Pith. sign in

REVIEW 3 major objections 5 minor 9 references

Random Lozenge Waterfall: Dimensional Collapse of Gibbs Measures

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read The paper proves that fixed-q q-Racah lozenge tilings collapse to a one-dimensional random barcode interface inside the waterfall region, with exponential concentration.

desk verdict Rigorous theorem is real, but the exponential concentration proof relies on a figure-backed inequality that must be settled. read the letter →

arxiv 2507.22011 v1 pith:UL237S5E submitted 2025-07-29 math.PR cond-mat.stat-mechmath-phmath.MP

classification math.PRcond-mat.stat-mechmath-phmath.MP MSC 60K3582B2033D45
keywords lozengetilingsq-Racahmeasurewaterfallphasebarcodeprocessexponentialconcentrationdeterminantalpointdimensionalcollapsefixed-qregime
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper is trying to establish that the fixed-q regime of q-Racah weighted lozenge tilings is a genuinely new macroscopic phase, not a minor deformation of the usual q-to-1 regime. As the hexagon side lengths grow with L while q in (0,1) and kappa in iR stay fixed, the two-dimensional Gibbs-type liquid structure collapses: inside the waterfall region W the fixed-slice correlation kernel converges to the identity, outside W it converges to zero, and the probability of a horizontal lozenge inside W or a square or vertical lozenge outside W decays exponentially in L. The surviving randomness is a one-dimensional random stepped interface, called a barcode, which the paper describes through a conjectural explicit determinantal kernel. If the proof is right, this is the first rigorous example of a deterministic-weight tiling model whose bulk phase is simultaneously random and one-dimensional.

What carries the argument

The argument runs on two tracks. On one vertical slice, the kernel is the orthogonal spectral projection onto the positive part of an explicit difference operator built from q-Racah polynomials; taking the fixed-q limit of its coefficients gives the identity operator on the center line and $(-\mathrm{Id})$ outside $W$, by strong resolvent convergence. Inside $W$ and off the center line these limits are no longer self-adjoint, so a second track estimates the q-Racah orthogonal polynomial ensemble directly: the probability ratio for moving a single particle from $x$ to $y$ is controlled by an exponent $E_k$, which after minimizing over the other $N-1$ particles becomes a discrete integral of a piecewise linear function $H_k$, and positivity of an explicit function (5.23) supplies the decisive lower bound. For the conjectural barcode kernel, the load-bearing object is the limiting two-diagonal operator $U_t^{\mathrm{barcode}}$ and the functions $F_n$ orthogonal in both lattice spaces, which satisfy the same inter-slice identity as the pre-limit polynomials.

What would settle it

Evaluate (5.23) symbolically or on a fine grid for $0<x<4/5$ and $0\le y\le \min\{x,(4-5x)/3\}$ and find any point where it is negative or zero; that would invalidate Lemma 5.10 and the exponential concentration theorem. Conversely, verifying the numerically observed geometric convergence of the pre-limit barcode kernel at rate $q$ over a longer sequence would support the conjecture but would not settle the proof.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is Theorem 1.4: for side lengths scaling linearly in L and fixed q and kappa, at every macroscopic location the fixed-slice kernel $K_{\lfloor Lt\rfloor}(\lfloor Lx\rfloor+\Delta x,\lfloor Lx\rfloor+\Delta y)$ tends to $\mathbf{1}_{\Delta x=\Delta y}$ when $(t,x)$ lies in the waterfall region $W$ and to $0$ otherwise. Equivalently, all local correlations on a slice inside $W$ converge to one, and outside $W$ they vanish, with exponential bounds on the probability of the disfavored lozenge types. The geometric content is that the nonintersecting paths cluster into a saturation band of width $N$, and their local structure reduces to a single random one-dimensional interface, the barcode. The paper further conjectures that this interface is a determinantal point process with an explicit symmetric kernel built from functions orthogonal in both $\ell^{2}(\mathbb{Z})$ and $\ell^{2}(\mathbb{Z}+\tfrac12)$, and that this kernel is $2\times 2$ block Toeplitz, giving invariance under even shifts but not odd shifts.

