pith. sign in

arxiv: 2606.19207 · v1 · pith:P64WZI5Enew · submitted 2026-06-17 · 🧮 math.PR

Extrema of microscopically slowed-down Gaussian fields

Zuodi Xie This is my paper

Pith reviewed 2026-06-26 19:49 UTC · model grok-4.3

classification 🧮 math.PR
keywords Gaussian fieldsextremamaximumtightnessphase transitionbranching Brownian motioncovariance slowdownmicroscopic interpolation
0
0 comments X

The pith

Gaussian fields with microscopically slowed covariance have maxima centered at order T^{1-α} with second-order phase transition at α=1/3.

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

The paper introduces a family of Gaussian fields whose covariance structure slows down microscopically in an inhomogeneous manner, with the slowdown tuned by a parameter α that runs from 0 to 1/2. It identifies the precise centering term for the maximum attained by time T and proves that the recentered maximum remains tight. The leading term in the centering grows as T to the power 1-α, with the exponent linear in α. The second-order correction changes form exactly when α equals 1/3. A reader would care because the construction unifies the study of extremes for models whose correlation profiles range continuously from ordinary logarithmic to double-logarithmic growth.

Core claim

For the family of Gaussian fields with inhomogeneous microscopic slowdown in covariance, the maximum M_T at time T admits a centering sequence such that M_T minus this sequence is tight; the centering has leading order T^{1-α} whose exponent varies linearly in α, while the second-order correction term undergoes a phase transition at α=1/3. The same centering and tightness hold for both the one-dimensional branching Brownian motions in a cooling environment and the higher-dimensional spatial versions.

What carries the argument

The family of Gaussian fields whose covariance exhibits an inhomogeneous microscopic slowdown parameterized by α, interpolating between log and log-log profiles.

If this is right

  • The leading growth of the maximum is of order T^{1-α}.
  • Tightness of the recentered maximum holds uniformly for α in the interval from 0 to 1/2.
  • The second-order term in the centering changes character at the critical value α=1/3.
  • The results apply equally to the one-dimensional cooling-environment branching Brownian motion and to the higher-dimensional spatial fields.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The phase transition at α=1/3 may mark a shift in how the microscopic slowdown interacts with the branching or correlation structure at finer scales.
  • The same scaling laws and transition could be checked numerically by direct simulation of the covariance for values of α straddling 1/3.
  • Analogous slowdown constructions in non-Gaussian fields might produce comparable centering and tightness statements.
  • The one-dimensional case offers a natural setting in which to compare the cooling-environment model against existing results on time-inhomogeneous branching processes.

Load-bearing premise

The covariance structure must exhibit an inhomogeneous microscopic slowdown that exactly interpolates between the log profile at α=0 and the log-log profile at α=1/2 while preserving the Gaussian property and allowing the stated centering and tightness to hold.

What would settle it

For a fixed α between 0 and 1/2, compute numerical realizations of the field up to large T, subtract the proposed centering of order T^{1-α} plus the predicted second-order term, and check whether the resulting sequence stays bounded in probability rather than drifting to plus or minus infinity.

Figures

Figures reproduced from arXiv: 2606.19207 by Zuodi Xie.

