{"id":"e65b3006-7f52-44d1-aaea-a07f0c40d8dd","arxiv_id":"2411.10829","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The Airy_beta line ensemble is constructed for all beta>0 via explicit multi-time Laplace transform formulas, and it is shown to be the edge scaling limit of both the Dyson Brownian Motion and the Gaussian beta corners process.","lead":"The paper defines a new universal object, the Airy_beta line ensemble, a collection of random curves that should describe the edge statistics of many random matrix models. It proves that two different random matrix processes converge to this ensemble and gives explicit formulas for its Laplace transform moments.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Uniqueness in Theorem 1.4 is the load-bearing point: the paper concedes the one-time moment problem is indeterminate, so equality of joint moments for subsequential limits cannot by itself identify a unique Airy_β line ensemble.","rationale":"The central claim of the paper is Theorem 1.4: existence and uniqueness of the Airy_β line ensemble characterized by the joint Laplace-transform moments L_beta. The convergence theorems 1.5 and 1.6 are meaningful only if this limit is unique; otherwise different subsequential limits could produce different line ensembles with identical moments. The reader correctly identifies the uniqueness step in Section 5.3 as the weakest assumption. My stress-test agrees: the paper explicitly acknowledges that the classical moment problem for the individual exponential variables is indeterminate, so the indirect argument must do more than match joint moments. The deterministic injectivity of the map from a configuration to its exponential sums is a red herring, because the paper provides the joint moments of the random sums, not their full law. Without a rigorous proof of moment determinacy for the process of these sums (or an external uniqueness theorem such as the SDE characterization in [HZ24] applied to all subsequential limits), Theorem 1.4 is unproven. I found no other internal inconsistency or obvious error in the walk-expansion machinery; the estimates in Propositions 4.38 and 4.44 are lengthy and were not independently checked, but they are not as structurally load-bearing as the uniqueness claim. The reader's conditional acceptance is therefore appropriate: the paper is impressive and likely correct, but the uniqueness assertion must be fully substantiated before unconditional acceptance.","tokens_in":85517,"tokens_out":17070,"duration_ms":193524,"concrete_test":"Obtain the complete proof of Section 5.3 and isolate the step that upgrades 'all subsequential limits have joint moments L_beta' to 'subsequential limits have the same law'. Then independently re-derive that step: prove that the joint moments (5) determine the law of an ordered, continuous, stationary process {A_i}. If the proof relies on moment determinacy of the exponential sums, test its hypotheses by checking the growth of E[S_k(τ)^n] using the bounds in Section 5.1 and Theorem 4.1; if the growth violates a classical sufficiency criterion such as Carleman's condition, the argument must explicitly use additional structure (ordering, stationarity, continuity in τ). If the proof instead invokes uniqueness of the SDE in [HZ24], verify that the subsequential limits satisfy the SDE and that its solution is unique for all β>0.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 5.3 must prove that the family of joint Laplace-transform moments in (5) determines a unique law on the space of ordered continuous curves. The paper itself warns (Section 5 introduction) that the classical moment problem for the variables exp(k A_i^β(τ)/2) is not determined because the moments grow too fast. Therefore, showing that every subsequential limit of DBM or GβE corners has the same joint moments L_beta is insufficient: two different laws for the curves could share all these moments unless an additional determinacy mechanism is supplied. The deterministic injectivity of the map from a point configuration to its exponential sums is not enough, because the paper provides only the joint moments of those sums, not their full Laplace functional. The indirect argument must either prove moment determinacy for the stochastic process S_k(τ)=Σ_i exp(k A_i^β(τ)/2) (despite the classical indeterminacy) or invoke an external uniqueness result (e.g., the SDE characterization in [HZ24]) and verify its hypotheses. If neither is done, Theorem 1.4's uniqueness assertion remains unsupported, and the convergence theorems 1.5–1.6 would only identify subsequential limits, not a unique Airy_β line ensemble.