Pith. sign in

REVIEW 2 major objections 7 minor 22 references

Morphology of frozen labyrinths from irreversible threshold dynamics

T0 review · 2 major / 7 minor · reviewed 2026-08-08 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Adding a one-flip cap to majority threshold dynamics freezes balanced, non-consensus labyrinth patterns where the reversible rule would coarsen to consensus or stripes.

desk verdict A clean, honest minimal model showing irreversible one-flip threshold dynamics freezes balanced labyrinths; the 1/σ window law is conditional on an unproved but clearly flagged σ-independence of cascade onset. read the letter →

arxiv 2608.05496 v1 pith:YQTXLRTW submitted 2026-08-06 cond-mat.stat-mech

Daniel Richard Levy (Princeton University) This is my paper · ORCID
classification cond-mat.stat-mech
keywords irreversiblethresholddynamicsone-fliplabyrinthpatternsarrestedcoarseningfrustratedspinsGaussiankernellatticemodelstandardizedinitialbiascompositionwindow
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

The paper studies what happens when the usual threshold dynamics of Merriman–Bence–Osher type is made irreversible: on $\mathbb{Z}^2$, each spin may flip out of its local Gaussian-weighted minority at most once. From Bernoulli initial conditions near equal balance, this one-flip rule freezes balanced, non-consensus labyrinth patterns, whereas the same rule without the one-flip cap coarsens to consensus or pinned stripes. The pattern survives only inside a composition window around $p_0=1/2$ whose half-width scales as $1/\sigma$, because the initial bias enters dynamics only through the standardized quantity $z_0\simeq 2\sqrt{\pi}(2p_0-1)\sigma$, with onset at a $\sigma$-independent value $z_0^*\approx0.31$\textendash$0.36$. If correct, the paper establishes a generic mechanism by which permanent, two-sided commitment arrests coarsening in local alignment systems, leaving a measurable frustrated-spin signature that reversible relaxation cannot produce.

What carries the argument

The load-bearing objects are the standardized initial bias and the frustrated spin. The bias, $z_0=\mathbb{E}[m]/\sqrt{\operatorname{Var}[m]}\simeq 2\sqrt{\pi}(2p_0-1)\sigma$, collapses the entire initial excursion-set geometry of the random field onto one scale-free variable, and the paper's window law follows from assuming cascade onset occurs at a fixed $z_0^*$. The frustrated spin, a site that has used its single flip and sits with $s_i m_i<0$, is the signature of irreversibility: no reversible relaxation can stop with a spin strictly against its local field, and in the frozen state these sites decorate the domain walls and lock them in place.

What would settle it

Run the one-flip dynamics at $\sigma=32$ and $64$, tuning $p_0$ to keep $z_0$ matched, and measure the cascade onset $z_0^*$ from the same magnetization or giant-cluster crossings used here; if the crossing moves by more than the bootstrap scatter observed at $\sigma\le16$, the claimed $\sigma$-independence and the derived $1/\sigma$ law are wrong.

Watch

Extended reading notes

Core claim

Adding a one-flip cap to asynchronous Gaussian threshold dynamics changes the absorbing-state selection. From random initial conditions close to balance, the system freezes into smooth bicontinuous labyrinths with low residual magnetization, instead of the near-consensus or flat pinned states reached by reversible descent. The control variable is the standardized initial bias $z_0 = \mathbb{E}[m]/\sqrt{\operatorname{Var}[m]} \simeq 2\sqrt{\pi}(2p_0-1)\sigma$; the cascade boundary sits at a fixed $z_0^*$ independent of $\sigma$ over a fourfold range, so the patterned window narrows as $1/\sigma$. The frozen morphology is arrested coarsening: no wavelength is selected, both phases span at balance, and the feature width grows sublinearly with $\sigma$, falling below the interaction range at large scales. The paper traces the mechanism by intervention: the coarse domain layout is deterministic, walls are placed by deterministic front advance with only a secondary first-passage race, and frustrated spins\textemdash spent sites whose spin opposes their local field\textemdash lock the walls without placing them.

