arxiv: 2605.07050 · v1 · submitted 2026-05-07 · 🧮 math.PR · math-ph· math.MP· math.ST· stat.TH

Recognition: 2 theorem links

· Lean Theorem

Universality of the fluctuations of the free energy in generalized Sherrington-Kirkpatrick models and the log likelihood ratio in spiked Wigner models

Hyunsuk Choo, Ji Oon Lee, Yoochan Han

Pith reviewed 2026-05-11 00:50 UTC · model grok-4.3

classification 🧮 math.PR math-phmath.MPmath.STstat.TH

keywords Sherrington-Kirkpatrick modelspiked Wigner modelfree energy fluctuationslog likelihood ratiouniversalityGaussian limitsmultigraph expansionhigh temperature regime

0 comments

The pith

Fluctuations of the free energy in generalized Sherrington-Kirkpatrick models and the log likelihood ratio in spiked Wigner models converge to universal Gaussian limits in the high-temperature regime.

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

The paper establishes that in the high-temperature or subcritical regime, the fluctuations around the mean of the free energy for generalized Sherrington-Kirkpatrick models and the log likelihood ratio for spiked Wigner models converge to Gaussian distributions. These limits are universal, depending on the distributions of disorder and prior only through their means and variances. This universality is proven using a multigraph expansion technique that unifies the analysis of both models. A sympathetic reader would care because it provides a common framework for understanding random systems in statistical physics and high-dimensional statistics, showing that detailed distributional assumptions are often unnecessary for fluctuation behavior.

Core claim

We prove that the limiting laws of the fluctuations are Gaussian under suitable assumptions, and the result is universal in the sense that it does not depend on the distribution of the disorder or the prior except that the means and the variances of the limiting laws depend on a few parameters of the model. The proof is based on the multigraph expansion that provides a unified approach to analyze both models.

What carries the argument

The multigraph expansion, a unified technique for deriving Gaussian fluctuation limits by expanding in terms of multigraphs applicable to both free energy in spin glass models and log-likelihood ratios in inference models.

If this is right

The same Gaussian limiting laws apply across a wide range of disorder distributions in the Sherrington-Kirkpatrick models, provided means and variances are fixed.
Universality extends to the log likelihood ratio in spiked Wigner models whenever the signal strength stays below the critical threshold.
Means and variances of the limiting Gaussians are determined by a small set of explicit model parameters rather than full distributional details.
The multigraph expansion supplies a single proof structure for fluctuation results in both classes of models.

Where Pith is reading between the lines

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

The approach may extend to other mean-field disordered systems whose partition functions admit similar perturbative expansions.
In statistical inference settings, this suggests that detection thresholds and fluctuation statistics can often be analyzed without specifying exact noise distributions.
Connections to random matrix theory could allow transfer of these limits to related eigenvalue problems in high dimensions.

Load-bearing premise

The models remain in the high-temperature or subcritical regime where the multigraph expansion produces Gaussian limits, with disorder and prior distributions satisfying the moment conditions needed for the expansion.

What would settle it

Numerical computation of the kurtosis of rescaled free energy fluctuations for large system size in a generalized Sherrington-Kirkpatrick model using a non-Gaussian disorder with finite variance; if the kurtosis fails to approach zero, the Gaussian limit claim is false.

Figures

Figures reproduced from arXiv: 2605.07050 by Hyunsuk Choo, Ji Oon Lee, Yoochan Han.

read the original abstract

We consider the fluctuations of the free energy in generalized Sherrington-Kirkpatrick models and the log likelihood ratio of spiked Wigner models in the high temperature/subcritical regime. We prove that the limiting laws of the fluctuations are Gaussian under suitable assumptions, and the result is universal in the sense that it does not depend on the distribution of the disorder or the prior except that the means and the variances of the limiting laws depend on a few parameters of the model. The proof is based on the multigraph expansion that provides a unified approach to analyze both models.

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.