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a candidate universal edge-scaling object, the Airy_β line ensemble, and characterizes it by explicit integral formulas for its joint multi-time Laplace-transform moments. Theorem 1.4 asserts that for every β>0 there is a unique ordered family of continuous stationary processes whose moments are given by the block integral L_β of Definition 2.12. Theorems 1.5 and 1.6 assert that the GβE corners process and the Dyson Brownian Motion converge to this object in the edge-scaling limit. The method computes high moments through Dunkl-type differential-difference operators, expands them as sums over discrete walks, and then passes to a Brownian scaling limit with cancellations and regularizations, encoded in the block combinatorics of Section 2.","tokens_in":85767,"tokens_out":8164,"duration_ms":83972,"significance":"If the main theorems are correct, this is a substantial contribution: it provides the first construction of a general-β Airy line ensemble with explicit multi-time Laplace-transform formulas, unifying the Tracy-Widom laws, the stochastic Airy operator, and the Airy_2 process, and it gives two separate convergence theorems for β>0. The paper has real strengths: the Dunkl-operator identities in Theorems 3.2 and 3.4 are elegant and exact; the limiting object is defined independently of both prelimit models, so there is no fitted-parameter circularity; the random-walk expansion and cancellation mechanism are developed with considerable care; and the claimed formulas are analytic in β, which is new for these quantities. However, the verification burden is exceptionally high, and two load-bearing parts of the proof — the uniqueness argument in §5.3 and the DBM adaptation in §4.10 — are not supplied at the same level of detail as the rest of the paper.","major_comments":[{"comment":"The uniqueness assertion is load-bearing and is not proved in the material supplied. The introduction to Section 5 explicitly notes that the classical moment problem for the variables exp(k A_i^β(τ)/2) is not determined because the moments grow too fast. Therefore convergence of the joint moments (5) for every subsequential limit cannot, by itself, identify a unique law on the space of ordered continuous curves; an additional determinacy mechanism is required. The text says §5.3 contains an “indirect argument,” but the version I received breaks off inside §5.2, so that argument could not be checked. If the argument only shows that all subsequential limits have the same joint moments L_β, then Theorems 1.5 and 1.6 would identify only subsequential limits, and Theorem 1.4 would not be established. Please supply the full uniqueness proof and specify whether it uses a process-level moment-determinacy theorem for S_k(τ)=∑_i exp(k A_i^β(τ)/2) or an external characterization such as the SDE result of [HZ24], with all hypotheses verified.","section":"§5.3, Theorem 1.4"},{"comment":"The Dyson Brownian Motion case is described as “almost the same” as the corners process, with a list of substitutions in the weight formula (81) and a statement that Sections 4.8 and 4.9 go through verbatim, mutatis mutandis. This is not sufficient for a proof at the level of rigor used elsewhere in the paper. The corners proof depends on delicate cancellation mechanisms for blow-up terms (Section 4.6), on elimination of type III and type B indices (Section 4.7), and on the discrete-blocks approximation of Proposition 4.38; these analyses are sensitive to the signs, weights, and index sets that change in the DBM case. Theorem 1.6 is one of the two central convergence theorems, so the DBM adaptation needs to be written out in enough detail to verify that Proposition 4.25 and its analogues hold unchanged.","section":"§4.10, Proposition 4.4"},{"comment":"The paper’s route from moment convergence to distributional convergence depends on both the fourth-moment estimates of Propositions 5.1 and 5.2 and on the topological statement of Proposition 5.5. The proofs of Propositions 5.1 and 5.2 are more compressed than the rest of the paper: they invoke Proposition 5.3 and Corollary 5.4, then assert that the double/composite estimates for the four mixed fourth-moment terms combine via Corollary 5.4 to give the required (τ_2−τ_1)^2 bound. Given that the tightness and continuity of the limiting process rest on these estimates, the intermediate bounds used in the linear combination should be displayed explicitly. As written, this part is plausible but not fully checkable.","section":"§5.1–5.3, moment-to-process upgrade"}],"minor_comments":[{"comment":"The exponent in the second line of (35) is typeset as N^{(H(Q_{\\ell-1})-H(Q_\\ell))/2-|\\{t:...