Figure 1
Figure 1. Figure 1: Branching explorers in a cooling medium tY α,v s : s P r0, tsu denote the trajectory of particle v before time t. It is well known that the number of BBM particles satisfies ErNts “ e t , PpNt “ nq “ e ´t p1 ´ e ´t q n´1 , n ě 1, P ` Nt ď e ct˘ ď e ´p1´cqt , @ c ă 1. (1.12) See, for example, the proof of Lemma 2.2 in [CHM23]. The following theorem shows that the rightmost positions of our microscopically s… view at source ↗
Figure 2
Figure 2. Figure 2: By the branching property and (2.25), P ˜ max v 1PN T0,v0 T ` Y α,v1 T ´ Y α,v0 T0 ˘ ě ˆ T T0 kαpsq ds ´ M ÿ´1 m“0 p2 mHq ´α p2 m´1Hq β ¸ ě M ź´1 m“0 ´ 1 ´ e ´cp2mHq β ¯ ě 1 ´ M ÿ´1 m“0 e ´cp2mHq β ě 1 ´ e ´c 1Hβ . (2.26) Since β ă α, M ÿ´1 m“0 p2 mHq ´α p2 m´1Hq β ď CHβ´α ÿ8 m“0 2 ´mpα´βq ď C. Therefore, for every v0 P NT0 , P ˜ max v 1PN T0,v0 T ` Y α,v1 T ´ Y α,v0 T0 ˘ ě ˆ T T0 kαpsq ds ´ C ¸ ě 1 ´ e ´c… view at source ↗
Figure 2
Figure 2. Figure 2: The leading particles in BBM Thus P ˆ max vPNT Y α,v T ď AαpTq ´ λ ˙ ď P ˆ N ˆ 0, ˆ T 0 σ 2 αpsq ds ˙ ď ´ λ 2 ˙ ď e ´cλ . It remains to consider λ ď 10T. Set H0 :“ 1 ` λ 50 . Choose M P Z` such that H :“ 2 ´Mp1 ` 2Tq P rH0, 2H0q. Then TM “ T for the mesh (2.20), and T0 “ H ´ 1 2 satisfies c1λ ď T0 ď c2λ for deterministic constants c1, c2 ą 0 and all large λ. For v P NT0 , define Dv :“ max wPN T0,v T ` Y α,… view at source ↗
Figure 3
Figure 3. Figure 3: The affine map L z ε,1 . and, for every z P Vε and every x, y P lz ε , }x ´ y}1 ε ď E “ pψ x ε ´ ψ y ε q 2 ‰ ď 2 d }x ´ y}1 ε . (3.12) Finally, we recall Lemma 2.2 of [Aco14], which follows from Fernique’s majorizing-measure criterion: there exist constants c, C ą 0, independent of λ and ε, such that sup vPVε P ˜ sup xPlv ε ψ x ε ě λ ¸ ď Ce´cλ2 , λ ě 0, ε ą 0. (3.13) 3.2. Upper Tail. In this subsection, we… view at source ↗
read the original abstract

We introduce a family of Gaussian fields whose covariance structure exhibits an inhomogeneous, microscopic slowdown and it interpolates between a $\log$ profile (for a certain interpolation parameter $\alpha=0$) and a $\log\log$ profile (when the interpolation parameter is $\alpha=1/2$). We consider both one dimensional such objects (which we call {\it Branching Brownian Motions in a cooling environment}) as well as higher dimensional, spatial fields. We identify the correct centering of the maximum at time $T$ and prove tightness of the recentered maximum. While the exponent in the first-order growth varies linearly with $\alpha$, giving a leading order of $T^{1-\alpha}$, the second-order correction exhibits a phase transition at $\alpha=1/3$.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript introduces a parametrized family of Gaussian fields (including a one-dimensional 'Branching Brownian Motion in a cooling environment' and higher-dimensional spatial versions) whose covariance exhibits an inhomogeneous microscopic slowdown. The covariance is constructed to interpolate continuously between the standard log-correlated case at α=0 and the log-log correlated case at α=1/2. The central results are an explicit identification of the centering sequence for the maximum at time T (leading term of order T^{1-α}) together with a proof of tightness for the recentered maximum; the second-order correction term is shown to undergo a phase transition at α=1/3.

