pith. sign in

arxiv: 2604.16288 · v1 · submitted 2026-04-17 · 🧮 math.AP · cond-mat.stat-mech· math-ph· math.MP· math.PR· stat.ML

Phase transitions in Doi-Onsager, Noisy Transformer, and other multimodal models

Kyunghoo Mun , Matthew Rosenzweig This is my paper

Pith reviewed 2026-05-10 07:31 UTC · model grok-4.3

classification 🧮 math.AP cond-mat.stat-mechmath-phmath.MPmath.PRstat.ML
keywords phase transitionsmean-field free energyDoi-Onsager modelnoisy transformerHegselmann-Krause modelLebedev-Milin inequalityperiodic interactionscontinuous transition
0
0 comments X

The pith

For 1/(n+1)-periodic interactions with decaying Fourier coefficients, the critical coupling equals the linear stability threshold and the phase transition is continuous.

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

The paper establishes that repulsive-attractive mean-field interactions on the circle that are periodic with period 1/(n+1) and whose Fourier coefficients decay at a sufficient rate have their critical coupling strength K_c exactly equal to the linear stability threshold K_# of the uniform distribution. At this critical value the uniform distribution remains the unique global minimizer of the free energy, so the transition to nonuniform states is continuous. The proof rests on a sharp coercivity lower bound for the free energy that follows from the constrained Lebedev-Milin inequality. The general result is applied to three concrete models, fixing the exact location and continuity type of their previously unresolved phase transitions.

Core claim

For a 1/(n+1)-periodic interaction whose Fourier coefficients satisfy a certain decay condition, the critical coupling strength K_c coincides with the linear stability threshold K_# of the uniform distribution and the phase transition is continuous, in the sense that the uniform distribution is the unique global minimizer at criticality. The proof is based on a sharp coercivity estimate for the free energy obtained from the constrained Lebedev-Milin inequality. The result determines that the two-dimensional Doi-Onsager model undergoes a continuous transition at K_c = K_# = 3π/4, while the noisy transformer and noisy Hegselmann-Krause models each exhibit a sharp parameter threshold separating

What carries the argument

The constrained Lebedev-Milin inequality, which supplies the sharp coercivity estimate for the mean-field free energy functional on the circle.

If this is right

  • In the Doi-Onsager model the transition is continuous precisely at K_c = K_# = 3π/4.
  • For the noisy transformer interaction there is a threshold β_* such that the transition is continuous with K_c = K_# for β ≤ β_* and discontinuous with K_c < K_# for β > β_*.
  • The noisy Hegselmann-Krause model obeys the same sharp dichotomy between continuous and discontinuous regimes.
  • Whenever the interaction satisfies the stated periodicity and decay conditions, linear stability analysis alone locates the onset of the phase transition.

Where Pith is reading between the lines

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

  • The same coercivity technique may classify the continuity of transitions for other multimodal potentials once their Fourier decay is verified.
  • Numerical checks of the Fourier decay condition could immediately extend the continuous-transition conclusion to additional models on the circle.
  • If analogous inequalities exist on higher-dimensional manifolds, the result would predict continuous transitions from linear thresholds in those settings as well.

Load-bearing premise

The interaction must be exactly 1/(n+1)-periodic and its Fourier coefficients must decay fast enough for the constrained Lebedev-Milin inequality to deliver a sharp coercivity bound on the free energy.