\\}|}, which is ambiguous: the subtraction should be displayed in parentheses or split into two factors.","section":"Eq. (35)"},{"comment":"The normalized edge processes in Theorems 1.5 and 1.6 carry an explicit factor 2N^{2/3}, while the processes y_i^{(N)}(τ) and Y_i^{(N)}(τ) defined in Section 5 are N^{2/3}(...−1) without that factor. Since the limiting moments in (5) are for A_i^β/2, the factor-of-two conventions should be aligned and stated once, near the statements of Theorems 1.5 and 1.6.","section":"Theorems 1.5/1.6 and Section 5"},{"comment":"In the parametrization of block processes, the sentence “we should choose ∑δ_{j,ℓ}+∑δ_{j,ℓ} reals” appears to duplicate the same sum; the intended counting of discontinuity positions and jump sizes should be written out without repetition.","section":"Definition 2.7"},{"comment":"Calling the triplet “‘Blocks’” in Definition 2.6 is stylistically unusual for a formal definition; a non-quoted name such as “block structure” would avoid confusion with the word “blocks” used informally in Figures and examples.","section":"Section 2.2, Definition 2.6"}],"recommendation":"major_revision","confidential_remarks":"The central construction and moment computations are impressive, and I do not see an internal contradiction in the main random-walk expansion. However, the uniqueness of the limiting process is exactly the point that makes the object well-defined, and the submitted text does not contain a verifiable proof of it. I would urge the editor to have the uniqueness section and the DBM adaptation checked independently before deciding; if the §5.3 argument exists in the full arXiv version and was truncated only in the review copy, my concern should be treated as a request to expand rather than as a claim that the result is false."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nBottom line: this is a large, serious paper and I'd take the main results seriously. For each beta>0 it defines an Airy_beta line ensemble via explicit joint multi-time Laplace-transform moments (Definition 2.12), proves existence and uniqueness (Theorem 1.4), and shows both the GbetaE corners process and DBM edge scaling limits converge to it (Theorems 1.5 and 1.6). The formulas are analytic in beta and new even for the one-time marginal. That is a real step: it unifies Tracy-Widom distributions, stochastic Airy operator spectra, and the Airy_2 process, and it gives the first general-beta edge convergence for the corners process.\n\nThe method matters too. The Dunkl-operator walk expansions leading to block structures and Brownian bridge weights are original, and the paper works hard to control the signed sums, dealing with blow-up via cancellations between type I/II and III/B indices. Sections 4.3-4.9 look genuinely detailed; the authors identify where absolute convergence holds and define L_beta as a principal value integral. No fitting, no circularity: the limiting object is defined first and the convergence theorems reproduce it.\n\nThe soft spots are where a referee should spend time. Section 4.10 (DBM adaptation) is sketched rather than fully proved; acceptable if the modifications really are parallel. More important: uniqueness in Theorem 1.4 relies on the indirect argument in Section 5.3, and the paper itself concedes the classical one-time moment problem for exp(k A_i^beta(tau)/2) is indeterminate. The stress-test note is fair: showing all subsequential limits have the same joint moments is not automatically enough unless the paper proves a determinacy result for the exponential sums or invokes an external uniqueness theorem, such as the SDE characterization in [HZ24]. I did not see that spelled out; it needs a close check. If that step is solid, this is a major contribution; if not, the convergence theorems only identify subsequential limits.\n\nThere are many technical estimates in Section 4 (Propositions 4.38 and 4.44 among others) that I cannot certify by reading once. The paper is careful about implicit constants and the structure is coherent, so I am not especially worried, but this is referee work.