Pith. sign in

REVIEW 1 major objections 1 minor 65 references

Replica theory for the rate functional of the empirical spectral distribution function of diluted Hermitian matrices

T0 review · 1 major / 1 minor · reviewed 2026-06-26 · grok-4.3

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

Pith's one-line read A replica-based construction under symmetry assumption produces a candidate rate functional for fluctuations of the empirical spectral distribution of diluted Hermitian matrices.

desk verdict Replica construction for the rate functional of spectral fluctuations in sparse Hermitian matrices, but the RS saddle point leaves large-deviation predictions untested. read the letter →

arxiv 2606.17868 v1 pith:QCLDV47U submitted 2026-06-16 cond-mat.dis-nn cond-mat.stat-mech

Edgar Guzmán-González , Isaac Pérez Castillo This is my paper · ORCID
classification cond-mat.dis-nncond-mat.stat-mech
keywords replicamethodratefunctionalempiricalspectraldistributiondilutedHermitianmatricesErdős-Rényigraphscumulantgeneratinglargedeviationslinearstatistics
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 develops a replica framework for the scaled cumulant-generating functional of the empirical spectral distribution i_C of diluted Hermitian random matrices. Within a replica-symmetric saddle-point assumption, the construction yields a candidate rate functional for fluctuations of i_C. Applied to adjacency matrices of Erdős-Rényi graphs with mean degree c, it supplies explicit first and second cumulants of i_C, shows how higher cumulants follow from functional derivatives, and gives the rate function for Fourier coefficients of selected linear spectral statistics. These predictions are compared to exact numerical diagonalization and agree in the accessible regime. The method supplies a route to rate functionals for spectral observables in sparse random-matrix ensembles.

What carries the argument

Replica-symmetric saddle-point approximation to the replica construction of the scaled cumulant-generating functional of the empirical spectral distribution i_C.

What would settle it

Explicit computation of the third cumulant via functional derivative followed by comparison to numerical diagonalization on large Erdős-Rényi graphs; significant mismatch would falsify the candidate rate functional.

Watch

Extended reading notes

Core claim

Within a replica-symmetric saddle-point assumption, the replica construction for the scaled cumulant-generating functional of i_C yields an explicit candidate rate functional whose functional derivatives produce the cumulants and the rate function of linear spectral statistics for diluted Hermitian matrices, including adjacency matrices of Erdős-Rényi graphs.

Load-bearing premise

The saddle point obtained from the replica construction is replica-symmetric.

Editorial extensions

If this is right

  • The first two cumulants of i_C admit closed-form expressions for Erdős-Rényi adjacency matrices.
  • Higher-order cumulants of i_C are obtained by repeated functional differentiation of the rate functional.
  • The rate function for any finite collection of Fourier coefficients of i_C follows directly from the same functional.
  • The construction extends in principle to the study of rate functionals for other spectral observables in sparse Hermitian ensembles.

Reading between the lines

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

  • The same replica construction could be revisited with one-step or full replica-symmetry-breaking ansätze to access regimes where the symmetric saddle point ceases to be stable.
  • The resulting rate functional supplies a concrete starting point for deriving large-deviation principles for eigenvalue statistics of other sparse network models.
  • Direct comparison of the predicted rate function against Monte-Carlo sampling of linear spectral statistics on very large graphs would test the functional beyond the small-fluctuation window accessible by diagonalization.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

1 major / 1 minor

Summary. The paper develops a replica-based framework for the scaled cumulant-generating functional of the empirical spectral distribution function i_C of diluted Hermitian random matrices. Within a replica-symmetric saddle-point assumption, this construction yields a candidate rate functional for fluctuations of i_C. As an illustrative application to adjacency matrices of unweighted Erdős-Rényi random graphs with mean degree c, it derives explicit expressions for the first two cumulants of i_C, indicates how higher cumulants follow from further functional derivatives, computes the rate function of Fourier coefficients (equivalently selected linear spectral statistics), and tests the replica-symmetric predictions against exact numerical diagonalization, finding good agreement in the accessible fluctuation regime.