Load-bearing premise

The exponential-concentration theorem rests on Lemma 5.10, whose proof asserts positivity of the explicit piecewise-linear function (5.23) on a certain triangle and justifies it only by appealing to a plot; if that positivity failed, the probability-ratio bound (5.19) and the hole-probability estimate Theorem 5.11 would not follow.

Editorial extensions

If this is right

  • In the fixed-q regime the two-dimensional Gibbs picture is replaced by dimensional reduction: fluctuations survive only along a one-dimensional interface.
  • All three lozenge types cannot coexist in the limit inside $W$: horizontal lozenges inside $W$, and square or vertical lozenges outside $W$, have exponentially small probability.
  • If the conjecture is right, the limiting barcode process is universal: it depends only on $q$ and $\kappa$, not on the hexagon side lengths or the observation point inside $W$.
  • The limiting kernel satisfies $K^{\mathrm{barcode}}(0,0)=1-K^{\mathrm{barcode}}(1,1)$, so the global density is $1/2$ and the process is invariant under shifts by $2\mathbb{Z}$ but not by $\mathbb{Z}$.
  • Numerical evidence indicates correlations in the barcode process decay exponentially in the separation $t$, in contrast to the polynomial decay seen in ergodic Gibbs measures.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the conjectural kernel is confirmed, the barcode process would be a rare determinantal point process with exponential correlation decay; a natural next step is to identify its scaling limits as $q\to 1$ or $q\to 0$.
  • The non-self-adjoint limiting operators on the boundary of $W$ suggest that a canonical self-adjoint extension, once identified, could produce the barcode kernel rigorously from spectral data rather than from a regularized divergent series.
  • The emergent period-two invariance may correspond to a hidden two-step Markov structure, since each vertical step adds exactly one horizontal lozenge that must be placed in one of the two frozen regions.
  • The fixed-q waterfall could also occur in other q-deformed dimer models with deterministic weights; testing the real and trigonometric q-Racah regimes could reveal how generic the phenomenon is.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper studies the q-Racah probability measure on lozenge tilings of a hexagon in the fixed-q regime, where the side lengths scale linearly with L while q and κ remain fixed. The main rigorous result, Theorem 1.4, asserts that on each vertical slice the fixed-slice correlation kernel converges to the identity inside the waterfall region W and to zero outside it, and that wrong-type lozenges have exponentially small probability. The proof combines a spectral projection argument for the center line and outside W with direct q-Racah orthogonal polynomial ensemble estimates for the interior of W. The paper also formulates Conjecture 1.5, an explicit determinantal barcode kernel for the conjectural one-dimensional limit, supported by numerical evaluation of exact formulas and by simulations from a ported perfect-sampling algorithm. The last section and Appendix A develop properties of the conjectural orthogonal functions F_n, including orthogonality and a q-Mehler completeness statement.

Significance. If Theorem 1.4 stands, the paper establishes a genuinely new macroscopic phase in an exactly solvable tiling model: dimensional collapse from two-dimensional Gibbs behavior to a one-dimensional random interface. The rigorous portion is built on exact q-Racah formulas with no fitted parameters, and the paper gives reproducible numerical evidence, Mathematica code, and a Python simulator. The conjectural barcode kernel is explicitly labelled as non-rigorous and is not needed for Theorem 1.4; its divergent-series regularization is clearly identified as an open technical obstacle. The main gap is a load-bearing but localized estimate in Lemma 5.10 whose proof currently rests on a numerical plot rather than an algebraic argument. This is a fixable gap within the manuscript's scope, provided the positivity assertion can be proven exactly.