Significance. If the technical steps hold, the work supplies a unified, continuous interpolation between two well-studied but qualitatively different regimes of extreme-value behavior for Gaussian fields. The explicit centering formula and the identification of the α=1/3 transition would constitute a concrete advance in the theory of extrema for non-stationary or modified Gaussian processes, with potential relevance to branching random walks and related models.

major comments (2)
  1. [Abstract / §1 (presumed)] The abstract states that the covariance 'exhibits an inhomogeneous, microscopic slowdown' that 'exactly interpolates' while 'preserving the Gaussian property,' yet no explicit formula, kernel definition, or verification of positive-definiteness for α ∈ [0,1/2] is supplied in the visible text. Because the centering T^{1-α} and the tightness proof rest on this construction, the absence of the explicit covariance prevents verification that the claimed interpolation is admissible.
  2. [Abstract] The phase-transition claim at α=1/3 for the second-order correction is asserted without any displayed equation, heuristic derivation, or reference to the section where the correction term is computed. This makes it impossible to assess whether the transition arises from the inhomogeneous slowdown or from an auxiliary approximation whose validity may break at that value.
minor comments (1)
  1. [Abstract] The one-dimensional object is called 'Branching Brownian Motions in a cooling environment'; a brief sentence clarifying how the cooling is realized via the covariance slowdown would help readers connect the model to existing literature on branching Brownian motion.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their careful reading and for identifying points where the presentation could be improved for clarity. We address the two major comments point by point below.

read point-by-point responses

  1. Referee: The abstract states that the covariance 'exhibits an inhomogeneous, microscopic slowdown' that 'exactly interpolates' while 'preserving the Gaussian property,' yet no explicit formula, kernel definition, or verification of positive-definiteness for α ∈ [0,1/2] is supplied in the visible text. Because the centering T^{1-α} and the tightness proof rest on this construction, the absence of the explicit covariance prevents verification that the claimed interpolation is admissible.

    Authors: The explicit covariance kernel is defined in Section 2 of the manuscript (Definition 2.1), where the inhomogeneous microscopic slowdown is constructed to interpolate continuously between the log-correlated case at α=0 and the log-log case at α=1/2. Positive-definiteness for α ∈ [0,1/2] is established in Proposition 2.2 by verifying that the kernel remains a valid covariance function under the continuous interpolation, building on the known positive-definiteness of the endpoint cases. We agree that the abstract lacks a reference to this construction and will revise it to include a brief description of the kernel together with a pointer to Section 2. revision: yes

  2. Referee: The phase-transition claim at α=1/3 for the second-order correction is asserted without any displayed equation, heuristic derivation, or reference to the section where the correction term is computed. This makes it impossible to assess whether the transition arises from the inhomogeneous slowdown or from an auxiliary approximation whose validity may break at that value.

    Authors: The second-order correction term and the phase transition at α=1/3 are derived in Section 4 (Theorem 4.3), where the explicit form of the correction is obtained from a variational problem that accounts for the slowdown parameter. The transition arises directly from the inhomogeneous covariance: for α < 1/3 the leading correction is governed by the microscopic slowdown, while for α > 1/3 it reverts to the adjusted log-log behavior. A heuristic derivation appears in the introduction (page 4). We will add a reference to Section 4 and a one-sentence indication of the origin of the transition in the abstract. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation self-contained against external benchmarks

full rationale

The provided abstract and context contain no equations, fitted parameters, self-citations, or explicit derivation steps. Claims of identifying centering (leading term T^{1-α}) and proving tightness rest on an unspecified covariance construction that is stated to interpolate between log and log-log profiles while preserving Gaussianity and positive-definiteness. No load-bearing step reduces by construction to its own inputs, as no internal definitions or predictions are visible that equate to fitted quantities. This is the normal case of an honest non-finding when the manuscript text supplies no reducible steps.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Review based solely on abstract; no explicit free parameters, axioms, or invented entities are stated. The result relies on standard properties of Gaussian fields and branching processes.

pith-pipeline@v0.9.1-grok · 5644 in / 1142 out tokens · 23946 ms · 2026-06-26T19:49:01.243442+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