What would settle it

Direct numerical minimization of the free energy at the predicted critical coupling for any interaction satisfying the periodicity and decay conditions, checking whether any non-uniform density achieves strictly lower energy than the uniform distribution.

read the original abstract

We study phase transitions for repulsive-attractive mean-field free energies on the circle. For a $\frac{1}{n+1}$-periodic interaction whose Fourier coefficients satisfy a certain decay condition, we prove that the critical coupling strength $K_c$ coincides with the linear stability threshold $K_\#$ of the uniform distribution and that the phase transition is continuous, in the sense that the uniform distribution is the unique global minimizer at criticality. The proof is based on a sharp coercivity estimate for the free energy obtained from the constrained Lebedev--Milin inequality. We apply this result to three motivating models for which the exact value of the phase transition and its (dis)continuity in terms of the model parameters was not fully known. For the two-dimensional Doi--Onsager model $W(\theta)=-|\sin(2\pi\theta)|$, we prove that the phase transition is continuous at $K_c=K_\#=3\pi/4$. For the noisy transformer model $W_\beta(\theta)=(e^{\beta\cos(2\pi\theta)}-1)/\beta$, we identify the sharp threshold $\beta_*$ such that $K_c(\beta) = K_\#(\beta)$ and the phase transition is continuous for $\beta \leq \beta_*$, while $K_c(\beta)<K_\#(\beta)$ and the phase transition is discontinuous for $\beta > \beta_*$. We also obtain the corresponding sharp dichotomy for the noisy Hegselmann--Krause model $W_{R}(\theta) = (R-2\pi|\theta|)_{+}^2$ .

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 / 2 minor

Summary. The manuscript proves that for 1/(n+1)-periodic interaction potentials on the circle whose Fourier coefficients satisfy a stated decay condition, the critical coupling K_c for the associated mean-field free energy coincides with the linear stability threshold K_# of the uniform distribution, and the phase transition is continuous (uniform distribution is the unique global minimizer at criticality). The proof rests on a sharp coercivity lower bound for the free energy obtained via the constrained Lebedev-Milin inequality. The result is applied to the 2D Doi-Onsager model W(θ) = -|sin(2πθ)| (yielding continuous transition at K_c = K_# = 3π/4), the noisy transformer model W_β(θ) = (e^{β cos(2πθ)} - 1)/β (identifying a threshold β_* separating continuous and discontinuous regimes), and the noisy Hegselmann-Krause model W_R(θ) = (R - 2π|θ|)_+² (obtaining the analogous sharp dichotomy).

Significance. If the central claims hold, the work supplies rigorous, parameter-explicit resolutions to the continuity/discontinuity and exact location of phase transitions in three concrete multimodal mean-field models whose behavior was previously only partially characterized. The general theorem via the constrained Lebedev-Milin inequality furnishes a reusable coercivity tool for periodic interactions meeting the Fourier-decay hypothesis, and the explicit model applications (including the concrete value 3π/4 for Doi-Onsager) are directly usable in statistical mechanics and machine-learning contexts.

major comments (2)
  1. [Main theorem and coercivity estimate (Section 2)] The central uniqueness claim at K = K_# rests on the coercivity estimate being sharp, i.e., the quadratic form controlling deviations from uniformity is positive semi-definite with kernel precisely the constants. The manuscript must explicitly verify that equality cases in the constrained Lebedev-Milin inequality are attained only for constant densities when the interaction is 1/(n+1)-periodic and the Fourier coefficients obey the decay condition; otherwise the uniqueness statement fails even if linear stability holds. This verification should appear in the proof of the main theorem (likely §2 or the statement following the inequality application).
  2. [Noisy transformer model (Section 4)] For the noisy transformer application, the threshold β_* is defined by the point at which the Fourier coefficients of W_β cease to satisfy the decay hypothesis required by the general theorem. The manuscript should supply the explicit computation of β_* together with a direct check that the decay condition holds for all β ≤ β_* (and fails for β > β_*), since this determines the switch from K_c = K_# (continuous) to K_c < K_# (discontinuous).
minor comments (2)
  1. [Introduction] The notation K_c versus K_# is introduced in the abstract but should be defined with a short sentence in the introduction before the statement of the main result.
  2. [Doi-Onsager application] In the Doi-Onsager application, the numerical value K_c = 3π/4 should be cross-referenced to the explicit Fourier coefficients of W(θ) = -|sin(2πθ)| to make the verification self-contained.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We are grateful to the referee for the thorough review and valuable suggestions. Below we address the major comments point by point, indicating the revisions we will make to the manuscript.