Significance. If the replica-symmetric saddle-point assumption is valid for the large-deviation regime, the work supplies an analytical route to rate functionals of spectral observables in sparse Hermitian ensembles, a setting where direct large-deviation analysis is otherwise intractable. The explicit first- and second-cumulant expressions and the reduction to linear statistics constitute concrete, usable outputs.

major comments (1)
  1. [Abstract] Abstract: the numerical validation is confined to the small-fluctuation regime accessible by direct diagonalization. Because the central object is a rate functional whose large-deviation predictions are the primary target, agreement only in the perturbative regime does not yet establish the functional outside the regime where the replica-symmetric ansatz is already known to be reliable for the typical density.
minor comments (1)
  1. The repeated use of the qualifier 'candidate' for the rate functional is appropriate and should be retained; it correctly signals that the result remains conditional on the replica-symmetric saddle-point closure.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the careful reading of the manuscript and the positive evaluation of its significance. We respond to the single major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the numerical validation is confined to the small-fluctuation regime accessible by direct diagonalization. Because the central object is a rate functional whose large-deviation predictions are the primary target, agreement only in the perturbative regime does not yet establish the functional outside the regime where the replica-symmetric ansatz is already known to be reliable for the typical density.

    Authors: The abstract already qualifies the numerical comparison by stating that the replica-symmetric predictions 'show good agreement in the accessible fluctuation regime.' We agree that this regime is limited to fluctuations small enough to be sampled by direct diagonalization. The primary contribution of the work is the analytical construction of the candidate rate functional under the replica-symmetric saddle-point assumption; this construction is derived without restriction to the perturbative regime and is intended to furnish large-deviation predictions. Because events in the far tails of the empirical spectral distribution are exponentially rare, direct numerical access to the large-deviation regime for the full distribution is computationally prohibitive for sparse matrices. The agreement obtained where exact comparison is possible therefore provides a non-trivial consistency check on the framework and the ansatz. We do not consider a revision of the abstract necessary, as its current wording accurately reflects both the scope of the validation and the analytical nature of the rate-functional derivation. revision: no

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: derivation conditional on standard replica-symmetric assumption with no self-referential reduction

full rationale

The paper explicitly frames its result as conditional on a replica-symmetric saddle-point assumption applied to the replica construction for the scaled cumulant-generating functional. This is presented as an approximation whose validity is to be checked separately (as confirmed by the numerical tests against diagonalization), not as a quantity derived from or equivalent to its own inputs by construction. No equations reduce a fitted parameter to a 'prediction,' no self-citations are invoked as load-bearing uniqueness theorems, and the central objects (rate functional, cumulants) are obtained via functional derivatives under the stated ansatz rather than by renaming or tautological re-expression of the input data.

Assumptions & free parameters 0 free parameters · 1 assumptions · 0 invented entities

The central claim rests on the replica method and the replica-symmetric saddle-point assumption, which are standard in the field but the specific application to rate functionals is the contribution.

assumptions (1)
  • domain assumption Replica-symmetric saddle-point assumption
    Invoked to obtain the rate functional from the replica construction for fluctuations of i_C.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Replica theory for the rate functional of the empirical spectral distribution function of diluted Hermitian matrices." pith.science (2026). https://pith.science/paper/QCLDV47U

@misc{pith2026260617868,
  author       = {Pith},
  title        = {Pith review of: Replica theory for the rate functional of the empirical spectral distribution function of diluted Hermitian matrices},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QCLDV47U}},
  note         = {Machine review of arXiv:2606.17868}
}
abstract