52 extracted references · 14 canonical work pages

  1. [1]

    arXiv preprint arXiv:2507.16056 , year=

    Conditional GMC within the stochastic heat flow , author=. arXiv preprint arXiv:2507.16056 , year=

    arXiv

  2. [2]

    arXiv preprint arXiv:2603.03246 , year=

    Log Log Fluctuations of the Stochastic Heat Flow , author=. arXiv preprint arXiv:2603.03246 , year=

    arXiv

  3. [3]

    Gaussian Free Field and Liouville Quantum Gravity , series =

    Berestycki, Nathana. Gaussian Free Field and Liouville Quantum Gravity , series =. 2025 , doi =

    2025

  4. [4]

    2007 , publisher=

    Random Fields and Geometry , author=. 2007 , publisher=

    2007

  5. [5]

    The Annals of Probability , volume=

    The two-dimensional KPZ equation in the entire subcritical regime , author=. The Annals of Probability , volume=. 2020 , publisher=

    2020

  6. [6]

    Stochastic Processes and their Applications , volume=

    Directed polymers in a random environment: a review of the phase transitions , author=. Stochastic Processes and their Applications , volume=. 2024 , publisher=

    2024

  7. [7]

    arXiv preprint arXiv:2511.08479 , year=

    From disordered systems to the Critical 2D Stochastic Heat Flow , author=. arXiv preprint arXiv:2511.08479 , year=

    arXiv

  8. [8]

    2017 , publisher=

    Directed polymers in random environments , author=. 2017 , publisher=

    2017

  9. [9]

    2007 , publisher=

    Random fields and geometry , author=. 2007 , publisher=

    2007

  10. [10]

    Probability and Statistical Physics in St

    Ofer Zeitouni , title =. Probability and Statistical Physics in St. Petersburg , series =. 2016 , doi =

    2016

  11. [11]

    The Annals of Applied Probability , number =

    Janne Junnila and Eero Saksman and Christian Webb , title =. The Annals of Applied Probability , number =. 2020 , doi =

    2020

  12. [12]

    Probability Surveys , volume=

    Gaussian multiplicative chaos and applications: A review , author=. Probability Surveys , volume=

  13. [13]

    Hu, Xiaoyu and Miller, Jason and Peres, Yuval , TITLE =. Ann. Probab. , FJOURNAL =. 2010 , NUMBER =. doi:10.1214/09-AOP498 , URL =

    work page doi:10.1214/09-aop498 2010

  14. [14]

    Annales de l'Institut Henri Poincar

    On intermediate level sets of two-dimensional discrete Gaussian free field , author=. Annales de l'Institut Henri Poincar. 2016 , url=

    2016

  15. [15]

    arXiv preprint arXiv:2503.20171 , year=

    Martingale measure associated with the critical 2 d stochastic heat flow , author=. arXiv preprint arXiv:2503.20171 , year=

    arXiv

  16. [16]

    arXiv preprint arXiv:2507.22029 , year=

    Sharp moment and upper tail asymptotics for the critical 2 d Stochastic Heat Flow , author=. arXiv preprint arXiv:2507.22029 , year=

    arXiv

  17. [17]

    arXiv preprint arXiv:2412.10311 , year=

    The critical 2d stochastic heat flow and related models , author=. arXiv preprint arXiv:2412.10311 , year=

    Pith/arXiv arXiv

  18. [18]

    arXiv preprint arXiv:2504.06128 , year=

    Singularity and regularity of the critical 2d stochastic heat flow , author=. arXiv preprint arXiv:2504.06128 , year=

    arXiv

  19. [19]

    arXiv preprint arXiv:2410.14601 , year=

    On the moments of the mass of shrinking balls under the Critical 2 d Stochastic Heat Flow , author=. arXiv preprint arXiv:2410.14601 , year=

    arXiv

  20. [20]