Load-bearing premise

The argument rests on the premise, supported only numerically for interaction ranges $\sigma=4$ through $16$, that cascade onset occurs at a fixed value $z_0^*$ of the standardized initial bias independent of $\sigma$; if $z_0^*$ drifts at larger $\sigma$ or system size, the $1/\sigma$ window law and the scale-free control claim would fail.

Editorial extensions

If this is right

  • The half-width of the patterned composition window obeys $\Delta p_0 = z_0^*/(4\sqrt{\pi}\sigma)\propto 1/\sigma$, with $z_0^*\approx0.31$\textendash$0.36$ depending on the crossing convention.
  • At balance the absorbing labyrinth is bicontinuous with both phases spanning in the great majority of large boxes, and its structure factor has no finite-wavenumber peak, ruling out wavelength selection.
  • The frozen feature width $\ell_w$ grows sublinearly, $\ell_w\approx10.6\,\sigma^{0.30}$, and $\ell_w/\sigma$ passes below unity near $\sigma=28$, so the frozen length is not set by the interaction range.
  • Holding the initial condition fixed and varying update order shows the coarse domains are deterministic; the update-order race fine-tunes only the last fraction of $\sigma$ of the wall positions, and frustration ablation leaves the coarse layout intact.
  • Frustrated-site density distinguishes irreversible from reversible absorption: reversible dynamics must stop with every spin aligned with its local field, while the one-flip absorbing state carries a finite density of frustrated spins, 97\textendash98% of them on the walls.

Reading between the lines

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

  • An extension the paper leaves implicit: if $z_0^*$ is indeed $\sigma$-independent for the Gaussian kernel, the same $1/\sigma$ window should occur for any positive, integrable kernel; only the numerical prefactor in $z_0$ would change.
  • The measured $\ell_w\propto\sigma^{0.3}$ may be a crossover rather than an asymptotic exponent; testing at $\sigma\ge32$ would show whether $\ell_w/\sigma$ keeps declining or saturates, which distinguishes true arrested coarsening from a slowly growing selected scale.
  • The deterministic-layout result suggests pattern reproducibility in committed-agent models is governed by initial conditions alone; a testable prediction is that duplicates of the same initial condition under independent clocks agree site-wise almost everywhere except near walls.
  • The wall-placement race invites a quantitative comparison with first-passage percolation: at fixed large $\sigma$, the scatter of individual-order walls about the consensus wall should follow the same scaling as competing-growth interfaces.
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

2 major / 7 minor

Summary. This paper studies an asynchronous majority-threshold dynamics on Z^2 with a Gaussian kernel and a one-flip ("irreversible") constraint, and contrasts it with the reversible MBO-type relaxation of the same rule. The main results are: (i) exact lattice pinning constants for flat interfaces and a closed-form critical stripe width; (ii) a curvature drive for droplets with coefficient sqrt(2π)σ^3 and a universal mobility function g(z) describing the eligible boundary fraction; (iii) the identification of a composition window around p0=1/2 in which random initial conditions freeze into balanced bicontinuous labyrinths, with a window half-width Δp0 ∝ 1/σ controlled by the standardized initial bias z0 ≈ 2√π(2p0−1)σ; (iv) a characterization of the frozen morphology as arrested coarsening with sublinear feature width; and (v) an intervention-based decomposition showing that the coarse domains and wall positions are set deterministically, with a secondary first-passage race and frustrated interfacial sites locking, but not placing, the walls.

Significance. The paper's strengths are its exact single-interface calculations, the clean t=0 standardization of the initial field, the careful mobility collapse, and the unusually explicit segregation of proved statements from numerically supported assumptions. The irreversibility mechanism and the frustration signature are conceptually clear and transferable. If the 1/σ window law holds beyond the tested range, the paper provides a generic route to absorbing-state labyrinth patterns in two-sided committed dynamics. The main limitation is the unproved σ-independence of the cascade-onset threshold z*0; the paper acknowledges this, but because Eq. (17) and the abstract's window-law statement depend on it, the quantitative central claim is conditional on numerical extrapolation.

