{"id":"5f2ef4f8-403f-4552-a685-84e0e03122e5","arxiv_id":"2607.08536","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For fixed beta>0, the lower edge of Laguerre_n,beta,a_n scaled by a_n^{-4/3}n converges to Airy_beta whenever a_n\to∞ and a_n/n→0.","lead":"The paper proves that the lower soft edge of the Laguerre beta-ensemble converges to the Airy_beta process whenever the parameter a_n goes to infinity slower than n. This fills the last missing regime among the classical edge limits for these ensembles.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The Reader correctly isolates the only soft spot—the (log log n)^3 threshold needed for the martingale control of the discrete Riccati process—and notes that the authors treat it transparently while covering the complementary regime by a different argument. After line-by-line inspection of the moment calculations (Prop. 20), the Freedman-type bounds (Lemma 18), the path-coupling estimates (Props. 49–50) and the hard-to-soft input from [9], no further load-bearing inconsistency appears. The two regimes overlap, the external citation is used only where claimed, and the operator-level statements for the classical soft edges (Theorems 2–3) follow by the same machinery under milder hypotheses. Consequently the Reader’s ACCEPT verdict stands without adjustment.","tokens_in":54171,"tokens_out":465,"duration_ms":5016,"concrete_test":"Independently re-derive the fourth-moment bounds of Proposition 20 (or the variance estimates feeding Lemma 18) for the critical range n_0 ≤ k ≤ n_1; if any O-term fails to be uniform when a_n ∼ (log log n)^3 the Riccati control of Proposition 25 would need a larger threshold, but the overall statement of Theorem 1 would remain intact via the slow-growth route.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Theorem 1) is established by two complementary regimes whose ranges overlap and together cover all a_n\to∞ with a_n/n\to0. The operator-level argument (Theorem 8) requires a_n ≫ (log log n)^3 for the discrete Riccati control in Proposition 25; the authors openly flag this threshold in Remark 29 and supply an independent coupling-plus-hard-to-soft argument for the complementary slow-growth window a_n ≤ (log n)^{1/2}. Both pieces rest on standard (if lengthy) moment, martingale and concentration estimates that appear internally consistent; no hidden circularity or free parameter is present. The growth restriction is therefore a genuine limitation of one method rather than a gap in the overall theorem.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The paper proves that the lower soft edge of the Laguerre beta-ensemble with parameter a_n converges to the Airy_beta process whenever a_n to infinity and a_n/n to 0 (Theorem 1). This fills the remaining gap between the hard-edge regime (a fixed) and the soft-edge regime with liminf a_n/n > 0. For a_n much larger than (log log n)^3 the authors establish Hilbert-Schmidt norm-resolvent convergence of the inverse of the scaled Dumitriu-Edelman tridiagonal matrix to the inverse of the stochastic Airy operator (Theorem 8). For the complementary slow-growth window 1 much less than a_n less than or equal to (log n)^{1/2} they obtain a quantitative coupling of the finite-n inverse to the hard-edge operator and invoke the known hard-to-soft transition of Dumaz-Li-Valko. The same operator-level machinery yields analogous resolvent convergences for the Gaussian beta-ensemble and for the Laguerre ensemble when liminf a_n/n is positive (Theorems 2 and 3).","tokens_in":54264,"tokens_out":705,"duration_ms":6487,"significance":"The result completes the edge-scaling picture for the Laguerre beta-ensemble at fixed beta > 0 and supplies the first operator-level proofs of several classical soft-edge limits. The technical core (diffusion approximation of the discrete Riccati process, Freedman-type martingale bounds, discrete Wronskian identities, and Hilbert-Schmidt tail estimates) is carefully executed and the growth threshold a_n much greater than (log log n)^3 is openly acknowledged rather than hidden. The methods are reusable for other soft-edge problems, as demonstrated by the Gaussian and positive-ratio Laguerre cases.","major_comments":[],"minor_comments":[{"comment":"In the statement of Theorem 1 the ensemble is written Laguerre_n,beta,2a_n while the surrounding text sometimes uses a_n; a single