major comments (3)
  1. [§5.3, Lemma 5.10, Eq. (5.23)] The proof of Lemma 5.10 reduces the key inequality (5.21) to positivity of the function f(x,y) = (3/4)(x-y) + (1-2x)_- - (1-2y)_- on the region 0 < x < 4/5, 0 ≤ y ≤ min{x, (4-5x)/3}, but the only justification is 'see Figure 8'. This inequality is load-bearing: (5.21) enters (5.19), which is used in Theorem 5.11 and Corollary 5.12 to prove the exponential concentration half of Theorem 1.4. An algebraic verification of (5.23) is required. Moreover, on the boundary segment y = (4-5x)/3 the limiting expression f is identically zero, so the word 'positive' cannot be interpreted as strict positivity there; the strict inequality (5.21) must be obtained either by an off-boundary argument plus uniformity, or by tracking the discrete corrections in G_k and the z0±1 terms in (5.14), rather than from the limiting function alone.
  2. [§5.3, proof of Lemma 5.10, Case 2] The step 'Since (5.23) is piecewise linear, we conclude that (5.21) holds for x > (S+t-N)/2 - γL' is not justified as written. The passage from the limiting expression (5.22)-(5.23) to the finite-L inequality (5.21) requires a uniform bound on the error terms introduced by replacing k with -1, by replacing z0-1 and z0+N-1 with the continuous boundary points, and by ignoring floors in the scaled variables. Because the limiting function vanishes on a boundary segment, the error terms cannot be absorbed by a uniform positive margin from (5.23) alone. The proof needs an explicit statement of the convergence rate and a demonstration that the O(L) replacement errors are strictly dominated by the 5/4 d(x;y) term in the regime considered.
  3. [§1.4 and §5.3] The exponential concentration statements for the region outside W are described in Section 1.4 as following from 'similar estimates' in Section 5.3, and the proof of Theorem 5.11 explicitly says that the analogous estimates for t outside [tl,tr] or for N > T are not formulated because they are not needed for the waterfall region. However, Theorem 1.4 as stated covers the entire hexagon, including the outside-W exponential decay of square and vertical lozenges. The paper should either state the missing estimates for the non-waterfall slices or explicitly restrict Theorem 1.4 to the waterfall regime, since the current formulation asserts more than the proof verifies.
minor comments (5)
  1. [Remark 2.4, page 14] The phrase 'These these factors are independent of x' contains a duplicated word and should be corrected.
  2. [Figure 7, left plot] The caption says 'the values close to zero are equal to ±L^{-1}', but the vertical axis is labeled L^{-1} H_k; please clarify whether the plotted quantity is the rescaled function or the raw discrete integral, so that the reader can reproduce the positivity/zero regions.
  3. [§5.3, Eq. (5.18)] In the displayed definition of H(u), the term 4F_{1/2(S+t-N),1/2(S+t+N)}(u) appears, but the following sentence refers to 'the last cdf' as if it were the term 2F_{...}(u) from (5.17); the notation should be aligned to avoid confusion about which cdf is order O(1).
  4. [§7.5.2 and Table 5] The parameter is sometimes written as 'κ = 4.3, i' or 'κ = 4.3, i' instead of the intended κ = 4.3i; this appears in the text before Eq. (7.15) and in the caption of Table 5.
  5. [§7.1] The sentence 'In this way one recovers the celebrated incomplete Beta kernel [OR03]' is grammatically unclear because the preceding sentence describes the Fourier-space inverse; please rephrase to state that the inverse Fourier transform yields the incomplete Beta kernel in the bounded-operator regime.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the rigorous Theorem 1.4 is derived from exact q-Racah formulas and standard spectral-projection theorems; the barcode kernel is an explicitly labeled conjecture tested against independent sampling. A non-circular proof gap in Lemma 5.10 is flagged.

full rationale

The derivation chain for Theorem 1.4 is self-contained: the fixed-slice kernel is expressed exactly through q-Racah orthogonal polynomials (Eqs. (2.25), (6.7)) and the limit of the rescaled difference operator is computed coefficient-wise (Lemmas 4.1, 4.3, 4.4), with spectral projection convergence supplied by Reed–Simon Theorem 4.7. No parameter is fitted to the target statement. The exponential concentration half rests on the exact probability ratio (5.1), with q-powers extracted in Lemmas 5.1–5.2 and the packed-configuration minimization in Lemmas 5.3–5.6; these are algebraic identities plus a minimization argument. The barcode-kernel part is openly conjectural: the divergent series (1.7) is regularized by pairing terms, and the resulting kernel is tested against independent perfect-sampling simulations and exact evaluations (Sections 7.4–7.5), so it is not a fitted input renamed as a prediction. The period-two statement is presented as a conjecture, not derived from the input weights. The only load-bearing concern is an unproven assertion inside Lemma 5.10: positivity of (5.23) is justified by 'see Figure 8' rather than by an algebraic estimate; this is a correctness gap on which Theorem 5.11 and the exponential part of Theorem 1.4 depend, but it is not circularity, since the claim does not presuppose the conclusion. Score 0 for circularity.