We develop a replica-based framework for the scaled cumulant-generating functional of the empirical spectral distribution function $i_C$ of diluted Hermitian random matrices. Within a replica-symmetric saddle-point assumption, this construction yields a candidate rate functional for fluctuations of $i_C$. As an illustrative application, we consider adjacency matrices of unweighted Erd\H{o}s-R\'enyi random graphs with mean degree $c$. We derive explicit expressions for the first two cumulants of $i_C$, indicate how higher cumulants can be obtained from further functional derivatives, and compute the rate function of Fourier coefficients, equivalently of selected linear spectral statistics. The replica-symmetric predictions are tested against exact numerical diagonalization and show good agreement in the accessible fluctuation regime. The approach provides a basis for studying rate functionals of spectral observables in sparse random matrix ensembles.

Figures

Figures reproduced from arXiv: 2606.17868 by the authors.

Figure 1
Figure 1. FIG. 1. Mean empirical spectral distribution function [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Scaled connected covariance [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Candidate rate function [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

65 extracted references · 5 canonical work pages

  1. [1]

    Meanempiricalspectraldistributionfunctionκ 1(x) = ⟨iC (x)⟩for Erdős–Rényi adjacency matrices with mean con- nectivityc= 2,3,5

    8 1 −6 −4 −2 0 2 4 6 c=2 c=3 c=5 κ 1 x FIG.1. Meanempiricalspectraldistributionfunctionκ 1(x) = ⟨iC (x)⟩for Erdős–Rényi adjacency matrices with mean con- nectivityc= 2,3,5. The solid curves are obtained from the population-dynamics solution of Eq. (13), while the markers are exact diagonalization results forN×Nrandom matrices. The theoretical curves use t...

  2. [2]

    015 y0 = 1

    03 −6 −4 −2 0 2 4 6 y0 = 1. 015 y0 = 1. 995 y0 = 2. 835 N κ2(x, y 0) x 0

  3. [3]

    06 −6 −4 −2 0 2 4 6 N κ2(x, x ) x FIG. 2. Scaled connected covarianceN κ2(x, y)of the empir- ical spectral distribution function for Erdős–Rényi adjacency matrices with mean connectivityc= 2. The top panel shows a density plot of the theoretical population-dynamics result obtained from Eq. (13). The middle panel shows cuts at the dashed horizontal lines i...

  4. [4]

    08 g ψ(g + ⟨gC⟩) 100 200 5003. 52

  5. [5]

    54 N ⟨gC⟩ FIG. 3. Candidate rate functionψfor the Fourier projection gC = R a −a dxsin(πx/a)i C (x), witha= 7, for Erdős–Rényi adjacency matrices with mean connectivityc= 5. The solid curve is the replica-symmetric prediction. The horizontal co- ordinate is shifted by the mean,g=u− ⟨g C ⟩, so that the minimum is plotted atg= 0, and the normalization is ch...

  6. [6]

    Wishart, The generalised product moment distribu- tion in samples from a normal multivariate population, Biometrika20A, 32 (1928)

    J. Wishart, The generalised product moment distribu- tion in samples from a normal multivariate population, Biometrika20A, 32 (1928)

    1928

  7. [7]

    E. P. Wigner, Characteristic vectors of bordered matrices with infinite dimensions, Ann. Math.62, 548 (1955)

    1955

  8. [8]

    F. J. Dyson, Statistical theory of the energy levels of complex systems. I, J. Math. Phys.3, 140 (1962)

    1962

Show all 65 references
  1. [9]

    Akemann, J

    G. Akemann, J. Baik, and P. Di Francesco, eds.,The Ox- ford Handbook of Random Matrix Theory(Oxford Uni- versity Press, Oxford, 2011)

    2011

  2. [10]

    L. A. Pastur, On the spectrum of random matrices, Theor. Math. Phys.10, 67 (1972)

    1972

  3. [11]

    M. L. Mehta,Random Matrices, 3rd ed. (Elsevier, 2004)

    2004

  4. [12]