read point-by-point responses

  1. Referee: [Main theorem and coercivity estimate (Section 2)] The central uniqueness claim at K = K_# rests on the coercivity estimate being sharp, i.e., the quadratic form controlling deviations from uniformity is positive semi-definite with kernel precisely the constants. The manuscript must explicitly verify that equality cases in the constrained Lebedev-Milin inequality are attained only for constant densities when the interaction is 1/(n+1)-periodic and the Fourier coefficients obey the decay condition; otherwise the uniqueness statement fails even if linear stability holds. This verification should appear in the proof of the main theorem (likely §2 or the statement following the inequality application).

    Authors: We agree with the referee that the equality cases require explicit verification to rigorously establish the sharpness of the coercivity bound and the uniqueness of the uniform minimizer at K = K_#. In the revised version, we will include this verification directly in the proof of the main theorem in Section 2. Under the 1/(n+1)-periodicity assumption, the Fourier series of the interaction potential has only every (n+1)th coefficient nonzero, which, combined with the decay condition, allows us to show that equality in the constrained Lebedev-Milin inequality holds solely for constant functions. This will be demonstrated by analyzing the associated quadratic form and using the strict inequality properties for non-constant perturbations satisfying the periodicity. revision: yes

  2. Referee: [Noisy transformer model (Section 4)] For the noisy transformer application, the threshold β_* is defined by the point at which the Fourier coefficients of W_β cease to satisfy the decay hypothesis required by the general theorem. The manuscript should supply the explicit computation of β_* together with a direct check that the decay condition holds for all β ≤ β_* (and fails for β > β_*), since this determines the switch from K_c = K_# (continuous) to K_c < K_# (discontinuous).

    Authors: We thank the referee for highlighting the need for explicit details on β_*. While the manuscript identifies β_* as the critical value separating the regimes, we will revise Section 4 to provide the explicit computation of β_* by determining the infimum of β where the Fourier coefficients of W_β violate the decay condition (specifically, the rate at which |ĉ_k(W_β)| decays). We will include direct checks: for β ≤ β_*, we verify the inequality holds with explicit constants derived from the series expansion of the exponential, and for β > β_*, we exhibit a specific mode where the decay fails. This will make the transition between continuous and discontinuous phase transitions fully explicit and verifiable. revision: yes

Circularity Check

0 steps flagged

Direct proof via Fourier analysis and established inequality; no reduction to inputs by construction

full rationale

The paper's central result is a theorem establishing K_c = K_# with uniqueness of the uniform minimizer at criticality, proved by deriving a sharp coercivity estimate for the free energy from the constrained Lebedev-Milin inequality under the stated periodicity and Fourier decay hypotheses. This inequality is an external, pre-existing result in complex analysis; the paper applies it rather than deriving or assuming the target conclusion. The applications to the Doi-Onsager, noisy transformer, and Hegselmann-Krause models consist of verifying that their interaction potentials satisfy the hypotheses, after which the general theorem applies directly. No parameters are fitted, no self-citations form a load-bearing chain, no ansatz is smuggled, and no known result is merely renamed. The derivation chain is therefore self-contained against external mathematical benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the validity of the constrained Lebedev-Milin inequality as a source of sharp coercivity and on the periodicity plus Fourier decay assumption for the interaction; no free parameters or new entities are introduced.

axioms (1)
  • standard math constrained Lebedev-Milin inequality
    Invoked to obtain the sharp coercivity estimate for the free energy in the general theorem.

pith-pipeline@v0.9.0 · 5614 in / 1382 out tokens · 74526 ms · 2026-05-10T07:31:12.732729+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

29 extracted references · 29 canonical work pages

  1. [1]

    On the Structure of Stationary Solutions to

    Krishnakumar Balasubramanian, Sayan Banerjee, and Philippe Rigollet. On the structure of stationary solutions to McKean-Vlasov equations with applications to noisy transformers. arXiv preprint arXiv:2510.20094 , 2025

    work page arXiv 2025

  2. [2]