Assumptions & free parameters 0 free parameters · 4 assumptions · 1 invented entities

No parameters are fitted: q and kappa are fixed inputs. The central theorem relies on standard spectral theory and on the exact algebra of q-Racah polynomials; the barcode conjecture adds an ad hoc regularization of a divergent series.

assumptions (4)
  • domain assumption Imaginary q-Racah case: kappa in iR_{>0}, q in (0,1), and N<T so the waterfall region is nonempty
    The paper restricts to the imaginary case and to N<T throughout (Section 3.2, Definition 3.1); other cases do not exhibit the waterfall.
  • standard math Strong resolvent convergence implies convergence of spectral projections (Reed-Simon Theorem VIII.24-25)
    Invoked as Theorem 4.7 to justify kernel convergence from coefficient convergence.
  • standard math q-Mehler Poisson kernel formula for continuous q^{-1}-Hermite polynomials from Ismail-Masson
    Used in Appendix A.5 to prove completeness of the functions F_n (Proposition A.8).
  • ad hoc to paper The two distinct summations of the divergent density series determine the even and odd densities
    Conjectures 7.2 and 7.7 introduce a specific regularization of a divergent series; this is not proved and is introduced to match the observed period-two behavior.
invented entities (1)
  • Barcode process (one-dimensional determinantal point process on Z) independent evidence
    purpose: Conjectured local limit of q-Racah tilings inside the waterfall region after dimensional collapse
    Its kernel K_barcode(s,t|q,kappa) yields explicit predictions for densities and correlations, which the paper tests against exact pre-limit kernels and perfect sampling simulations without fitting parameters.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Random Lozenge Waterfall: Dimensional Collapse of Gibbs Measures." pith.science (2026). https://pith.science/paper/UL237S5E

@misc{pith2026250722011,
  author       = {Pith},
  title        = {Pith review of: Random Lozenge Waterfall: Dimensional Collapse of Gibbs Measures},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UL237S5E}},
  note         = {Machine review of arXiv:2507.22011}
}
abstract

We investigate the asymptotic behavior of the q-Racah probability measure on lozenge tilings of a hexagon whose side lengths scale linearly with a large parameter $L$, while the parameters $q\in(0,1)$ and $\kappa\in \mathbf{i}\mathbb{R}$ remain fixed. This regime differs fundamentally from the traditional case $q\sim e^{-c/L}\to1$, in which random tilings are locally governed by two-dimensional translation-invariant ergodic Gibbs measures. In the fixed-q regime we uncover a new macroscopic phase, the waterfall (previously only observed experimentally), where the two-dimensional Gibbs structure collapses into a one-dimensional random stepped interface that we call a barcode. We prove a law of large numbers and exponential concentration, showing that the random tilings converge to a deterministic waterfall profile. We further conjecture an explicit correlation kernel of the one-dimensional barcode process arising in the limit. Remarkably, the limit is invariant under shifts by $2\mathbb{Z}$ but not by $\mathbb{Z}$, exhibiting an emergent period-two structure absent from the original weights. Our conjectures are supported by extensive numerical evidence and perfect sampling simulations. The kernel is built from a family of functions orthogonal in both spaces $\ell^{2}(\mathbb{Z})$ and $\ell^{2}(\mathbb{Z}+\frac12)$, that may be of independent interest. Our proofs adapt the spectral projection method of Borodin-Gorin-Rains (arXiv:0905.0679) to the regime with fixed~q. The resulting asymptotic analysis is substantially more involved, and leads to non-self-adjoint operators. We overcome these challenges in the exponential concentration result by a separate argument based on sharp bounds for the ratios of probabilities under the q-Racah orthogonal polynomial ensemble.

Figures

Figures reproduced from arXiv: 2507.22011 by the authors.