    P. J. Forrester,Log-Gases and Random Matrices, London Mathematical Society Monographs, Vol. 34 (Princeton University Press, 2010)

    2010

  5. [13]

    Erdős and H.-T

    L. Erdős and H.-T. Yau, Universality of local spectral statistics of random matrices, Bull. Am. Math. Soc.49, 377 (2012)

    2012

  6. [14]

    Tao,Topics in Random Matrix Theory, Graduate Studies in Mathematics, Vol

    T. Tao,Topics in Random Matrix Theory, Graduate Studies in Mathematics, Vol. 132 (American Mathemat- ical Society, 2012)

    2012

  7. [15]

    van Mieghem,Graph Spectra for Complex Networks (Cambridge University Press, 2010)

    P. van Mieghem,Graph Spectra for Complex Networks (Cambridge University Press, 2010)

    2010

  8. [16]

    G.BiroliandR.Monasson,Asingledefectapproximation for localized states on random lattices, J. Phys. A: Math. Gen.32, L255 (1999)

    1999

  9. [17]

    G. J. Rodgers andA. J. Bray, Density of statesof asparse random matrix, Phys. Rev. B37, 3557 (1988)

    1988

  10. [18]

    Bauer and O

    M. Bauer and O. Golinelli, Random incidence matrices: Moments of the spectral density, J. Stat. Phys.103, 301 (2001)

    2001

  11. [19]

    Khorunzhy, M

    O. Khorunzhy, M. Shcherbina, and V. Vengerovsky, Eigenvalue distribution of large weighted random graphs, J. Math. Phys.45, 1648 (2004)

    2004

  12. [20]

    Semerjian and L

    G. Semerjian and L. F. Cugliandolo, Sparse random ma- trices: the eigenvalue spectrum revisited, J. Phys. A: Math. Gen.35, 4837 (2002). 7

    2002

  13. [21]

    Abou-Chacra, D

    R. Abou-Chacra, D. J. Thouless, and P. W. Anderson, A selfconsistent theory of localization, J. Phys. C: Solid State Phys.6, 1734 (1973)

    1973

  14. [22]

    Fyodorov and A

    Y. Fyodorov and A. Mirlin, Localization in ensemble of sparse random matrices, Phys. Rev. Lett.67, 2049 (1991)

  15. [23]

    A. D. Mirlin, Y. V. Fyodorov, F.-M. Dittes, J. Quezada, and T. H. Seligman, Transition from localized to ex- tended eigenstates in the ensemble of power-law random banded matrices, Phys. Rev. E54, 3221 (1996)

    1996

  16. [24]

    Golinelli, Statistics of delta peaks in the spectral density of large random trees, arXiv:cond-mat/0301437 (2003), preprint

    O. Golinelli, Statistics of delta peaks in the spectral density of large random trees, arXiv:cond-mat/0301437 (2003), preprint

    2003 arXiv

  17. [25]

    Benaych-Georges, C

    F. Benaych-Georges, C. Bordenave, and A. Knowles, Spectral radii of sparse random matrices, Ann. Inst. Henri Poincaré Probab. Stat.56, 2141 (2020)

    2020

  18. [26]

    J. O. Lee and K. Schnelli, Local law and Tracy–Widom limit for sparse random matrices, Probab. Theory Relat. Fields171, 543 (2018)

    2018

  19. [27]

    Maritan, Predicting the stability of large structured food webs, Nat

    S.Allesina, J.Grilli, G.Barabás, S.Tang, J.Aljadeff,and A. Maritan, Predicting the stability of large structured food webs, Nat. Commun.6, 7842 (2015)

    2015

  20. [28]

    Grilli, T

    J. Grilli, T. Rogers, and S. Allesina, Modularity and sta- bility in ecological communities, Nat. Commun.7, 12031 (2016)

    2016

  21. [29]

    Gibbs, J

    T. Gibbs, J. Grilli, T. Rogers, and S. Allesina, Effect of population abundances on the stability of large random ecosystems, Phys. Rev. E98, 022410 (2018)