major comments (2)
  1. [Sec. V B/C, Eq. (17)] The central law Δp0 = z*0/(4√π σ) rests on the assumption, stated in Sec. V B, that cascade onset occurs at a fixed value z*0 independent of σ. The evidence in Sec. V C (slopes −0.004±0.003 over σ∈[4,16], 15 realizations per point) is genuine, and the matched-z0 collapse is visually good, but it does not bound a slow drift at larger σ or larger L/σ. The independent direct fit α=1.05±0.09 over σ∈[3,8] is consistent with 1/σ, but an effective exponent near 1 would also result from a slowly varying z*0(σ), for example z*0 ∝ σ^ε with small ε. Because the abstract and Sec. VIII present the 1/σ law as a quantitative result, I ask that the manuscript either extend the z*0 measurements to σ=32 and 64 with a few L/σ values, or explicitly rephrase the law as an empirical statement valid over the tested fourfold range, with the asymptotic form left as a stated conjecture.
  2. [Sec. V C] The σ-independence regression is performed at a single box ratio L=26σ and with 15 realizations per point. Since the onset is a smooth crossover, the measured z*0 could in principle depend on L/σ; the paper does not report a finite-size check for z*0, although the morphology section does such a check for ℓ_w. A brief statement of the dependence of z*0 on L/σ at a representative σ would close an otherwise open route to explaining the apparent σ-independence by a coincidental finite-size effect.
minor comments (7)
  1. [Sec. II] There is a typo in "irrerversible model" in the paragraph introducing the reversible variant.
  2. [Sec. IV D] The word "wihout" should read "without" in "leaves the boundary slightly non-circular wihout introducing fine-scale structure."
  3. [Sec. VII C] The phrase "far from what the the initial field alone implies" contains a duplicated "the."
  4. [Reference [12]] The title "Sharp metastability threhold" contains a misspelling; it should be "threshold."
  5. [Sec. VI] The phrase "described by σ ≈0.3" is an incomplete shorthand for the feature-width law; it should read "described by an exponent ℓ_w ∝ σ^0.30" or similar, both in the text and in the Fig. 4 caption.
  6. [Sec. V A] The heading "An exact control variable" is potentially misleading because Eq. (16) is the large-σ, near-balance limit of the exact ratio E[m]/√Var[m]; consider renaming it "A control variable" or "An asymptotically exact control variable."
  7. [Appendix A] The paper gives detailed numerical protocols but no data- or code-availability statement; a sentence on availability of simulation code would aid reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central 1/σ window law is an explicitly conditional consequence of an exactly derived z0 control plus a clearly stated, independently measured numerical assumption, not a hidden fit or self-citation.

full rationale

The paper's derivation chain is self-contained and does not reduce its predictions to its inputs by construction. The exact control variable z0 = E[m]/sqrt(Var[m]) in Eq. (16) is computed directly from the Gaussian kernel row sums and i.i.d. initial spins, with no reference to the window width or the onset location. The window law Eq. (17), Δp0 = z*0/(4√π σ), is explicitly conditional: the paper states, 'The assumption is that cascade onset occurs at a fixed value z0 = z*0 independent of σ' (Sec. V B) and labels the fixed-z*0 boundary as numerically supported but not proved. z*0 is not a fitted parameter chosen to force the window law; it is measured independently from two order-parameter crossing conventions, and the paper reports the slope of z*0 versus σ as −0.004±0.003 for both definitions. Moreover, the 1/σ scaling is confirmed by an independent raw-p0 sweep giving α=1.05±0.09, which does not presuppose the standardized-bias scaling. The universal mobility collapse f_elig ≈ g(z) is a data collapse over measured ring means and spreads, explicitly presented as empirical ('None of δ, the distribution shape, or the correlation is derived here'), not as a prediction from fitted parameters. The frustration signature is a direct logical consequence of the reversible dynamics' absorbing-state condition (s_i m_i ≥ 0 for all i), so the claim that reversible relaxation leaves no frustrated spins is a theorem about the model, not a circular definition. There are no load-bearing self-citations, no imported uniqueness theorems, and no ansatz smuggled in by citation. The only genuinely fragile element is the unproved σ-independence of z*0 at larger σ and L/σ; the paper acknowledges this openly, and that is a scaling/correctness risk rather than circular reasoning.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