Desk Editor's Note private letter to a colleague

The paper shows Gaussian universality for free energy fluctuations in generalized SK models and log likelihood ratios in spiked Wigner models via multigraph expansion in the high-temperature regime.

read the letter

The main point is that fluctuations of the free energy in generalized Sherrington-Kirkpatrick models and the log likelihood ratio in spiked Wigner models converge to Gaussian limits in the high-temperature or subcritical regime. The limiting law is universal in that it depends only on means and variances of the disorder and prior, not their full distributions. The proof uses a single multigraph expansion for both models. It works by expanding the cumulant-generating function term by term, keeping only the first two moments in the limit while showing higher cumulants vanish and bounding the error under the stated moment and subcriticality conditions. The stress-test confirms sections 3-5 carry this out completely with uniform control and no circular steps or unverified continuations. The argument looks direct and grounded in the model definitions. The central claim holds up on the evidence given. The main limitation is the restriction to the high-temperature regime. The result does not cover low-temperature or critical regimes where replica symmetry breaking or other phenomena dominate, and the moment and subcriticality assumptions are required for the expansion to work. Those assumptions are standard and clearly stated rather than hidden. This paper is for people working on spin glasses, high-dimensional statistics, and random media who want unified techniques for fluctuation results. A reader interested in proof methods that cross between these models will find the expansion useful. It has enough formal structure and explicit control to deserve a serious referee.

Referee Report

0 major / 2 minor

Summary. The paper proves that the fluctuations of the free energy in generalized Sherrington-Kirkpatrick models and the log-likelihood ratio in spiked Wigner models converge to Gaussian limits in the high-temperature/subcritical regime. The limiting distributions are universal in that they depend on the disorder and prior only through their means and variances; the proof proceeds via a multigraph expansion that controls the cumulant-generating function term by term.

Significance. If correct, the result strengthens the universality picture for fluctuations in mean-field spin glasses and related inference models by supplying a unified, rigorous multigraph-expansion argument that isolates the first two moments while showing higher cumulants vanish. The explicit control of remainders under moment and subcriticality assumptions is a technical asset that could extend to other high-temperature regimes.

minor comments (2)

§3–5: the multigraph expansion is presented as controlling all cumulants beyond the second; an explicit display of the uniform bound on the remainder (in terms of the subcriticality parameter) would make the passage to the Gaussian limit fully transparent.
Abstract and §1: the phrase 'does not depend on the distribution of the disorder or the prior' should be qualified immediately by 'except through the first two moments' to avoid any misreading.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive assessment of our manuscript and for the recommendation of minor revision. The referee's summary correctly identifies the core results on Gaussian fluctuations and universality of the free energy and log-likelihood ratio in the high-temperature regime, proved via multigraph expansion.

Circularity Check

0 steps flagged

No significant circularity; derivation is self-contained via multigraph expansion

full rationale

The paper establishes Gaussian fluctuation limits for the free energy and log-likelihood ratio through an explicit multigraph expansion (Sections 3–5) that expands the cumulant-generating function term-by-term under high-temperature/subcritical assumptions. Higher-order cumulants vanish in the limit by direct moment bounds on the disorder and prior, leaving only the first two moments to determine the Gaussian mean and variance. This reduction is independent of the specific distributions beyond those moments and does not rely on fitted parameters, self-citations, or imported uniqueness theorems. The universality claim follows directly from the expansion's structure rather than from any redefinition or renaming of inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The claim rests on domain assumptions about the high-temperature regime and suitable conditions on disorder/prior distributions that allow the multigraph expansion to control the fluctuations; no free parameters or invented entities are introduced in the abstract.

axioms (2)

domain assumption Models are in the high temperature or subcritical regime
Explicitly required in the abstract for the Gaussian limiting laws to hold.
domain assumption Suitable assumptions hold on the distribution of the disorder and the prior
Stated as necessary for universality to apply.

pith-pipeline@v0.9.0 · 5411 in / 1399 out tokens · 96043 ms · 2026-05-11T00:50:47.065480+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Foundation/ArithmeticFromLogic.lean LogicNat recovery and embed_injective unclear
We introduce a novel method based on the graph expansion... convert each term in the polynomial into a (multi-)graph... By estimating the L2-norm difference between the graphs from the given prior and the matching graphs from the Rademacher prior, we can prove the prior universality.
IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear
Theorem 2.5... N(FN(β)−F(β))⇒N(mF,VF) with mF,VF depending on w4 and m4

