REVIEW 4 major objections 4 minor 26 references
Quantum chaos, localization and phase transitions in random graphs
T0 review · 4 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash
Pith's one-line read Uniform random graphs, treated as quantum tight-binding lattices, undergo a phase transition from chaotic, extended states to localized, integrable states as the edge-to-vertex ratio is lowered, with a semi-Poisson critical regime in…
desk verdict Plausible numerical crossover, but the phase-transition claim outruns the data: no scaling analysis, no error bars, and the critical ratios are hand-picked. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central object is the uniform random graph $G(n,m)$ with $m$ edges on $n$ vertices, whose adjacency matrix serves directly as the tight-binding Hamiltonian. The control parameter is the ratio $R=m/n$, which acts as the disorder strength: it sets how many hopping paths a wavefunction can take and, through the giant-component transition at $R=0.5$, determines whether a macroscopic connected lattice exists at all. The diagnostic is the unfolded level-spacing distribution $P(S)$, compared against the Wigner-Dyson, Poisson, and semi-Poisson forms, supplemented by the inverse participation ratio for the wavefunctions. The argument's load-bearing link is the relation between $R$ and the emergent spatial dimension $D$, computed by the scaling method of the authors' earlier work, so that the metallic-to-localized transition is recast as a dimensional crossover from $D\approx 3$ to $D\approx 2$.
What would settle it
A decisive check is to measure the spacing distribution and the inverse participation ratio scaling at a fixed critical ratio, say R=0.82 in the window 0.1<E<0.2, as the graph size doubles to n=36000 and n=72000: true criticality requires the semi-Poisson form to persist and the inverse participation ratio to scale with a nontrivial fractal exponent, while drift toward the Wigner or Poisson form (or IPR scaling like 1/n) would falsify the phase-transition interpretation.
Extended reading notes
Core claim
The central discovery is that the level-spacing statistics of uniform random graphs realize the three classic regimes of the Anderson transition, controlled not by a random potential but by the graph's edge-to-vertex ratio $R=m/n$. For a fixed energy window near the band center, dense graphs ($R\approx 1$–$2$) yield Wigner-Dyson spacing $P(s)=(\pi/2)s\,e^{-\pi s^2/4}$, the signature of chaotic, extended wavefunctions; sparse graphs near the structural transition $R=0.5$ yield Poisson spacing $P(s)=e^{-s}$, the signature of localized, integrable wavefunctions; intermediate ratios (e.g., $R\approx 0.8$ at $0.1<E<0.2$) yield semi-Poisson $P(s)=4s\,e^{-2s}$, which the paper reads as the critical point of a phase transition. The critical ratio depends on energy: in the window $0.5<E<0.6$ the semi-Poisson distribution appears at $R_c\approx 0.66$. The inverse participation ratio confirms that wavefunctions become more localized as $R$ decreases, and the authors connect the transition to a drop in the emergent spatial dimension from about 3.15 at $R=1$ to 1.90 at $R=0.6$.
Load-bearing premise
The load-bearing premise is that the visually identified Wigner, Poisson, and semi-Poisson spacing distributions at only two graph sizes (n=9000 and n=18000) represent distinct thermodynamic phases in the infinite-size limit, rather than a smooth finite-size crossover of a single phase.
Editorial extensions
If this is right
- If the transition is a genuine phase transition, uniform random graphs provide a minimal model in which disorder is purely geometric: no on-site potentials or magnetic fields are needed to produce Anderson-like localization.
- The energy-dependent critical ratio $R_c(E)$ implies a mobility edge in the $(E,R)$ plane, so a complete phase diagram could be mapped by spectral window.
- The appearance of semi-Poisson statistics identifies the critical regime with the universality class of the conventional Anderson transition, suggesting common critical behavior despite the absence of a fixed spatial dimension.
- The dimensional-crossing interpretation predicts that other random geometries with tunable connectivity will show analogous Wigner-to-Poisson transitions.
- Wavefunction localization in graphs could be directly probed via transport or conductance calculations, connecting spectral statistics to physical observables.