No new physical entities are introduced. Frustrated sites are a defined observable, not a postulated entity. The model's free parameters are the measured or fitted constants: the empirical onset z*_0, the descriptive feature-width fit, and the auxiliary numerical constants C(σ) and δ. The main load-bearing axioms are the unproved fixed-z*_0 assumption and the standard Poisson summation and curvature-flow background.

free parameters (4)
  • z*_0, cascade onset standardized bias = 0.36±0.03 (magnetization definition); 0.31±0.03 (giant-cluster fraction)
    The central 1/σ window law relies on cascade onset occurring at a fixed, σ-independent value of z0. This is measured by two operational order parameters and is not derived; the paper states it is an assumption supported numerically but not proved (Secs. V B and V C).
  • Feature-width prefactor and exponent in ℓ_w ≈ 10.6 σ^0.30 = 10.6, 0.30
    Descriptive power-law fit over σ∈[4,32]; the paper states it is not claimed as an asymptotic exponent (Sec. VI).
  • Registration spread coefficient δ ≈ 1.26σ = 1.26
    Measured, not derived; used in the universal mobility collapse but not in the central freezing result. The paper states its derivation is open (Sec. IV B, Appendix A).
  • Ring offset coefficient C(σ) ≈ 0.23σ = 0.23
    Characterized numerically, not derived; affects the crossover-radius prefactor only (Sec. IV A).
assumptions (5)
  • standard math Poisson summation and Gaussian row-sum approximation give ℓ_k = √(2π)σ e^{-k^2/2σ^2} up to a correction below 1e-8 for σ≥1.
    Used in Eq. (4) to derive the exact flat-interface pinning constant and stripe threshold.
  • domain assumption The model definition with self-exclusion W_ii=0 and asynchronous updates ensures energy descent for every single flip.
    Model definition in Sec. II; it is what makes the reversible variant halt only at single-spin-flip stable configurations.
  • domain assumption The MBO scheme's continuum limit gives mean-curvature motion with drive coefficient √(2π)σ^3 under the same kernel normalization.
    Cited to Merriman-Bence-Osher and used in Sec. IV A to anchor the curvature coefficient and crossover radius scaling.
  • ad hoc to paper Cascade onset occurs at a fixed, σ-independent standardized bias z*_0.
    Sec. V B states: 'The assumption is that cascade onset occurs at a fixed value z0 = z*_0 independent of σ... supported numerically in Sec. V C but not proved.' This is load-bearing for Eq. (17).
  • ad hoc to paper The standardized registration statistics of rasterized circles are σ-independent, with δ ∝ σ and a platykurtic, anticorrelated distribution.
    Used for the universal mobility collapse in Sec. IV C; the paper states the derivation of δ and the distribution shape is open (Sec. IV B).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Morphology of frozen labyrinths from irreversible threshold dynamics." pith.science (2026). https://pith.science/paper/YQTXLRTW

@misc{pith2026260805496,
  author       = {Pith},
  title        = {Pith review of: Morphology of frozen labyrinths from irreversible threshold dynamics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YQTXLRTW}},
  note         = {Machine review of arXiv:2608.05496}
}
read the original abstract

Majority threshold dynamics, in which each agent adopts the dominant state in a weighted neighborhood, relaxes a binary field toward consensus or stripes. We study what happens when this rule is made irreversible: each agent, interacting through a Gaussian kernel on a lattice, may flip out of its local weighted minority at most once. The reversible form is threshold dynamics of Merriman-Bence-Osher type, an exactly solvable calibration in which interfaces move by mean curvature with closed-form lattice pinning and mobility. Irreversibility changes the outcome. From random initial conditions, the one-flip rule freezes balanced non-consensus labyrinths that reversible relaxation drives away. The patterned regime is a window of initial spin compositions around equal balance, narrowing as the interaction range grows, controlled by a standardized bias whose onset is independent of scale over a fourfold range. The frozen morphology is arrested coarsening: bicontinuous at balance, with a feature width that grows sublinearly and falls below the interaction range at large scales. We determine the mechanism by intervention. Fixing the initial condition while varying the update order shows that the coarse domain layout is deterministic, while the update order enters only in a secondary first-passage race that fine-tunes the wall positions the deterministic dynamics has already set. Those walls are marked by frustration: agents frozen against their own local field. This signature is identically absent from any reversible relaxation, is overwhelmingly interfacial, and carries almost none of the pattern's large-scale shape.