    2018

  22. [30]

    K.RajanandL.F.Abbott,Eigenvaluespectraofrandom matrices for neural networks, Phys. Rev. Lett.97, 188104 (2006)

    2006

  23. [31]

    Ahmadian, F

    Y. Ahmadian, F. Fumarola, and K. D. Miller, Proper- ties of networks with partially structured and partially random connectivity, Phys. Rev. E91, 012820 (2015)

    2015

  24. [32]

    Diaconis and S

    P. Diaconis and S. N. Evans, Linear functionals of eigen- values of random matrices, Trans. Am. Math. Soc.353, 2615 (2001)

    2001

  25. [33]

    D. S. Dean and S. N. Majumdar, Large deviations of extreme eigenvalues of random matrices, Phys. Rev. Lett. 97, 160201 (2006)

    2006

  26. [34]

    S. N. Majumdar and G. Schehr, Top eigenvalue of a ran- dom matrix: large deviations and third order phase tran- sition, J. Stat. Mech.: Theory Exp.2014(1), P01012

    2014

  27. [35]

    Ben Arous and A

    G. Ben Arous and A. Guionnet, Large deviations for Wigner’s law and Voiculescu’s non-commutative entropy, Probab. Theory Relat. Fields108, 517 (1997)

    1997

  28. [36]

    Hiai and D

    F. Hiai and D. Petz, A large deviation theorem for the empirical eigenvalue distribution of random unitary ma- trices, Ann. Inst. Henri Poincaré Probab. Stat.36, 71 (2000)

    2000

  29. [37]

    B. B. Bhattacharya, S. Bhattacharya, and S. Ganguly, Spectral edge in sparse random graphs: Upper and lower tail large deviations, Ann. Probab.49, 1847 (2021)

    2021

  30. [38]

    Ganguly, E

    S. Ganguly, E. Hiesmayr, and K. Nam, Spectral large deviations of sparse random matrices, J. London Math. Soc.110, e12954 (2024), arXiv:2206.06954 [math.PR]

    2024

  31. [39]

    F.Augeri,Largedeviationsoftheempiricalspectralmea- sure of supercritical sparse Wigner matrices, Adv. Math. 466, 110156 (2025), arXiv:2401.11925 [math.PR]

    2025

  32. [40]

    Zhu and Y

    X. Zhu and Y. Zhu, Central limit theorems for linear spectral statistics of inhomogeneous random graphs with graphon limits (2024), arXiv:2412.19352 [math.PR]

    2024

  33. [41]

    S. F. Edwards and R. C. Jones, The eigenvalue spectrum of a large symmetric random matrix, J. Phys. A: Math. Gen.9, 1595 (1976)

    1976

  34. [42]

    Mézard, G

    M. Mézard, G. Parisi, and M. A. Virasoro,Spin Glass Theory and Beyond, World Scientific Lecture Notes in Physics, Vol. 9 (World Scientific, Singapore, 1987)

    1987

  35. [43]

    Mézard and A

    M. Mézard and A. Montanari,Information, Physics, and Computation(Oxford University Press, 2009)

    2009

  36. [44]

    Mézard and G

    M. Mézard and G. Parisi, The Bethe lattice spin glass revisited, Eur. Phys. J. B20, 217 (2001)

    2001

  37. [45]

    Kühn, Spectra of sparse random matrices, J

    R. Kühn, Spectra of sparse random matrices, J. Phys. A: Math. Theor.41, 295002 (2008)

    2008

  38. [46]

    Rogers, I

    T. Rogers, I. Pérez Castillo, R. Kühn, and K. Takeda, Cavity approach to the spectral density of sparse sym- metric random matrices, Phys. Rev. E78, 031116 (2008)

    2008

  39. [47]