\n\nWho is this for? Specialists in beta ensembles, KPZ universality, and random matrix edge limits. It deserves a serious referee and should not be desk-rejected. My recommendation: send it to peer review, with a specific request that one referee check Section 5.3's uniqueness argument and the DBM outline.","headline":"Important and careful: a first construction of the general-beta Airy line ensemble with explicit moment formulas and two convergence theorems; the technical core is credible, and the only soft spot that matters is the outlined uniqueness step in Section 5.3.","tokens_in":86298,"tokens_out":2754,"would_cite":true,"duration_ms":30925,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60B20","60F17","60J65","82B23"],"pacs":[],"model":"deepseek-v4-flash","headline":"For every $\\beta>0$ there is a unique ordered family of continuous curves whose joint Laplace-transform moments equal the explicit integral $L_\\beta$, and both the $G\\beta E$ corners process and Dyson Brownian Motion converge to it.","keywords":["Airy beta line ensemble","random matrix edge universality","multi-time Laplace transform","Gaussian beta ensemble corners","Dyson Brownian Motion","Dunkl operators","random walk expansion","Brownian bridge blocks"],"falsifier":"Construct, for fixed $\\beta>0$ (say $\\beta=3$), two distinct ordered stationary continuous processes whose joint Laplace-transform moments both equal $L_\\beta(\\vec{k},\\vec{\\tau})$ for all $m$, $\\vec{k}$, $\\vec{\\tau}$; that would directly falsify Theorem 1.4. A numerical check is to evaluate $L_\\beta$ for $m=2$ with Brownian-bridge kernels and compare with high-$N$ simulation of the DBM edge at two times: a mismatch would show the principal-value integral does not encode the actual limit.","tokens_in":85309,"feed_emoji":"📈","tokens_out":9652,"duration_ms":94775,"temperature":0.7,"pith_summary":"The paper defines, for every $\\beta>0$, a single infinite family of random ordered continuous curves — the Airy$\\beta$ line ensemble — and proves that it is the edge scaling limit of the largest eigenvalues in two different two-dimensional extensions of the Gaussian $\\beta$-ensemble. The object is characterized by exact joint multi-time Laplace-transform moments, given by the explicit integral $L_\\beta(\\vec{k},\\vec{\\tau})$ of Definition 2.12, rather than by an infinite system of stochastic differential equations. The main theorems show that both the $G\\beta E$ corners process and the Dyson Brownian Motion converge to this same family in distribution on compact time intervals. If correct, this genuinely unifies the general-$\\beta$ Tracy–Widom distribution, the eigenvalues of the stochastic Airy operator, and the Airy$_2$ process from KPZ theory as marginals of one object that interpolates smoothly in $\\beta$.","feed_headline":"One line ensemble unifies random-matrix edges for every beta","feed_subtitle":"Explicit multi-time Laplace formulas define the Airy-beta ensemble; two major models provably converge to it.","key_machinery":"Three mechanisms carry the argument. Dunkl differential-difference operators $D_i^N$ act diagonally on multivariate Bessel functions, turning expectations of products of eigenvalue power sums into algebraic expressions (Theorems 3.2 and 3.4). Expanding high powers of these operators produces signed sums over decorated random walks; the edge rescaling sends the walks to Brownian bridges, with cancellations among large terms — encoded in the type I/II and III/B classifications — being exactly what makes the limiting sums conditionally convergent. The surviving terms are organized into 'blocks': a block process, a virtual block process, and a block height, with weights built from Brownian-bridge, Bessel-3, and Brownian-excursion kernels $I$, $I_0$, $I_{0,0}$. The block integral in Definition 2.12 is the continuum limit of the discrete walk sums, and it directly defines $L_\\beta$.","core_discovery":"The central assertion is Theorem 1.4: for each $\\beta>0$ there is a unique ordered family of stationary continuous processes $\\{A_i^\\beta(\\tau)\\}_{i=1}^\\infty$ whose joint Laplace-transform moments $$\\mathbb{E}\\left[\\prod_{\\ell=1}^m\\left(\\sum_{i=1}^\\infty \\exp(k_\\ell A_i^\\$\\beta$(\\tau_\\ell)/2)\\right)\\right]$$ equal the explicitly defined function $L_\\beta(\\vec{k},\\vec{\\tau})$, and the trajectories are almost surely real-valued. The same family is the distributional limit, uniformly on compact time sets and jointly in finitely many labels $i$, of the largest particles in the $G\\beta E$ corners process (Theorem 1.5) and of the Dyson Brownian Motion (Theorem 1.6). The coincidence of these two limits, previously known only for special $\\beta$, is taken as evidence that the Airy$\\beta$ line ensemble is the universal edge object for general-$\\beta$ random-matrix and 2d statistical mechanics models.","pith_inferences":["A direct test of the conjectured broader universality would be to run the same Dunkl-operator expansion for Laguerre or Jacobi $\\beta$-corners and check