pith. sign in

arxiv: 2602.18399 · v2 · submitted 2026-02-20 · ❄️ cond-mat.dis-nn

Overlap locking and non-perturbative effects in spin glasses

Silvio Franz , Giorgio Parisi , Federico Ricci-Tersenghi This is my paper

Pith reviewed 2026-05-15 20:40 UTC · model grok-4.3

classification ❄️ cond-mat.dis-nn
keywords spin glassesoverlap lockingnon-perturbative effectsDyson hierarchical modelfinite-size correctionscritical exponentsmean-field theorysynchronization
0
0 comments X

The pith

Non-perturbative effects lock spin glass overlaps for couplings where 1 ≪ ΔH ≪ N.

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

The paper examines how the order parameter synchronizes when two spin glass systems with independent disorders are coupled at low temperatures. It identifies non-perturbative phenomena that arise specifically in the intermediate-coupling window 1 ≪ ΔH ≪ N. This regime connects the locking effect to finite-size free-energy corrections and to correlation properties in the Dyson hierarchical spin glass, a model that approximates finite-dimensional behavior. The authors derive critical exponents for finite-volume corrections in mean-field theory and for correlation decay in the Dyson model.

Core claim

When two spin glass systems are coupled by a Hamiltonian of strength ΔH, their overlaps lock and become similar. In the intermediate range 1 ≪ ΔH ≪ N, this locking is a non-perturbative phenomenon driven by finite-size free-energy corrections and by the correlations that exist in the Dyson hierarchical spin glass. The work computes the associated critical exponents for finite-volume corrections within mean-field theory and for the spatial decay of correlations in the Dyson hierarchical model.

What carries the argument

Overlap locking in the intermediate-coupling regime (1 ≪ ΔH ≪ N), studied via mean-field theory combined with the Dyson hierarchical spin glass.

Load-bearing premise

The mean-field approach together with the Dyson hierarchical model captures the essential non-perturbative physics of the intermediate-coupling regime in finite-dimensional spin glasses.

What would settle it

A direct numerical check in a finite-dimensional spin glass that measures whether overlap similarity collapses or persists when the coupling strength lies strictly between 1 and a value proportional to system size N.

Figures

Figures reproduced from arXiv: 2602.18399 by Federico Ricci-Tersenghi, Giorgio Parisi, Silvio Franz.

Figure 1
Figure 1. Figure 1: Scaling of the function D(t = z/N, N) according to the theory that predicts the decays D(t, N) ∝ t −3/2 at fixed N and D(z) ∝ z −1/2 . The scaled data in the figure follow the function Cb(x)/x1/2 with x = z/N1/2 . where, for simplicity, we have neglected a logarithm in the behavior for t near zero. The crucial difference from the ferromagnet is that, in this case, the large-z and small-t behaviors do not m… view at source ↗
Figure 2
Figure 2. Figure 2: The normalized χ 2 of the fit in Eq. 18 to the energy corrections in mean-field spin glasses as a function of the exponent ω. Using the data with N ≥ 512 and N ≥ 1024 the analytically predicted value ω = 3/4 is statistically acceptable and preferred to previous estimates. Introducing replicas to average the free energy over the disorder, we obtain precisely the mode-locking free energy, Eq. 6, with z = β 2… view at source ↗
Figure 3
Figure 3. Figure 3: Our prediction for the exponent α(q, ρ) in the one dimensional long range model as function of ρ E(⟨s 2 |q⟩t,N ). For t = N1−ρ this correlation should decay as a power of the distance N, D(N1−ρ , N|q) = A(q)/Nα(q) with an exponent α(q) that depends on q. D(t, N|q) is a proxy for E(⟨q(x)q(y)|q⟩c) on distances |x − y| ∼ N, as explicitly shown in the SI. In the small t region for N → ∞, D(t, N|q) is given by … view at source ↗
Figure 4
Figure 4. Figure 4: The semi-analytic predictions for z 1/2D(z) plotted versus z 1/2 for M = 104 , 105 , 106 , 107 . A constant behaviour at large z of this quantity implies that D(z) ∝ z −1/2 . where ⟨q k⟩ = R dq P(q) q k and P(q) is the average overlap distribution. In principle, one can compute each order of the expansion in z; however, we need a high number of them to extract the large z behavior. This may be possible, bu… view at source ↗
Figure 5
Figure 5. Figure 5: Data for D(t = z/N, N) as a function of z for several system sizes. The dashed line is the predicted power behavior for D(z) = limN→∞ D(z/N, N). In the insert, we show the same data zoomed into the small-z region, where size dependence is nearly absent, and the function D(z) can be measured directly. Neglecting logs, this result implies the t−3/2 divergence of the function B(t) at small argument. In what f… view at source ↗
Figure 6
Figure 6. Figure 6: The semi-analytic predictions for D(z) versus z 1/2 M = 104 , 105 , 106 , 107 . The black line is proportional to z 1/2 0.27 0.28 0.29 0.3 0.31 0.32 0.33 0.34 100 1000 10000 100000 1x106 1x107 z1/2 〈(qL-q R) 2〉 M z1/2=32 z1/2=64 z1/2=128 z1/2=256 [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: The semi-analytic predictions for z 1/2D(z) versus M for z 1/2 = 32, 64, 128, 256. the fit and keep only the points at z ≥ 64, we obtain a better fit, with a = 0.52 ± 0.005. We think that we can safely conclude that a is close to 1/2, and it is certainly different from zero. For large values of z it is not easy to do a precise extrapolation to M = ∞, however, the M dependence of D(z) is weak, as clearly sh… view at source ↗
Figure 8
Figure 8. Figure 8: The semi-analytic results for P (q, z) at z 1/2 = 64 for M = 104 , 105 , 106 , 107 . The horizontal line is at 1/3, i.e., the supposed asymptotic value for P (q). distribution of q is symmetric around the origin. Up to now, we have studied the quantity Z dq ds P(q, s; N) s 2 . [64] There are other quantities that can be studied, notably P(q; N) ≡ Z ds P(q, s) D(t, N|q) ≡ R ds P(q, s; N)s 2 R ds P(q, s; N) … view at source ↗
Figure 10
Figure 10. Figure 10: A tree with ten leaves. The overlap of leave (1) with the leaves (2 · · · 5) is q(x2) and with the leaves (6 · · · 10) is q(x0); the overlaps of leave (2) with the other leaves are the same as leave (1), with the only difference that the overlap with the leave (3) is now q(x4) (taken from (37). with qα,α = qEA. It is easy to see, using the properties of Gaussian variables, that I(z) verifies dI(z) dz = β … view at source ↗
read the original abstract