    Caravenna, Francesco and Sun, Rongfeng and Zygouras, Nikos , TITLE =. Invent. Math. , FJOURNAL =. 2023 , NUMBER =. doi:10.1007/s00222-023-01184-7 , URL =

    work page doi:10.1007/s00222-023-01184-7 2023

  21. [21]

    Caravenna, Francesco and Sun, Rongfeng and Zygouras, Nikos , TITLE =. Ann. Probab. , FJOURNAL =. 2023 , NUMBER =. doi:10.1214/23-aop1648 , URL =

    work page doi:10.1214/23-aop1648 2023

  22. [22]

    2024 , eprint=

    Stochastic heat flow by moments , author=. 2024 , eprint=

    2024

  23. [23]

    Communications on Pure and Applied Mathematics , volume=

    Maximal displacement of branching Brownian motion , author=. Communications on Pure and Applied Mathematics , volume=. 1978 , publisher=

    1978

  24. [24]

    Probability Theory and related fields , volume=

    The extremal process of branching Brownian motion , author=. Probability Theory and related fields , volume=. 2013 , publisher=

    2013

  25. [25]

    PIMS-CRM Summer School in Probability , pages=

    Extrema of the two-dimensional discrete Gaussian free field , author=. PIMS-CRM Summer School in Probability , pages=. 2017 , publisher=

    2017

  26. [26]

    Advances in Disordered Systems, Random Processes and Some Applications , editor =

    Louis-Pierre Arguin , title =. Advances in Disordered Systems, Random Processes and Some Applications , editor =. 2017 , publisher =

    2017

  27. [27]

    Bramson, Maury , TITLE =. Mem. Amer. Math. Soc. , FJOURNAL =. 1983 , NUMBER =. doi:10.1090/memo/0285 , URL =

    work page doi:10.1090/memo/0285 1983

  28. [28]

    Probability Theory and Related Fields , volume=

    Branching Brownian motion seen from its tip , author=. Probability Theory and Related Fields , volume=. 2013 , publisher=

    2013

  29. [29]

    ALEA: Latin American Journal of Probability and Mathematical Statistics , volume=

    Branching Brownian motion conditioned on small maximum , author=. ALEA: Latin American Journal of Probability and Mathematical Statistics , volume=

  30. [30]

    Electron

    Tightness of the recentered maximum of log-correlated Gaussian fields , author=. Electron. J. Probab , volume=

  31. [31]

    Bramson, Maury and Zeitouni, Ofer , TITLE =. Comm. Pure Appl. Math. , FJOURNAL =. 2012 , NUMBER =. doi:10.1002/cpa.20390 , URL =

    work page doi:10.1002/cpa.20390 2012

  32. [32]

    Bramson, Maury and Ding, Jian and Zeitouni, Ofer , TITLE =. Comm. Pure Appl. Math. , FJOURNAL =. 2016 , NUMBER =. doi:10.1002/cpa.21621 , URL =

    work page doi:10.1002/cpa.21621 2016

  33. [33]

    Annales de l'IHP Probabilit

    Maximum of a log-correlated Gaussian field , author=. Annales de l'IHP Probabilit

  34. [34]

    Fang, Ming and Zeitouni, Ofer , TITLE =. J. Stat. Phys. , FJOURNAL =. 2012 , NUMBER =. doi:10.1007/s10955-012-0581-z , URL =

    work page doi:10.1007/s10955-012-0581-z 2012

  35. [35]

    Maillard, Pascal and Zeitouni, Ofer , TITLE =. Ann. Inst. Henri Poincar\'e. 2016 , NUMBER =. doi:10.1214/15-AIHP675 , URL =

    work page doi:10.1214/15-aihp675 2016

  36. [36]

    Electron

    Bovier, Anton and Hartung, Lisa , TITLE =. Electron. J. Probab. , FJOURNAL =. 2014 , PAGES =. doi:10.1214/EJP.v19-2982 , URL =

    work page doi:10.1214/ejp.v19-2982 2014

  37. [37]