Reading between the lines
- A testable consequence beyond the paper: random regular graphs with fixed degree should show the same Wigner-to-Poisson transition as the degree is lowered, which would show that connectivity density, not degree fluctuations, drives the transition.
- One step further, one could measure the multifractal spectrum of the wavefunctions at the claimed critical ratios; genuine criticality would give a nontrivial spectrum of inverse participation ratio exponents, and a trivial one would instead indicate a finite-size crossover.
- The energy dependence of the critical ratio suggests that the full spectral phase diagram in the $(E,R)$ plane could have a mobility-edge structure worth mapping explicitly, a prediction the paper leaves implicit.
- If the semi-Poisson regime is truly critical, it should appear at a well-defined critical value of $R$ in the thermodynamic limit; testing the drift of $R_c$ with $n$ would separate a true transition from a crossover.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript studies the energy-level statistics of uniform random graphs G(n,m) treated as tight-binding Hamiltonians. For several edge-to-vertex ratios R and graph sizes n=9000 and n=18000, the authors compute the density of states, the unfolded level-spacing distribution P(S), level crossings, and the inverse participation ratio. They report Wigner-Dyson statistics for dense graphs, Poisson statistics near R≈0.5, and semi-Poisson statistics at intermediate ratios, and they interpret these regimes as metallic, localized, and critical phases of a quantum phase transition whose critical ratio depends on energy.
Significance. If established, the result would connect random-graph geometry with quantum-chaos universality and would be of interest to both the quantum-chaos and network-science communities. The paper has clear strengths: the model is simple and well defined, the P(S) data are obtained by direct numerical unfolding over 1000 configurations rather than by fitting to the claimed analytic curves, and the IPR provides an independent look at wavefunction localization. However, the paper's central claim of a genuine thermodynamic phase transition is not supported by the presented evidence; the data are consistent with a finite-size crossover. The missing finite-size scaling and order-parameter analysis are in principle obtainable, so I regard the paper as requiring major revision rather than being irreparable.
major comments (4)
- [Fig. 2 and surrounding text] The central claim of a phase transition rests entirely on the visual agreement of P(S) with the Wigner, Poisson, and semi-Poisson curves at n=9000 and n=18000. No goodness-of-fit test, no error bars, and no finite-size scaling or scaling collapse are reported. The statement that 'the Wigner distribution remains invariant under scaling as shown from the two cases n=9000 and n=18000' is not a demonstration of scale invariance: agreement at two system sizes is also the expected behavior of a finite-size crossover in a model with a large but finite localization length. The same data could be obtained in a single phase, so the phase-transition interpretation is not established.
- [Discussion after Eq. (4), specifically Rc=0.66] The identification of the critical ratio Rc=0.66 for the window 0.5<E<0.6 is not based on any stated criterion. The text says that 'a phase transition ... at the critical ratio Rc=0.66' is found because the semi-Poisson distribution appears, but 'appears' is not defined. The values of Rc and the energy windows are therefore free parameters of the analysis, and the claim that Rc depends on energy is not quantitative. A defined order parameter with a crossing criterion, e.g., the integrated level-spacing ratio or the scaling of the IPR, is needed.
- [Fig. 4 and IPR discussion] The IPR data are presented for a single system size (n=9000) without error bars and without finite-size scaling. The text itself characterizes the behavior as an 'onset to localization' as R decreases, which is a crossover statement rather than evidence for distinct thermodynamic phases. In particular, the IPR curves in Fig. 4 do not distinguish localized wavefunctions in the thermodynamic limit from finite-size localization precursors, so they cannot support the phase-transition claim.
- [Emergent-dimension explanation near the conclusion] The explanation that the transition is driven by the emergent spatial dimension D relies on average values ⟨D⟩=3.15, 2.64, and 1.90 for three values of R, but the method for computing D is not described in this manuscript and no direct quantitative relation between D and the observed P(S) or IPR is established. This is a speculative interpretive step, not evidence for the phase transition, and it should be presented as such or omitted from the main claims.
minor comments (4)
- [Title and text typos] The title contains a typo: 'rando m graphs' should be 'random graphs'. There are also several typographical errors in the text, including 'Furhter', 'wavefucntion', and 'occuring', which should be corrected.