Figure 1
Figure 1. Left: the three types of lozenges. Center: an example of a lozenge tiling of a hexagon [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Left: An example of a lozenge tiling of a hexagon with sides [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Zones inside the hexagon in which X(t) (the path configuration at a slice t = const) has the q-Racah OPE distribution with different choices of parameters (see Theorem 2.6). For all T, S, N, the zones (2.15) and (2.18) are present inside the hexagon. The zone (2.16) is present if and only if S ≤ T − S, and otherwise we have the zone (2.17). On a border slice between two zones, the q-Racah parameters can be chosen in… view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: Examples of the waterfall region W inside the hexagon in the coordinate system (t, x), together with an exact sample produced by the shuffling algorithm from [BGR10]. On the left, we have displayed the large-scale limit of the trajectory of the nonintersecting path sta…
Figure 5
Figure 5. Figure 5: Exact samples from the q-Racah measure generated by the shuffling algorithm of [BGR10] in the regime N > T. The parameters are q = 0.85, κ = i, and (N, T, S) = (180, 120, 80) (left) and (N, T, S) = (100, 90, 60) (right). In the left-hand picture no waterfall behavior i…
Figure 6
Figure 6. Figure 6: The plot of Gk(x, y, z | A) as a function of z, where x = 4, y = 10, A = 15, and k = 4.3. Proof of Lemma 5.3. Under the assumptions on x, y, Gk is nonpositive as a function of z. More￾over, it weakly decreases for z ≤ 1 2 (A − k − 1), weakly increases afterwards, and s…
Figure 7
Figure 7. Figure 7: Left: The plot of L −1Hk(⌊Lu⌋, z0 | T, S, N, t) as a function of u. Right: The plot of its continuous analogue H(u) on a larger interval. The parameters are (T, S, N) = (4, 2, 1), t = 2, k = −2.2, and L = 25, and the saturation band is (1.5, 2.5). In the left plot, the…
Figure 8
Figure 8. Figure 8: The plot of the function (5.23) for x ≥ y > 0, and its horizontal cross-section at height zero. We see that the positive part of the function lies above the triangle with vertices (0, 0),( 4 5 , 0), and ( 1 2 , 1 2 ). Since (5.23) is piecewise linear, we conclude that …
Figure 9
Figure 9. Figure 9: Local structure of the one-dimensional random stepped interface in the waterfall region [PITH_FULL_IMAGE:figures/full_fig_p042_9.png]
Figure 10
Figure 10. Figure 10: The even density function ρ barcode even = Kbarcode (M) (0, 0) as a function of the parameters q and κ/i. Here we take M = 10, 1 10 ≤ q ≤ 9 10 , 1 13 ≤ κ/i ≤ 4, and the discretization in both parameters is 1/100. Note that the theta functions in the denominator of (A.…
Figure 11
Figure 11. Figure 11: Cross-sections of the surface ρ barcode even as one of the parameters q or κ/i is fixed. 51 [PITH_FULL_IMAGE:figures/full_fig_p051_11.png]
Figure 12
Figure 12. Figure 12: Two-point correlations in the barcode process as a function of [PITH_FULL_IMAGE:figures/full_fig_p052_12.png]
Figure 13
Figure 13. Figure 13: Graphical output of the Python script. Here T = 50, S = 10, N = 20, q = 7 10 , and κ = 3i. The barcode process realization is 11111111111111101001111001111010011011011111111111. In the rest of this subsection, we compare empirical pattern counts in the barcode process…
Figure 14
Figure 14. Figure 14: Fluctuations h(t) − t 2 of the height function of the barcode process arising in the hexagon with parameters (T, S, N) = (8L, 4L, 4L) for L = 100 and 200, where q = 4 5 , κ = 3i. The offset is K = L/10. On the left, the middle (bold) graph is the moving average, and t…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

9 extracted references · 8 canonical work pages

  1. [1]

    Universality for Lozenge Tiling Local Statistics

    [Agg23] A. Aggarwal, Universality for Lozenge Tiling Local Statistics , Ann. of Math. (2) 198 (2023), no. 3, 881–1012. arXiv:1907.09991 [math.PR]. ↑1, 17 [AG22] A. Aggarwal and V. Gorin, Gaussian unitary ensemble in random lozenge tilings , Prob. Theory Relat. Fields 184 (2022), no. 4, 1139–1166. arXiv:2106.07589 [math.PR]. ↑1 [AH25] A. Aggarwal and J. Hu...