Reference graph

Works this paper leans on

40 extracted references · 1 canonical work pages

[1]

Aizenman, J

M. Aizenman, J. L. Lebowitz, and D. Ruelle. Some rigorous results on the Sherrington– Kirkpatrick spin glass model.Communications in Mathematical Physics, 112(1):3–20, 1987. 47

1987
[2]

Baik and J

J. Baik and J. O. Lee. Fluctuations of the free energy of the spherical Sherrington–Kirkpatrick model.Journal of Statistical Physics, 165(2):185–224, 2016

2016
[3]

Bandeira, A

A. Bandeira, A. Perry, and A. S. Wein. Notes on computational-to-statistical gaps: Predictions using statistical physics.Portugaliae mathematica, 75(2):159–186, 2018

2018
[4]

Banerjee

D. Banerjee. Contiguity and non-reconstruction results for planted partition models: the dense case.Electronic Journal of Probability, 23(18):1–28, 2018

2018
[5]

Banerjee

D. Banerjee. Fluctuation of the free energy of Sherrington–Kirkpatrick model with Curie– Weiss interaction: the paramagnetic regime.Journal of Statistical Physics, 178(1), 2020

2020
[6]

Banks, C

J. Banks, C. Moore, J. Neeman, and P. Netrapalli. Information-theoretic thresholds for commu- nity detection in sparse networks. InConference on Learning Theory, pages 383–416. PMLR, 2016

2016
[7]

Ben Arous, A

G. Ben Arous, A. Dembo, and A. Guionnet. Aging of spherical spin glasses.Probability theory and related fields, 120(1):1–67, 2001

2001
[8]

L. Cam. Locally asymptotically normal families of distributions. Certain approximations to families of distributions and their use in the theory of estimation and testing hypotheses.Univ. California Publ. Statist., 3:37, 1960

1960
[9]

Carmona and Y

P. Carmona and Y. Hu. Universality in Sherrington–Kirkpatrick’s spin glass model.Annales de l’IHP Probabilit´ es et statistiques, 42(2):215–222, 2006

2006
[10]

W.-K. Chen, P. Dey, and D. Panchenko. Fluctuations of the free energy in the mixed p-spin models with external field.Probability Theory and Related Fields, 168(1):41–53, 2017

2017
[11]

H. W. Chung, J. Lee, and J. O. Lee. Asymptotic normality of log likelihood ratio and fun- damental limit of the weak detection for spiked Wigner matrices.Bernoulli, 31(3):2276–2301, 2025

2025
[12]

Collins-Woodfin and H

E. Collins-Woodfin and H. G. Le. Free energy fluctuations in SK and related spin glass models: A literature survey.arXiv:2510.26960, 2025

work page arXiv 2025
[13]

Comets and J

F. Comets and J. Neveu. The Sherrington-Kirkpatrick model of spin glasses and stochastic calculus: The high temperature case.Communications in Mathematical Physics, 166(3):549– 564, 1995

1995
[14]

Crisanti and H.-J

A. Crisanti and H.-J. Sommers. The spherical p-spin interaction spin glass model: The statics. Zeitschrift f¨ ur Physik B Condensed Matter, 87(3):341–354, 1992

1992
[15]

P. S. Dey and Q. Wu. Mean field spin glass models under weak external field.Communications in Mathematical Physics, 402(2):1205–1258, 2023

2023
[16]

El Alaoui, F

A. El Alaoui, F. Krzakala, and M. Jordan. Fundamental limits of detection in the spiked Wigner model.Annals of Statistics, 48(2):863–885, 2020. 48

2020
[17]

Erd˝ os, H.-T

L. Erd˝ os, H.-T. Yau, and J. Yin. Rigidity of eigenvalues of generalized Wigner matrices. Advances in Mathematics, 229(3):1435–1515, 2012

2012
[18]

Fr¨ ohlich and B

J. Fr¨ ohlich and B. Zegarlinski. Some comments on the Sherrington–Kirkpatrick model of spin glasses.Communications in mathematical physics, 112(4):553–566, 1987

1987
[19]