Figures

Figures reproduced from arXiv: 2608.05496 by the authors.

Figure 1
Figure 1. FIG. 1. Critical stripe width. Solid curve: the exact thresh [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Curvature drive and universal mobility. (a) Fraction [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. The composition window and its control by [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Frozen morphology at balance ( [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. The frozen state of our one-flip model (left) and a [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. Interior/border dichotomy: agreement as a function [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: FIG. 7. Irreversibility, frustration, and the structure of the [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

22 extracted references · 21 canonical work pages

  1. [1]

    Throughout, the front remains close to circular and does not develop concave pockets that could arrest it; small boundary fluctuations stay bounded rather than amplifying (App

    still erode, at the slow rateg(z)>0 supplied by the registration tail. Throughout, the front remains close to circular and does not develop concave pockets that could arrest it; small boundary fluctuations stay bounded rather than amplifying (App. A). D. F ront stability and shape relaxation Under the asynchronous one-flip dynamics, a droplet’s boundary i...

    work page

  2. [2]

    The standardized initial bias is therefore z0 = E[m]p Var[m] − − − − − − − − → σ≫1, p0→ 1 2 2√π(2p 0 −1)σ,(16) withO(1/σ 2) corrections from the exact forms (14)–(15)

    Both quantities hold up to Poisson corrections below 10 −8. The standardized initial bias is therefore z0 = E[m]p Var[m] − − − − − − − − → σ≫1, p0→ 1 2 2√π(2p 0 −1)σ,(16) withO(1/σ 2) corrections from the exact forms (14)–(15). The initial bias, measured in units of the initial fluctua- tions, is thus (2p 0 −1)σup to a constant. The same standardizationz ...

    work page

  3. [3]

    Since the frozen state is already smooth at that scale (Sec

    Interventions and reference constructions We resolve each site by its distance to the nearest coarse domain wall, defined as the zero level of the frozen configuration coarse-grained at the interaction scaleσ. Since the frozen state is already smooth at that scale (Sec. VII D), this coarse-graining alters only the sub- σfrustrated skin and leaves the doma...

    work page

  4. [4]

    The three stages The frozen pattern is built in three stages: (i) the coarse domains are laid down by the initial condition and the deterministic dynamics; (ii) their walls are then placed, predominantly deterministically with a secondary stochastic contribution; and (iii) frustration locks the placed walls without moving them. Domain formation.Deep insid...

    work page

  5. [5]

    The pat- terned window is controlled by the exact standardized biasz 0 ≃2 √π(2p0 −1)σ, with a critical valuez ∗ 0 in- dependent ofσwithin resolution, and hence giving a 1/σwindow

    It admits a quantitative phase structure. The pat- terned window is controlled by the exact standardized biasz 0 ≃2 √π(2p0 −1)σ, with a critical valuez ∗ 0 in- dependent ofσwithin resolution, and hence giving a 1/σwindow. This boundary appears to be genuinely dynamical: it lies beyond every static percolation thresh- old of the initial field we examined. ...

    work page 2026

  6. [6]

    Bence, and Stanley J

    Barry Merriman, James K. Bence, and Stanley J. Osher. Motion of Multiple Junctions: A Level Set Approach. Journal of Computational Physics, 112(2):334–363, June 1994

    work page 1994

  7. [7]

    Propagation and Its Failure in Coupled Systems of Discrete Excitable Cells.SIAM Journal on Applied Mathematics, 47(3), 1987

    James Keener. Propagation and Its Failure in Coupled Systems of Discrete Excitable Cells.SIAM Journal on Applied Mathematics, 47(3), 1987

    work page 1987

  8. [8]