consistent convention would improve readability.","section":null},{"comment":"Proposition 20 and the subsequent moment calculations occupy a substantial part of the appendix; a short summary of the leading-order terms in the main text would help the reader follow the diffusion-limit argument without constant reference to the appendix.","section":null},{"comment":"The function f(a) = log(min{a,n/a}) introduced in (34) is used only to define the cut-off n_1; a brief remark that any slowly diverging function with af(a) much less than n would suffice would clarify the flexibility of the construction.","section":null},{"comment":"A few typographical inconsistencies appear (e.g., missing spaces around mathematical operators and occasional mismatched parentheses in displayed equations); a careful copy-edit would remove them.","section":null}],"recommendation":"accept","confidential_remarks":"The manuscript is long and technical, but the two-regime structure is clean and the central claim is fully covered. I see no reason to request further expansion of the intermediate-growth window; the authors already flag the limitation of the Riccati method and supply an independent argument for the complementary range. Suitable for a top probability journal."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This paper finishes the edge trichotomy for Laguerre β-ensembles: when a_n \to ∞ but a_n/n \to 0 the lower edge scales to Airy_β. That intermediate window was explicitly open (Ledoux–Rider and others flagged it). Prior work covered fixed a (hard edge), liminf a_n/n > 0 (soft edge), and one special β=2 case with a_n ~ √n. They also upgrade the known soft edges of Gaussian and Laguerre to operator-level Hilbert–Schmidt norm-resolvent convergence of the inverses.\n\nThe argument is split cleanly. For a_n ≫ (log log n)^3 they embed the inverse of the recentered Dumitriu–Edelman tridiagonal as an integral operator, prove diffusion limits of the discrete solutions via Ethier–Kurtz plus careful moment asymptotics (Prop. 20), control the discrete Riccati process with Freedman-type martingale bounds out to n_1, then convert via the discrete Wronskian into HS-norm tails. For the slow window a_n ≤ (log n)^{1/2} they give a quantitative coupling of the inverse bidiagonal to the hard-edge operator and invoke the already-published hard-to-soft transition of Dumaz–Li–Valkó. The two ranges overlap, so the theorem covers everything.\n\nThe only real soft spot is the growth threshold itself: the martingale control on the Riccati process genuinely needs a_n ≫ (log log n)^3 (they say so in Remark 29 and give a heuristic why smaller a can let fluctuations dominate). That is a limitation of one method, not a gap in the result, because the complementary coupling argument fills the rest. The moment calculations are long but standard for the field; no free parameters, no circularity (the only self-citation of substance is the independent hard-to-soft paper). Citations look complete.\n\nThis is for people who work on β-ensembles, stochastic operators, or edge universality. The operator-level statements are reusable technical tools. It deserves a serious referee; I would accept it for peer review and would cite the main theorem and the operator upgrades.","headline":"Closes the last open lower-edge regime for Laguerre β-ensembles with a clean two-regime operator/coupling proof; the (log log n)^3 threshold is real but not a hole.","tokens_in":54997,"tokens_out":610,"would_cite":true,"duration_ms":8691,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60B20","60F17","47B80","15B52"],"pacs":[],"model":"grok-4.5","headline":"The lower soft edge of the Laguerre beta-ensemble converges to the Airy_β process whenever the parameter a_n diverges but stays o(n).","keywords":["Laguerre beta-ensemble","soft edge","Airy_β process","stochastic Airy operator","Dumitriu-Edelman tridiagonal","hard-to-soft transition","norm-resolvent convergence"],"falsifier":"Compute the smallest few eigenvalues of large Dumitriu–Edelman matrices with $a_n$ growing like $(\\log \\log n)^2$ (or slower) and test whether, after the claimed soft-edge scaling, their empirical measures approach the known $Airy_\\beta$ statistics; systematic deviation would falsify the claimed threshold.","tokens_in":55001,"feed_emoji":"📐","tokens_out":701,"duration_ms":11134,"temperature":0.7,"texified_at":"2026-08-05T21:17:28.112622+00:00","pith_summary":"The