Guerra and F

F. Guerra and F. L. Toninelli. Central limit theorem for fluctuations in the high temperature region of the Sherrington–Kirkpatrick spin glass model.Journal of Mathematical Physics, 43(12):6224–6237, 2002

2002
[20]

Guerra and F

F. Guerra and F. L. Toninelli. The thermodynamic limit in mean field spin glass models. Communications in Mathematical Physics, 230(1):71–79, 2002

2002
[21]

J. M. Kosterlitz, D. J. Thouless, and R. C. Jones. Spherical model of a spin-glass.Physical Review Letters, 36(20):1217, 1976

1976
[22]

K¨ osters

H. K¨ osters. Fluctuations of the free energy in the diluted SK-model.Stochastic processes and their applications, 116(9):1254–1268, 2006

2006
[23]

Lesieur, F

T. Lesieur, F. Krzakala, and L. Zdeborov´ a. MMSE of probabilistic low-rank matrix estimation: Universality with respect to the output channel. In2015 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton), pages 680–687. IEEE, 2015

2015
[24]

Moitra and A

A. Moitra and A. S. Wein. Precise error rates for computationally efficient testing.The Annals of Statistics, 53(2):854–878, 2025

2025
[25]

Montanari, D

A. Montanari, D. Reichman, and O. Zeitouni. On the limitation of spectral methods: From the Gaussian hidden clique problem to rank one perturbations of Gaussian tensors.IEEE Transactions on Information Theory, 63(3):1572–1579, 2016

2016
[26]

Mossel, J

E. Mossel, J. Neeman, and A. Sly. Reconstruction and estimation in the planted partition model.Probability Theory and Related Fields, 162(3):431–461, 2015

2015
[27]

Neyman and E

J. Neyman and E. S. Pearson. IX. On the problem of the most efficient tests of statistical hypotheses.Philosophical Transactions of the Royal Society of London. Series A, Containing Papers of a Mathematical or Physical Character, 231(694-706):289–337, 1933

1933
[28]

Panchenko

D. Panchenko. Free energy in the generalized Sherrington–Kirkpatrick mean field model. Reviews in Mathematical Physics, 17(07):793–857, 2005

2005
[29]

Panchenko

D. Panchenko. Free energy in the Potts spin glass.The Annals of Probability, 46(2):829–864, 2018

2018
[30]

G. Parisi. Infinite number of order parameters for spin-glasses.Physical Review Letters, 43(23):1754, 1979

1979
[31]

G. Parisi. A sequence of approximated solutions to the SK model for spin glasses.Journal of Physics A: Mathematical and General, 13(4):L115–L121, 1980. 49

1980
[32]

Perry, A

A. Perry, A. S. Wein, A. S. Bandeira, and A. Moitra. Optimality and sub-optimality of PCA I: Spiked random matrix models.The Annals of Statistics, 46(5):2416–2451, 2018

2018
[33]

Sherrington and S

D. Sherrington and S. Kirkpatrick. Solvable model of a spin-glass.Physical review letters, 35(26):1792, 1975

1975
[34]

Sherrington and S

D. Sherrington and S. Kirkpatrick. 50 years of spin glass theory.Nature Reviews Physics, 7(10):528–529, 2025

2025
[35]

Talagrand

M. Talagrand. Free energy of the spherical mean field model.Probability theory and related fields, 134(3):339–382, 2006

2006
[36]

Talagrand

M. Talagrand. The Parisi formula.Annals of mathematics, 163(1):221–263, 2006

2006
[37]

Talagrand.Mean field models for spin glasses

M. Talagrand.Mean field models for spin glasses. Volume II: Advanced replica-symmetry and low temperature, volume 55. Springer Science & Business Media, 2011

2011
[38]

A. W. Van der Vaart.Asymptotic statistics, volume 3. Cambridge university press, 2000

2000
[39]

Vershynin.High-dimensional probability

R. Vershynin.High-dimensional probability. Cambridge university press, 2020

2020
[40]

Zdeborov´ a and F

L. Zdeborov´ a and F. Krzakala. Statistical physics of inference: Thresholds and algorithms. Advances in Physics, 65(5):453–552, 2016. 50

2016