produces the closed hierarchy (5.3), and because this hierarchy arises for every β after the time rescaling, the large-N limit is β-independent. A second coordinate change, x↦sinh x, brings the freezing CLT covariance into a matrix with known eigen","core_discovery":"The paper establishes that the β-Hua-Pickrell diffusions form a single family across β∈[1,∞] after the time change t↦t/β. For β=∞ the SDE degenerates into a deterministic ODE whose solutions exist uniquely, never collide, and—for a>0—converge to the ordered zeros of the pseudo-Jacobi polynomial P_N(·;−(N+a),b). For finite β the same SDE has unique strong solutions and a unique invariant measure, the Hua-Pickrell (Cauchy) measure with density proportional to ∏(1+x_j²)^{β(1−N−a)/2−1} e^{βb arctan x_j}∏|x_j−x_k|^β. The central large-N result is Theorem 5.1: under mild scaling of a,b and convergence of the initial empirical measures, the empirical measures μ_{N,t} converge weakly almost surely t","pith_inferences":["The β-independence of the N→∞ moment hierarchy suggests that all β-dependence in the bulk is confined to a lower-order fluctuation scale; one could test whether the covariance of √N(μ_{N,t}−μ_t) scales inversely with β in the manner predicted by the freezing CLT.","The moment ODEs (5.3) are the traced form of a free Itô equation for a self-adjoint operator; this hints that the limiting measures are not only limits but exact spectral distributions of a free Hua-Pickrell process, so the free-convolution formula should hold as an identity in free probability rather than only asymptotically.","The same deterministic flow could serve as a numerical benchmark: solve the triangular moment ODEs and compare with SDE simulations to locate where finite-β corrections first appear as functions of N and β.","The pseudo-Jacobi-zero parametrisation of the stationary states may carry over to finite N and finite β via the Hua-Pickrell measures' density, suggesting an orthogonal-polynomial duality that could yield exact transition probabilities or correlation functions."],"forward_implications":["For fixed β∈[1,∞] and large N, the empirical spectrum of the Hua-Pickrell diffusion follows the same deterministic moment ODEs; the bulk limit is a law of large numbers with no β dependence.","When â>0, the large-N empirical measures converge as t→∞ to the explicit equilibrium Hua-Pickrell measure μ_{HP,â,b̂}, with a relaxation rate determined by the exponents in the triangular moment system.","The freezing CLT gives explicit eigenvalues k(a+(k−1)/2) for the inverse covariance in trigonometric coordinates, so the β→∞ fluctuations of Hua-Pickrell ensembles are quantitatively tied to pseudo-Jacobi zeros and their discrete orthogonal polynomials.","In the symmetric case â=−1,b̂=0, the limiting measures are given by explicit free additive and multiplicative convolutions, providing closed-form predictions for simulation.","The Cauchy transforms of μ_t solve the PDE ∂_t G = −∂_z((−2(a+1)z+2b)G+(z²+1)G²), giving a continuum description that preserves compact support on finite time intervals."],"fun_headline_variants":["One ODE links Hua-Pickrell across all β","Infinite-β freezing reveals pseudo-Jacobi zeros","Unified moment system spans β=1 to ∞","Hua-Pickrell diffusions: same ODE at β=∞ and β<∞","Pseudo-Jacobi zeros emerge from frozen Hua-Pickrell"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"In Section 6, the proof of existence and uniqueness of solutions invokes an existing non-colliding SDE criterion whose assumptions require a certain interaction kernel to be nonnegative; here the kernel is 2(1+xy), which changes sign, and the paper's repair of the criterion's identities is only partially verified.","fun_headline_variants_meta":{"raw":{"variants":["One ODE links Hua-Pickrell across all β","Infinite-β freezing reveals pseudo-Jacobi zeros","Unified moment system spans β=1 to ∞","Hua-Pickrell diffusions: same ODE at β=∞ and β<∞","Pseudo-Jacobi zeros emerge from frozen Hua-Pickrell"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001262,"raw_usage":{"total_tokens":5130,"prompt_tokens":998,"completion_tokens":4132,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":742,"completion_tokens_details":{"reasoning_tokens":4058}},"tokens_in":742,"tokens_out":4132,"duration_ms":33112,"temperature":1.0,"reasoning_tokens":4058,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T05:58:17.657439+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate the SDE for β=1 and solve the frozen ODE for β=∞ from the same large-N initial empirical measure (e.g., N=200, a_N=2N, b_N=N) and compare empirical moments m_2(t) and m_4(t) at t=1 and t=2; if the β=1 and β=∞ curves do not approach the same limit within finite-N and Monte Carlo error, the β-independence claim of Theorem 5.1 is false.","supporting_citations":[],"review_version":1}