We study the phenomenon of the locking of the order parameter (or synchronization) in spin glasses at low temperatures. When two systems with independent disorders are coupled, their overlaps become similar. A crucial question is how this effect depends on the strength of the coupling between the two systems. Non-perturbative phenomena are present when $1 \ll \Delta H \ll N$, being $\Delta H$ the coupling Hamiltonian and $N$ the size of the system. In this intermediate-coupling region, the effect is related to finite-size free-energy corrections and to the correlations in the Dyson hierarchical spin glass, a model that mimics the physics of finite-dimensional systems. We study this phenomenon in the mean-field approach, both analytically and numerically, and we finally compute the critical exponents for finite-volume corrections in mean-field theory and for the decay of correlations in the Dyson hierarchical model.

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 studies overlap locking (synchronization of order parameters) in spin glasses when two replicas with independent disorders are coupled by a term of strength ΔH. It focuses on the intermediate regime 1 ≪ ΔH ≪ N, where non-perturbative effects are claimed to arise from finite-size free-energy corrections in mean-field theory and from correlation decay in the Dyson hierarchical spin glass; the authors perform an analytical and numerical mean-field analysis and compute the associated critical exponents for finite-volume corrections and for the decay of correlations.

Significance. If the central mapping holds, the work would supply quantitative exponents that link mean-field finite-size effects to hierarchical-model correlations, offering a concrete route to characterize non-perturbative overlap locking beyond perturbative regimes. The explicit computation of these exponents constitutes a falsifiable prediction that could be checked in larger-scale simulations.

major comments (2)
  1. [Abstract / Dyson hierarchical model section] Abstract and the section introducing the Dyson hierarchical model: the central claim that this construction faithfully captures the non-perturbative physics of the intermediate-coupling regime in finite-dimensional Edwards-Anderson glasses rests on an unverified mimicry assumption. The manuscript computes exponents for correlation decay but does not demonstrate that these exponents control the locking amplitude in a non-hierarchical short-range model, leaving the extrapolation to d=3 untested.
  2. [Mean-field analysis] Mean-field analysis section: the isolation of finite-size free-energy corrections that supposedly drive the locking for 1 ≪ ΔH ≪ N is not shown with explicit equations or numerical protocols that separate these corrections from perturbative contributions; without this separation the non-perturbative interpretation cannot be verified from the reported data.
minor comments (2)
  1. [Introduction] Notation for the coupling strength ΔH and system size N should be introduced with a clear definition in the first paragraph of the main text rather than only in the abstract.
  2. [Numerical results] Figure captions for the numerical mean-field data should state the system sizes used and the number of disorder realizations so that the scaling of finite-size corrections can be assessed directly.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive feedback on our manuscript. We provide point-by-point responses to the major comments below, indicating where revisions have been made.