  2. [6]

    Plane partitions with 2-periodic weights

    ↑15 [Mkr14] S. Mkrtchyan, Plane partitions with 2-periodic weights , Letters in Mathematical Physics 104 (2014), no. 9, 1053–1078. arXiv:1309.4825 [math.PR]. ↑3 [Mkr21] S. Mkrtchyan, Turning Point Processes in Plane Partitions with Periodic Weights of Arbitrary Period , Prog. Math. 340 (2021), 497–527. arXiv:1908.01246 [math.PR]. ↑3 [MP17] S. Mkrtchyan an...

  3. [1972]

    Random Surfaces

    ↑27, 45 [She05] S. Sheffield, Random surfaces, Ast´ erisque304 (2005). arXiv:math/0304049 [math.PR]. ↑1, 17 [Tao12] T. Tao, Topics in random matrix theory , Graduate Studies in Mathematics, vol. 132, AMS,

  4. [1996]

    Orthogonal polynomial ensembles in probability theory

    Updated version available at https://fa.ewi.tudelft.nl/~koekoek/documents/as98.pdf. ↑2, 7, 12, 26, 56, 57 [K¨ on05] W. K¨ onig,Orthogonal polynomial ensembles in probability theory , Probab. Surv. 2 (2005), 385–447. arXiv:math/0403090 [math.PR]. ↑12, 14, 15 64 [Meh04] M.L. Mehta, Random matrices, Academic press,

  5. [2004]

    Non-intersecting paths and Hahn orthogonal polynomial ensemble

    ↑12, 43 [Gor08] V. Gorin, Nonintersecting paths and the Hahn orthogonal polynomial ensemble , Funct. Anal. Appl. 42 (2008), no. 3, 180–197. arXiv:0708.2349 [math.PR]. ↑1 [Gor21] V. Gorin, Lectures on random lozenge tilings, Cambridge Studies in Advanced Mathematics. Cambridge University Press (2021). ↑1, 17 [GH24] V. Gorin and J. Huang, Dynamical loop equ...

  6. [2007]

    Uniform Asymptotics for Polynomials Orthogonal With Respect to a General Class of Discrete Weights and Universality Results for Associated Ensembles

    arXiv:math/0310278 [math.CA]. ↑1 [BB23] T. Berggren and A. Borodin, Geometry of the doubly periodic Aztec dimer model, arXiv preprint (2023). arXiv:2306.07482 [math.PR]. ↑3 [BNR24] T. Berggren, M. Nicoletti, and M. Russkikh, Perfect t-embeddings and Lozenge Tilings, arXiv preprint (2024). arXiv:2408.05441 [math.PR]. ↑1 [Bor11] A. Borodin, Determinantal po...

  7. [2011]

    ↑14, 38 [BG13] A

    arXiv:0911.1153 [math.PR]. ↑14, 38 [BG13] A. Borodin and V. Gorin, Markov processes of infinitely many nonintersecting random walks , Probab. Theory Relat. Fields 155 (2013), no. 3-4, 935–997. arXiv:1106.1299 [math.PR]. ↑17 [BGR10] A. Borodin, V. Gorin, and E. Rains, q-Distributions on boxed plane partitions, Selecta Math. 16 (2010), no. 4, 731–789. arXiv...

  8. [2012]

    ↑3, 16 [WMTB76] T. T. Wu, B. M. McCoy, C. A. Tracy, and E. Barouch, Spin-spin correlation functions for the two- dimensional Ising model: Exact theory in the scaling region , Phys. Rev. B 13 (1976), no. 1, 316–374. ↑4 a. knizel, barnard college, columbia university, new york, ny, usa e-mail: aknizel@barnard.edu l. petrov, university of virginia, charlotte...

Show all 9 references
  1. [2025]

    Petrov and M

    ↑52 [PT23] L. Petrov and M. Tikhonov, Asymptotics of noncolliding q-exchangeable random walks , J. Phys. A: Math. Theor. 56 (2023), no. 36, 365203. arXiv:2303.02380 [math.PR]. ↑17 [RS72] M. Reed and B. Simon, Methods of Modern Mathematical Physics, Vol. I. Functional Analysis ...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.