    A Sharp Inequality on the Exponentiation of Functions on the Sphere

    Sun-Yung Alice Chang and Changfeng Gui. A Sharp Inequality on the Exponentiation of Functions on the Sphere . Communications on Pure and Applied Mathematics , 76(6):1303--1326, 2023

    work page 2023

  3. [3]

    J. A. Carrillo, R. S. Gvalani, G. A. Pavliotis, and A. Schlichting. Long-time behaviour and phase transitions for the McKean-Vlasov equation on the torus. Arch. Ration. Mech. Anal. , 235(1):635--690, 2020

    work page 2020

  4. [4]

    Wasserstein-\ L \ ojasiewicz inequalities and asymptotics of McKean - Vlasov equation, November 2025

    Beomjun Choi, Seunghoon Jeong, and Geuntaek Seo. Wasserstein-\ L \ ojasiewicz inequalities and asymptotics of McKean - Vlasov equation, November 2025

    work page 2025

  5. [5]

    Brownian particles with electrostatic repulsion on the circle: Dyson 's model for unitary random matrices revisited

    Emmanuel C \'e pa and Dominique L \'e pingle. Brownian particles with electrostatic repulsion on the circle: Dyson 's model for unitary random matrices revisited. ESAIM Probab. Statist. , 5:203--224, 2001

    work page 2001

  6. [6]

    On the stationary solutions of the 2D Doi -- Onsager model

    Wenxiong Chen, Congming Li, and Guofang Wang. On the stationary solutions of the 2D Doi -- Onsager model. Nonlinear Analysis: Theory, Methods & Applications , 73(8):2410--2425, 2010

    work page 2010

  7. [7]

    Chayes and V

    L. Chayes and V. Panferov. The McKean-Vlasov equation in finite volume. J. Stat. Phys. , 138(1-3):351--380, 2010

    work page 2010

  8. [8]

    Exact solvability of the macroscopic circular log gas

    Antonin Chodron de Courcel , Matthew Rosenzweig, and Sylvia Serfaty. Exact solvability of the macroscopic circular log gas. In preparation

    work page

  9. [9]

    The attractive log gas: Stability , uniqueness, and propagation of chaos

    Antonin Chodron de Courcel , Matthew Rosenzweig, and Sylvia Serfaty. The attractive log gas: Stability , uniqueness, and propagation of chaos. Commun. Am. Math. Soc. , 5:695--773, 2025

    work page 2025

  10. [10]

    Masao Doi and Samuel F. Edwards. The Theory of Polymer Dynamics . Cambridge University Press, Cambridge, 1986

    work page 1986

  11. [11]

    Phase transitions, logarithmic Sobolev inequalities, and uniform-in-time propagation of chaos for weakly interacting diffusions

    Mat \' as G Delgadino, Rishabh S Gvalani, Grigorios A Pavliotis, and Scott A Smith. Phase transitions, logarithmic Sobolev inequalities, and uniform-in-time propagation of chaos for weakly interacting diffusions. Communications in Mathematical Physics , 401(1):275--323, 2023

    work page 2023

  12. [12]

    Univalent functions , volume 259

    Peter L Duren. Univalent functions , volume 259. Springer Science & Business Media, 2001

    work page 2001

  13. [13]

    Log-Gases and Random Matrices , volume 34 of London Mathematical Society Monographs

    Peter (Peter John) Forrester. Log-Gases and Random Matrices , volume 34 of London Mathematical Society Monographs . Princeton University Press,, Princeton :, course book edition, 2010

    work page 2010

  14. [14]

    A mathematical perspective on transformers

    Borjan Geshkovski, Cyril Letrouit, Yury Polyanskiy, and Philippe Rigollet. A mathematical perspective on transformers. Bulletin of the American Mathematical Society , 62(3):427--479, 2025

    work page 2025

  15. [15]