numerically that their high moments converge to the same $L_\\beta$; the paper does not prove this, but its method is built to extend that way.","The Section 5.3 uniqueness argument exists because the classical moment problem is indeterminate here; an implication is that additional structural axioms, such as a generalized Brownian-Gibbs property, may be needed to characterize the Airy$\\beta$ line ensemble independently of its realization as a specific limit.","The paper only handles monotone two-dimensional sections of the three-dimensional evolution indexed by $(N,\\tau)$; proving convergence of the full three-dimensional object would be a stronger universality statement and remains open."],"forward_implications":["For $\\beta=2$ the Airy$\\beta$ line ensemble is the Airy$_2$ line ensemble of KPZ theory, so the new Laplace formulas give a direct integral representation of multi-time Airy$_2$ statistics.","The one-time marginal $\\{A_i^\\beta(0)\\}$ recovers the general-$\\beta$ Tracy–Widom law and the eigenvalues of the stochastic Airy operator, embedding those classical edge objects in a single $\\beta$-continuous family.","Because the same limit appears from the corners process and from the Dyson Brownian Motion, edge universality for general $\\beta$ holds simultaneously in the matrix-size direction and in the time direction; the paper lists Laguerre and Jacobi corners, Macdonald processes, Jack–Gibbs measures, and non-intersecting walk models as expected further instances.","The principal-value integral defining $L_\\beta$ is proven finite for every $\\beta>0$, which gives a well-defined analytic formula whose dependence on $\\beta$ is explicit."],"supporting_citations":[{"why":"Supplies the commuting Dunkl operators used to expand eigenvalue power sums.","marker":"[Dun89]"},{"why":"Gives the eigenrelation of Dunkl operators on multivariate Bessel functions and the analytic continuation used in the moment formulas.","marker":"[Opd93]"},{"why":"Provides the Bessel generating function identity for the Gaussian beta-ensemble used to start the corners-process moment computation.","marker":"[BGCG22]"},{"why":"Defines the beta-corners process whose edge limit is Theorem 1.5.","marker":"[GS15]"},{"why":"Provides the DBM transition density representation used to express DBM multi-time moments through Bessel functions.","marker":"[BF97]"}],"fun_headline_variants":["Airy-beta line ensemble defined by Laplace transform for all beta","DBM and G beta E corners converge to same Airy-beta ensemble","Laplace formulas give universal Airy-beta edge for every repulsion","Airy-beta: explicit moments, two random-matrix limits coincide","Airy-beta ensemble: one universal edge via Laplace, proven for all beta"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing step is the indirect uniqueness proof in Section 5.3: the moments of the individual exponentials $\\exp(k A_i^\\beta(\\tau)/2)$ grow too fast for the classical moment problem to apply, so the paper must show that any subsequential limit sharing the joint Laplace transform $L_\\beta$ has the same law on the space of ordered continuous curves; if that uniqueness failed, the convergence theorems would only identify subsequential limits.","fun_headline_variants_meta":{"raw":{"variants":["Airy-beta line ensemble defined by Laplace transform for all beta","DBM and G beta E corners converge to same Airy-beta ensemble","Laplace formulas give universal Airy-beta edge for every repulsion","Airy-beta: explicit moments, two random-matrix limits coincide","Airy-beta ensemble: one universal edge via Laplace, proven for all beta"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00026,"raw_usage":{"total_tokens":1603,"prompt_tokens":973,"completion_tokens":630,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":589,"completion_tokens_details":{"reasoning_tokens":537}},"tokens_in":589,"tokens_out":630,"duration_ms":6919,"temperature":1.0,"reasoning_tokens":537,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T19:14:32.108791+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct, for fixed $\\beta>0$ (say $\\beta=3$), two distinct ordered stationary continuous processes whose joint Laplace-transform moments both equal $L_\\beta(\\vec{k},\\vec{\\tau})$ for all $m$, $\\vec{k}$, $\\vec{\\tau}$; that would directly falsify Theorem 1.4. A numerical check is to evaluate $L_\\beta$ for $m=2$ with Brownian-bridge kernels and compare with high-$N$ simulation of the DBM edge at two times: a mismatch would show the principal-value integral does not encode the actual limit.","supporting_citations":[],"review_version":1}