    ALEA Lat

    Bovier, Anton and Hartung, Lisa , TITLE =. ALEA Lat. Am. J. Probab. Math. Stat. , FJOURNAL =. 2015 , NUMBER =

    2015

  38. [38]

    Electron

    Fang, Ming and Zeitouni, Ofer , TITLE =. Electron. J. Probab. , FJOURNAL =. 2012 , PAGES =. doi:10.1214/EJP.v17-2253 , URL =

    work page doi:10.1214/ejp.v17-2253 2012

  39. [39]

    Stochastic Process

    Mallein, Bastien , TITLE =. Stochastic Process. Appl. , FJOURNAL =. 2015 , NUMBER =. doi:10.1016/j.spa.2015.05.011 , URL =

    work page doi:10.1016/j.spa.2015.05.011 2015

  40. [40]

    Ouimet, Fr\'ed\'eric , TITLE =. Braz. J. Probab. Stat. , FJOURNAL =. 2018 , NUMBER =. doi:10.1214/17-BJPS358 , URL =

    work page doi:10.1214/17-bjps358 2018

  41. [41]

    arXiv preprint arXiv:2503.05994 , year=

    The extremal process of two-speed branching random walk , author=. arXiv preprint arXiv:2503.05994 , year=

    arXiv

  42. [42]

    ALEA Lat

    Arguin, Louis-Pierre and Ouimet, Fr\'ed\'eric , TITLE =. ALEA Lat. Am. J. Probab. Math. Stat. , FJOURNAL =. 2016 , NUMBER =. doi:10.30757/alea.v13-31 , URL =

    work page doi:10.30757/alea.v13-31 2016

  43. [43]

    ALEA Lat

    Ouimet, Fr\'ed\'eric , TITLE =. ALEA Lat. Am. J. Probab. Math. Stat. , FJOURNAL =. 2017 , NUMBER =

    2017

  44. [44]

    ALEA , volume=

    Extremes of the 2d scale-inhomogeneous discrete Gaussian free field: Sub-leading order and exponential tails , author=. ALEA , volume=

  45. [45]

    arXiv preprint arXiv:1912.13184 , year=

    Extremes of the 2d scale-inhomogeneous discrete Gaussian free field: Convergence of the maximum in the regime of weak correlations , author=. arXiv preprint arXiv:1912.13184 , year=

    arXiv 1912

  46. [46]

    arXiv preprint arXiv:2002.00925 , year=

    Extremes of the 2d scale-inhomogeneous discrete Gaussian free field: Extremal process in the weakly correlated regime , author=. arXiv preprint arXiv:2002.00925 , year=

    arXiv 2002

  47. [47]

    arXiv preprint arXiv:2503.17236 , year=

    The maximum of the two dimensional Gaussian directed polymer in the subcritical regime , author=. arXiv preprint arXiv:2503.17236 , year=

    arXiv

  48. [48]

    The Annals of Probability , year=

    Maximum and coupling of the sine-Gordon field , author=. The Annals of Probability , year=

  49. [49]

    Communications in Mathematical Physics , volume=

    Multiscale Coupling and the Maximum of P ( ) 2 Models on the Torus , author=. Communications in Mathematical Physics , volume=. 2023 , publisher=

    2023

  50. [50]

    arXiv preprint arXiv:2508.15689 , year=

    The Liouville model in the L1 phase: coupling and extreme values , author=. arXiv preprint arXiv:2508.15689 , year=

    arXiv

  51. [51]

    A probabilistic proof of

    K\"uhn, Franziska and Schilling, Ren\'e. A probabilistic proof of. J. Math. Anal. Appl. , FJOURNAL =. 2019 , NUMBER =. doi:10.1016/j.jmaa.2018.11.046 , URL =

    work page doi:10.1016/j.jmaa.2018.11.046 2019

  52. [52]

    2012 , publisher=

    Bernstein functions: theory and applications , author=. 2012 , publisher=

    2012