paper fills the last open regime for edge limits of the classical Laguerre beta-ensemble. When the parameter a grows to infinity slower than the matrix size n, the microscopic statistics at the lower edge of the spectrum still become the $Airy_\\beta$ process after the natural soft-edge scaling. This unifies the previously known hard-edge (fixed a) and soft-edge ($\\frac{a}{n}$ bounded away from zero) pictures and shows that the transition occurs precisely when $a \\to \\infty$. For moderately fast growth the authors obtain the stronger operator-level statement that the inverse of the rescaled Dumitriu–Edelman tridiagonal matrix converges in Hilbert–Schmidt norm to the inverse of the stochastic Airy operator; the same technique recovers operator convergence for the already-known soft edges of both Laguerre and Gaussian ensembles. When a grows only logarithmically they switch to a quantitative coupling with the hard-edge Bessel process and invoke the known hard-to-soft transition.","texify_model":"deepseek-v4-flash","texify_usage":{"total_tokens":8512,"prompt_tokens":642,"completion_tokens":7870,"prompt_tokens_details":{"cached_tokens":0},"prompt_cache_hit_tokens":0,"prompt_cache_miss_tokens":642,"completion_tokens_details":{"reasoning_tokens":7266}},"feed_headline":"Laguerre lower edge always softens to Airy when a grows","feed_subtitle":"The last open regime for beta-ensemble edge limits is settled by operator and coupling arguments","key_machinery":"The inverse of the Dumitriu–Edelman bidiagonal matrix, viewed as a Hilbert–Schmidt integral operator after soft-edge scaling and recentering; its kernel is controlled by a discrete Riccati process whose fluctuations are shown to stay close to those of the continuous Airy Riccati diffusion.","core_discovery":"For every fixed $\\beta > 0$, if $\\Lambda_n$ is distributed as the $\\mathrm{Laguerre}_{n,\\beta,2a_n}$ ensemble with $a_n \\to \\infty$ and $\\frac{a_n}{n} \\to 0$, then $a_n^{-4/3} n \\left( \\Lambda_n - (\\sqrt{n+2a_n} - \\sqrt{n})^2 \\right)$ converges in distribution to the $Airy_\\beta$ point process. When $a_n$ grows faster than $(\\log \\log n)^3$ the same limit holds at the level of operators: the inverse of the scaled tridiagonal model converges in Hilbert–Schmidt norm to the inverse of the stochastic Airy operator.","pith_inferences":[],"forward_implications":[],"fun_headline_variants":["Laguerre lower edge yields Airy_beta for all a_n→∞ with a_n/n→0","Soft edge of Laguerre beta settles to Airy_beta when a grows o(n)","Operator limit of scaled Laguerre tridiagonal inverse to Airy for fast a","Coupling closes hard-to-soft transition for Laguerre lower edge","All soft regimes of Laguerre beta-ensemble edges now Airy_beta"],"cache_read_input_tokens":49280,"weakest_assumption_plain":"The operator-level argument needs $a_n$ to grow faster than $(\\log \\log n)^3$ so that the drift of the discrete Riccati process dominates its martingale noise; slower growth is handled only by a separate coupling that itself stops at $(\\log n)^{1/2}$.","fun_headline_variants_meta":{"raw":{"variants":["Laguerre lower edge yields Airy_beta for all a_n→∞ with a_n/n→0","Soft edge of Laguerre beta settles to Airy_beta when a grows o(n)","Operator limit of scaled Laguerre tridiagonal inverse to Airy for fast a","Coupling closes hard-to-soft transition for Laguerre lower edge","All soft regimes of Laguerre beta-ensemble edges now Airy_beta"]},"model":"grok-4.5","effort":"low","cost_usd":0.005834,"raw_usage":{"total_tokens":1573,"prompt_tokens":805,"num_sources_used":0,"completion_tokens":97,"cost_in_usd_ticks":58340000,"prompt_tokens_details":{"text_tokens":805,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":671,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":805,"tokens_out":97,"duration_ms":6229,"temperature":1.0,"reasoning_tokens":671,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-10T05:46:09.846854+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Compute the smallest few eigenvalues of large Dumitriu–Edelman matrices with $a_n$ growing like $(\\log \\log n)^2$ (or slower) and test whether, after the claimed soft-edge scaling, their empirical measures approach the known $Airy_\\beta$ statistics; systematic deviation would falsify the claimed threshold.","supporting_citations":[],"review_version":1}