    V. A. R. Susca, P. Vivo, and R. Kühn, Cavity and replica methods for the spectral density of sparse sym- metric random matrices, SciPost Phys. Lect. Notes33, 10.21468/SciPostPhysLectNotes.33 (2021)

    2021 doi

  40. [48]

    Bordenave and M

    C. Bordenave and M. Lelarge, Resolvent of large random graphs,RandomStructures&Algorithms37,332(2010)

    2010

  41. [49]

    K. B. Efetov,Supersymmetry in Disorder and Chaos (Cambridge University Press, Cambridge, 1996)

    1996

  42. [50]

    F. L. Metz and I. Pérez Castillo, Large deviation function for the number of eigenvalues of sparse random graphs inside an interval, Phys. Rev. Lett.117, 104101 (2016)

    2016

  43. [51]

    Pérez Castillo and F

    I. Pérez Castillo and F. L. Metz, Large-deviation theory for diluted Wishart random matrices, Phys. Rev. E97, 032124 (2018)

    2018

  44. [52]

    Dembo and O

    A. Dembo and O. Zeitouni,Large Deviations Techniques and Applications, 2nd ed. (Springer, Berlin, Heidelberg, 2010)

    2010

  45. [53]

    Touchette, The large deviation approach to statistical mechanics, Phys

    H. Touchette, The large deviation approach to statistical mechanics, Phys. Rep.478, 1 (2009)

    2009

  46. [54]

    S. F. Edwards and P. W. Anderson, Theory of spin glasses, J. Phys. F: Met. Phys.5, 965 (1975)

    1975

  47. [55]

    Debrecen6, 290 (1959)

    P.ErdősandA.Rényi,Onrandomgraphs.I,Publ.Math. Debrecen6, 290 (1959)

    1959

  48. [56]

    M.Newman,Networks,2nded.(OxfordUniversityPress, 2018)

    2018

  49. [57]

    Guzmán-González and I

    E. Guzmán-González and I. Pérez Castillo, Supplemental material (2026), see the Supplemental Material for de- tails on the computation of the rate functional using the replica method, explicit expressions for the first and sec- ond cumulants and the cumulant-generating functio...

    2026

  50. [58]

    T. Guhr, A. Müller-Groeling, and H. A. Weidenmüller, Random-matrix theories in quantum physics: common concepts, Phys. Rep.299, 189 (1998)

    1998

  51. [59]

    Pérez Castillo and F

    I. Pérez Castillo and F. L. Metz, Theory for the condi- tioned spectral density of noninvariant random matrices, Phys. Rev. E98, 020102 (2018)

    2018

  52. [60]

    A. K. Hartmann, Sampling rare events: Statistics of local sequence alignments, Phys. Rev. E65, 056102 (2002)

    2002

  53. [61]

    Replica theory for the rate functional of the empirical spectral distribution function of diluted Hermitian matrices

    A. K. Hartmann, Large-deviation properties of largest component for random graphs, Eur. Phys. J. B84, 627 (2011). S1 Supplemental Material for “Replica theory for the rate functional of the empirical spectral distribution function of diluted Hermitian matrices” E. Guzmán-Gonzá...

    2011

  54. [62]

    Details of the procedure used to compute the rate functional ofi C using the replica method

  55. [63]

    Explicit expressions for the first and second cumu- lants ofi C and for its cumulant-generating func- tional

  56. [64]

    A stationarity argument showing when the implicit dependence of the functionalwonµcan be omitted in differentiatingF[µ], with applications toκ1,κ 2, andf ′(β)

  57. [65]

    n+ lY a=1 δ(yla −y laj) #

    An explanation of how the rate functional can be used to obtain the candidate rate function of the Fourier coefficients ofiC. S1. USING THE REPLICA METHOD TO COMPUTE THE RA TE FUNCTIONAL In this section, we present the details to obtain the rate functional using the replica me...

Pith tools

Reviewed June 26, 2026 · model on record in the stance chip above.