    Spirin, P.L

    V. Spirin, P.L. Krapivsky, and S. Redner. Freezing in Ising ferromagnets.Physical Review E, 65(1):016119, De- cember 2001

    work page 2001

Show all 22 references
  1. [9]

    Holm, James A

    Elizabeth A. Holm, James A. Glazier, David J. Srolovitz, and Gary S. Grest. Effects of lattice anisotropy and tem- perature on domain growth in the two-dimensional Potts model.Physical Review A, 43(6):2662–2668, March 1991

    1991

  2. [10]

    Duncan J. Watts. A simple model of global cascades on random networks.Proceedings of the National Academy of Sciences, 99(9):5766–5771, April 2002

    2002

  3. [11]

    Gleeson and Diarmuid J

    James P. Gleeson and Diarmuid J. Cahalane. Seed size strongly affects cascades on random networks.Physical Review E, 75(5):056103, May 2007

    2007

  4. [12]

    Complex Contagions and the Weakness of Long Ties1.American Journal of Sociology, 113(3):702–734, November 2007

    Damon Centola and Michael Macy. Complex Contagions and the Weakness of Long Ties1.American Journal of Sociology, 113(3):702–734, November 2007

    2007

  5. [13]

    Rational group decision making: A random field Ising model at T = 0.Physica A: Statistical Mechan- ics and its Applications, 238(1-4):66–80, April 1997

    Serge Galam. Rational group decision making: A random field Ising model at T = 0.Physica A: Statistical Mechan- ics and its Applications, 238(1-4):66–80, April 1997

    1997

  6. [14]

    On the role of zealotry in the voter model.Journal of Statistical Me- chanics: Theory and Experiment, 2007(08):P08029, Au- gust 2007

    M Mobilia, A Petersen, and S Redner. On the role of zealotry in the voter model.Journal of Statistical Me- chanics: Theory and Experiment, 2007(08):P08029, Au- gust 2007

    2007

  7. [15]

    Latoski, W

    Lu ´ ıs Carlos F. Latoski, W. G. Dantas, and Jeferson J. Arenzon. Opinion inertia and coarsening in the persis- tent voter model.Physical Review E, 109(5):054115, May 2024

    2024

  8. [16]

    Aizenman and J

    M. Aizenman and J. L. Lebowitz. Metastability effects in bootstrap percolation.Journal of Physics A: Mathe- matical and General, 21(19):3801, October 1988

    1988

  9. [17]

    Sharp metastability threhold for two-dimensional bootstrap percolation.Probability The- ory and Related Fields, 125(2):195–224, 2003

    Alexander Holroyd. Sharp metastability threhold for two-dimensional bootstrap percolation.Probability The- ory and Related Fields, 125(2):195–224, 2003

    2003

  10. [18]

    First passage percolation and a model for competing spatial growth

    Olle H¨ aggstr¨ om and Robin Pemantle. First passage percolation and a model for competing spatial growth. Journal of Applied Probability, 35(3):683–692, Septem- ber 1998

    1998

  11. [19]

    Random growth models with polygonal shapes.The Annals of Probability, 34(1):181–218, January 2006

    Janko Gravner and David Griffeath. Random growth models with polygonal shapes.The Annals of Probability, 34(1):181–218, January 2006

    2006

  12. [20]

    Gage and R

    M. Gage and R. S. Hamilton. The heat equation shrink- ing convex plane curves.Journal of Differential Geome- try, 23(1):69–96, January 1986

    1986

  13. [21]

    Matthew A. Grayson. The heat equation shrinks embed- ded plane curves to round points.Journal of Differential Geometry, 26(2):285–314, January 1987

    1987

  14. [22]

    Contagion.The Review of Economic Studies, 67(1):57–78, 2000

    Stephen Morris. Contagion.The Review of Economic Studies, 67(1):57–78, 2000. Appendix A: Numerical methods Dynamics.Exact Gillespie simulation on anL×L torus: the eligible set is maintained incrementally, an eligible site is chosen uniformly, and time advances by an exponentia...

    2000

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.