    The van der Waals limit for classical systems: III

    DJ Gates and O Penrose. The van der Waals limit for classical systems: III . Deviation from the van der Waals-Maxwell theory. Communications in Mathematical Physics , 17(3):194--209, 1970

    work page 1970

  16. [16]

    Toeplitz Forms and Their Applications

    Ulf Grenander and Gabor Szego. Toeplitz Forms and Their Applications . Chelsea Publishing Co., New York, 2 edition, 1984

    work page 1984

  17. [17]

    Opinion Dynamics and Bounded Confidence Models , Analysis and Simulation

    Rainer Hegselmann and Ulrich Krause. Opinion Dynamics and Bounded Confidence Models , Analysis and Simulation . Journal of Artificial Societies and Social Simulation , 5(3):1--2, 2002

    work page 2002

  18. [18]

    N. A. Lebedev and I. M. Milin. On one inequality. Vestnik Leningrad. Univ. , 20(19):157--158, 1965. Russian, with English summary. Original Russian title: ``Ob odnom neravenstve''

    work page 1965

  19. [19]

    Phase transitions and linear stability for the mean-field Kuramoto - Daido model, February 2026

    Kyunghoo Mun and Matthew Rosenzweig. Phase transitions and linear stability for the mean-field Kuramoto - Daido model, February 2026. arXiv:2602.14954 [math]

    work page arXiv 2026

  20. [20]

    On stationary solutions of the 2D Doi -- Onsager model

    Mohammad Ali Niksirat and Xinwei Yu. On stationary solutions of the 2D Doi -- Onsager model. Journal of Mathematical Analysis and Applications , 430(1):152--165, 2015

    work page 2015

  21. [21]

    The effects of shape on the interaction of colloidal particles

    Lars Onsager. The effects of shape on the interaction of colloidal particles. Ann. N. Y. Acad. Sci. , 51(4):627--659, 1949

    work page 1949

  22. [22]

    Extremals of determinants of Laplacians

    B Osgood, R Phillips, and P Sarnak. Extremals of determinants of Laplacians . Journal of Functional Analysis , 80(1):148--211, September 1988

    work page 1988

  23. [23]

    On an inequality for modified bessel functions

    RP Soni. On an inequality for modified bessel functions. Journal of Mathematics and Physics , 44(1-4):406--407, 1965

    work page 1965

  24. [24]

    Solutions of stationary McKean- Vlasov equation on a high-dimensional sphere and other Riemannian manifolds.arXiv preprint arXiv:2412.14813,

    Anna Shalova and Andr \'e Schlichting. Solutions of stationary McKean-Vlasov equation on a high-dimensional sphere and other Riemannian manifolds. arXiv preprint arXiv:2412.14813 , 2024

    work page arXiv 2024

  25. [25]

    Note on Uniqueness of the Isotropic Solution of the 2D Doi-Onsager Model

    Mitchell A Taylor and Xinwei Yu. Note on Uniqueness of the Isotropic Solution of the 2D Doi-Onsager Model

    work page

  26. [26]

    Attention is all you need

    A Vaswani. Attention is all you need. Advances in Neural Information Processing Systems , 2017

    work page 2017

  27. [27]

    A Treatise on the Theory of Bessel Functions , volume 2

    George Neville Watson. A Treatise on the Theory of Bessel Functions , volume 2. The University Press, 1922

    work page 1922

  28. [28]

    On an inequality of Osgood , Phillips and Sarnak

    Harold Widom. On an inequality of Osgood , Phillips and Sarnak . Proc. Amer. Math. Soc. , 102(3):773--774, 1988

    work page 1988

  29. [29]

    Noisy Hegselmann - Krause Systems : Phase Transition and the 2R - Conjecture

    Chu Wang, Qianxiao Li, Weinan E, and Bernard Chazelle. Noisy Hegselmann - Krause Systems : Phase Transition and the 2R - Conjecture . Journal of Statistical Physics , 166(5):1209--1225, March 2017

    work page 2017