read point-by-point responses

  1. Referee: [Abstract / Dyson hierarchical model section] Abstract and the section introducing the Dyson hierarchical model: the central claim that this construction faithfully captures the non-perturbative physics of the intermediate-coupling regime in finite-dimensional Edwards-Anderson glasses rests on an unverified mimicry assumption. The manuscript computes exponents for correlation decay but does not demonstrate that these exponents control the locking amplitude in a non-hierarchical short-range model, leaving the extrapolation to d=3 untested.

    Authors: We acknowledge that the Dyson hierarchical model is employed as a controlled proxy whose correlation properties are known from prior work to mimic certain aspects of short-range spin glasses. Our manuscript computes the relevant exponents within this setting but does not include direct simulations of non-hierarchical short-range models in d=3. We have revised the abstract and the Dyson section to state explicitly that the connection to finite-dimensional Edwards-Anderson glasses is conjectural and rests on established mimicry properties, without claiming a verified extrapolation of the locking amplitude to d=3. revision: partial

  2. Referee: [Mean-field analysis] Mean-field analysis section: the isolation of finite-size free-energy corrections that supposedly drive the locking for 1 ≪ ΔH ≪ N is not shown with explicit equations or numerical protocols that separate these corrections from perturbative contributions; without this separation the non-perturbative interpretation cannot be verified from the reported data.

    Authors: We appreciate this observation. In the revised manuscript we have added explicit equations in the Mean-field analysis section that isolate the finite-size free-energy corrections from the perturbative terms, together with a description of the numerical fitting and subtraction protocols used to extract these corrections from the data. These additions allow independent verification of the non-perturbative contribution to the locking. revision: yes

Circularity Check

0 steps flagged

Minor self-citation in adopting Dyson hierarchical model as finite-d mimic

full rationale

The derivation computes critical exponents for finite-volume free-energy corrections in mean-field theory and correlation decay in the Dyson hierarchical model using standard analytical and numerical techniques on those models. These steps are self-contained and do not reduce by construction to fitted inputs or self-referential definitions. The statement that the Dyson model 'mimics the physics of finite-dimensional systems' relies on prior literature (including possible self-citations from the Parisi group), but this is not load-bearing for the exponent calculations themselves and does not force the central results tautologically. No equations or claims exhibit the specific reductions required for higher circularity scores.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The analysis rests on standard mean-field assumptions for spin glasses and the validity of the Dyson hierarchical model as a proxy for finite-dimensional effects; no new free parameters or invented entities are introduced in the abstract.

axioms (2)
  • domain assumption Mean-field theory accurately describes the low-temperature phase of spin glasses for the purpose of overlap statistics.
    Invoked throughout the mean-field approach described in the abstract.
  • domain assumption The Dyson hierarchical spin glass captures the essential correlations and finite-size corrections relevant to the locking phenomenon.
    Used to relate the effect to finite-dimensional physics.

pith-pipeline@v0.9.0 · 5446 in / 1297 out tokens · 20295 ms · 2026-05-15T20:40:23.153774+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

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

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

44 extracted references · 44 canonical work pages

  1. [2]

    The corrections arise from fluctuations in the order parameter around the saddle- point value

    Finite size corrections in mean-field spin glasses In theO(3) (or more generallyO(n)) ferromagnetic model, up to logarithmic terms, the finite-volume corrections are simply proportional to 1/N. The corrections arise from fluctuations in the order parameter around the saddle- point value. The computations are straightforward; they are reported in the SI fo...

    work page doi:10.1073/pnas.xxxxxxxxxx3 2026

  2. [3]

    National Center for HPC, Big Data and Quantum Computing

    Correlations in the Hierarchical Model A long-standing problem in finite-dimensional spin glasses is the determination of the distance dependence of the overlap connected correlation function E(⟨q(x)q(y)|q⟩c) = E(⟨q(x)q(y)|q⟩)−q2 for fixed value of the total overlap q (3, 30, 31). This has a large distance power law behavior E(⟨q(x)q(y)|q⟩c)≃|x−y|−α(q)wit...

    work page doi:10.1073/pnas.xxxxxxxxxx5 2026

  3. [4]

    (Addison-Wesley), (1988)

    G Parisi,Statistical Field Theory. (Addison-Wesley), (1988)

    work page 1988

  4. [5]

    (Oxford university press) Vol

    J Zinn-Justin,Quantum field theory and critical phenomena. (Oxford university press) Vol. 171, (2021)

    work page 2021

  5. [6]

    de Physique Lettres45, 205–210 (1984)

    C de Dominicis, I Kondor, On spin glass fluctuations.J. de Physique Lettres45, 205–210 (1984)

    work page 1984

  6. [7]

    T Temesvari, I Kondor, C De Dominicis, Long-wavelength fluctuations in the ising spin glass. J. Phys. A: Math. Gen.21, L1145 (1988)

    work page 1988

  7. [8]

    S Franz, G Parisi, MA Virasoro, Interfaces and louver critical dimension in a spin glass model. J. de Physique I4, 1657–1667 (1994)

    work page 1994

  8. [9]

    V Astuti, S Franz, G Parisi, New analysis of the free energy cost of interfaces in spin glasses. J. Phys. A: Math. Theor.52, 294001 (2019)

    work page 2019

  9. [10]

    S Franz, T J¨org, G Parisi, Overlap interfaces in hierarchical spin-glass models.J. Stat. Mech. Theory Exp.2009, P02002 (2009)

    work page 2009

  10. [11]

    Reports12, 75 – 199 (1974)

    KG Wilson, J Kogut, The renormalization group and theϵ-expansion.Phys. Reports12, 75 – 199 (1974)

    work page 1974

  11. [12]

    P Bleher, YG Sinai, Critical indices for dyson’s asymptotically-hierarchical models.Commun. Math. Phys.45, 247–278 (1975)

    work page 1975

  12. [13]

    H Katzgraber, AP Y oung, Monte carlo studies of the one-dimensional ising spin glass with power-law interactions.Phys. Rev. B67, 134410 (2003)

    work page 2003

  13. [14]

    L Leuzzi, G Parisi, F Ricci-Tersenghi, JJ Ruiz-Lorenzo, Dilute one-dimensional spin glasses with power law decaying interactions.Phys. Rev. Lett.101, 107203 (2008)

    work page 2008

  14. [15]

    L Leuzzi, G Parisi, F Ricci-Tersenghi, JJ Ruiz-Lorenzo, Ising spin-glass transition in a magnetic field outside the limit of validity of mean-field theory.Phys. Rev. Lett.103, 267201 (2009)

    work page 2009

  15. [16]

    S Jensen, N Read, AP Y oung, Nontrivial maturation metastate-average state in a one-dimensional long-range ising spin glass: Above and below the upper critical range.Phys. Rev. E104, 034105 (2021)

    work page 2021

  16. [17]

    Europhys

    S Franz, G Parisi, MA Virasoro, Ultrametricity in an inhomogeneous simplest spin glass model. Europhys. Lett.17, 5 (1992)

    work page 1992

  17. [18]

    D Panchenko, The free energy in a multi-species sherrington–kirkpatrick model.The Annals Probab.43, 3494—-3513 (2015)

    work page 2015

  18. [19]

    HE Castillo, C Chamon, LF Cugliandolo, MP Kennett, Heterogeneous aging in spin glasses. Phys. Rev. Lett.88, 237201 (2002)

    work page 2002

  19. [20]

    G Parisi, Local overlaps, heterogeneities and the local fluctuation dissipation relations.J. Phys. A: Math. Gen.36, 10773–10789 (2003)

    work page 2003

  20. [21]

    SciPost Phys

    J Kurchan, Time-reparametrization invariances, multithermalization and the parisi scheme. SciPost Phys. Core6, 001 (2023)

    work page 2023

  21. [22]

    (World Scientific), pp

    A Crisanti, S Franz, J Kurchan, A Maiorano, Dynamical mean-field theory and the aging dynamics inSpin Glass Theory and Far Beyond: Replica Symmetry Breaking After 40 Y ears. (World Scientific), pp. 157–186 (2023)

    work page 2023

  22. [23]

    E Bolthausen, AS Sznitman, On Ruelle’s probability cascades and an abstract cavity method. Comm. Math. Phys.197, 247–276 (1998)

    work page 1998

  23. [24]

    S Boettcher, Extremal optimization for sherrington-kirkpatrick spin glasses.The Eur. Phys. J. B-Condensed Matter Complex Syst.46, 501–505 (2005)

    work page 2005

  24. [25]

    M Palassini, Ground-state energy fluctuations in the sherrington–kirkpatrick model.J. Stat. Mech. Theory Exp.2008, P10005 (2008)

    work page 2008

  25. [26]

    S Boettcher, Simulations of ground state fluctuations in mean-field ising spin glasses.J. Stat. Mech. Theory Exp.2010, P07002 (2010)

    work page 2010

  26. [27]

    A Billoire, Numerical estimate of the finite-size corrections to the free energy of the sherrington-kirkpatrick model using guerra-toninelli interpolation.Phys. Rev. B73, 132201 (2006)

    work page 2006

  27. [28]

    T Aspelmeier, A Billoire, E Marinari, MA Moore, Finite-size corrections in the sherrington–kirkpatrick model.J. Phys. A: Math. Theor.41, 324008 (2008)

    work page 2008

  28. [29]

    G Parisi, L Sarra, L Talamanca, Study of longitudinal fluctuations of the sherrington–kirkpatrick model.J. Stat. Mech. Theory Exp.2019, 033302 (2019)

    work page 2019

  29. [30]

    F Guerra, FL Toninelli, The thermodynamic limit in mean field spin glass models.Commun. Math. Phys.230, 71–79 (2002)

    work page 2002

  30. [31]

    D Panchenko, Hierarchical exchangeability of pure states in mean field spin glass models. Probab. Theory Relat. Fields161, 619–650 (2015)

    work page 2015

  31. [32]

    statistical physics162, 1–42 (2016)

    D Panchenko, Structure of finite-rsb asymptotic gibbs measures in the diluted spin glass models.J. statistical physics162, 1–42 (2016)

    work page 2016

  32. [33]

    P Contucci, C Giardin `a, C Giberti, G Parisi, C Vernia, Structure of correlations in three dimensional spin glasses.Phys. Rev. Lett103, 017201 (2009)

    work page 2009

  33. [34]

    R Alvarez Ba˜nos, et al., Nature of the spin-glass phase at experimental length scales.J. Stat. Mech.2010, P06026 (2010)

    work page 2010

  34. [35]

    G Kotliar, PW Anderson, DL Stein, One-dimensional spin-glass model with long-range random interactions.Phys. Rev. B27, 602 (1983)

    work page 1983

  35. [36]

    RA Ba ˜nos, LA Fernandez, V Martin-Mayor, AP Y oung, Correspondence between long-range and short-range spin glasses.Phys. Rev. B86, 134416 (2012)

    work page 2012

  36. [37]

    MC Angelini, G Parisi, F Ricci-Tersenghi, Relations between short-range and long-range ising models.Phys. Rev. E89, 062120 (2014)

    work page 2014

  37. [38]

    de Physique Lettres46, 1037–1043 (1985)

    C De Dominicis, I Kondor, Gaussian propagators for the ising spin glass below tc.J. de Physique Lettres46, 1037–1043 (1985)

    work page 1985

  38. [39]

    G Parisi, P Biscari, Finite-volume corrections to the mean-field solution of the sk model.J. Phys. A: Math. Gen.25, 4787 (1992)

    work page 1992

  39. [40]

    M Aguilar-Janita, et al., Small field chaos in spin glasses: Universal predictions from the ultrametric tree and comparison with numerical simulations.Proc. Natl. Acad. Sci.121, e2404973121 (2024)

    work page 2024

  40. [41]

    R Brunetti, G Parisi, F Ritort, Asymmetric little spin-glass model.Phys. Rev. B46, 5339 (1992)

    work page 1992

  41. [42]

    S Franz, M Leone, Replica bounds for optimization problems and diluted spin systems.J. Stat. Phys.111, 535–564 (2003)

    work page 2003

  42. [43]

    S Franz, M Leone, FL Toninelli, Replica bounds for diluted non-poissonian spin systems.J. Phys. A: Math. Gen.36, 10967 (2003)

    work page 2003

  43. [44]

    M Picco, N Sourlas, On the phase transition of the 3d random field ising model.J. Stat. Mech. Theory Exp.2014, P03019 (2014)

    work page 2014

  44. [45]

    M M ´ezard, G Parisi, N Sourlas, G Toulouse, M Virasoro, Replica symmetry breaking and the nature of the spin glass phase.J. Phys. France45, 843–854 (1984). 8 of 13www.pnas.org/cgi/doi/10.1073/pnas.XXXXXXXXXX pnas.org Supplementary Information Some technical details Finite volume corrections for the O(n)ferromagnetic model In the high-temperature region, ...

    work page doi:10.1073/pnas.xxxxxxxxxx 1984