- [Fig. 2d] In Fig. 2d the panel includes a curve for n=18000-R=1 alongside the R=0.66 and R=0.55 curves, but the role of this reference curve is not explained in the caption or text.
- [Fig. 3] The vertical axis of Fig. 3 appears to lack tick labels in the printed figure, and the caption does not state how many levels are shown or how the unfolding was performed for this plot.
- [References [8,9]] The emergent-dimension scaling method is cited to the authors' own previous work but is not summarized; a brief description of the method would make the paper more self-contained.
Circularity Check
No significant circularity: level-spacing distributions are direct numerical outputs compared to independent analytic benchmarks; the self-cited dimension calculation is only an ancillary explanation.
full rationale
The paper's core observation is the numerically computed level-spacing distribution P(S) for G(n,m) random graphs, unfolded from the spectrum and compared directly to the fixed analytic Wigner-Dyson (Eq. 2), Poisson (Eq. 3), and semi-Poisson (Eq. 4) distributions. These comparisons are not fits: the functional forms and parameters are set by the standard random-matrix/Anderson-transition benchmarks, and no parameter of the model is adjusted to force the match. The phase labels ('chaotic', 'localized', 'critical') are operational names for which benchmark P(S) is closest, so the observation that the statistics change with R is not circular. The only self-citations that could raise concern are Refs. [8,9], used to quote emergent-dimension values (<D>=3.15, 2.64, 1.90) in the final 'partial explanation' paragraph; that paragraph explicitly frames the dimension argument as 'could provide a partial explanation', and it is not the derivation of the transition. The numerical result stands independently of the self-cited dimension calculation. Lack of finite-size scaling or error bars is an evidence/correctness limitation, not a circularity.
Assumptions & free parameters
free parameters (2)
- Critical ratio Rc =
0.82 for 0.1<E<0.2; 0.66 for 0.5<E<0.6; also 0.8 shown as critical
- Energy windows =
0.1<E<0.2 and 0.5<E<0.6
assumptions (4)
- domain assumption The adjacency matrix (Eq. 1) with unit hopping on a uniform random graph G(n,m) captures the essential quantum physics of random-geometry systems.
- domain assumption For R>0.5 the giant component dominates the spectral statistics; smaller components can be ignored.
- ad hoc to paper The semi-Poisson distribution identifies the critical point of the transition.
- ad hoc to paper The emergent dimension D from Refs. [8,9] controls the localization behavior.
Cite this review
Pith. "Pith review of Quantum chaos, localization and phase transitions in random graphs." pith.science (2026). https://pith.science/paper/S2IXCIBN
@misc{pith2026241214722,
author = {Pith},
title = {Pith review of: Quantum chaos, localization and phase transitions in random graphs},
year = {2026},
howpublished = {\url{https://pith.science/paper/S2IXCIBN}},
note = {Machine review of arXiv:2412.14722}
}
read the original abstract
The energy level statistics of uniform random graphs are studied, by treating the graphs as random tight-binding lattices. The inherent random geometry of the graphs and their dynamical spatial dimensionality, leads to various quantum chaotic and localized phases and transitions between them. Essentially the random geometry acts as disorder, whose strength is characterized by the ratio of edges over vertices R in the graphs. For dense graphs, with large ratio R, the spacing between successive energy levels follows the Wigner-Dyson distribution, leading to a quantum chaotic behavior and a metallic phase, characterized by level repulsion. For ratios near R=0.5, where a large dominating component in the graph appears, the level spacing follows the Poisson distribution with level crossings and a localized phase for the respective wavefunctions lying on the graph. For intermediate ratios R we observe a phase transition between the quantum chaotic and localized phases characterized by a semi-Poisson distribution. The values R of the critical regime where the phase transition occurs depend on the energy of the system. Our analysis shows that physical systems with random geometry, for example ones with a fluctuating/dynamical spatial dimension, contain novel universal phase transition properties, similar to those occuring in more traditional phase transitions based on symmetry breaking mechanisms, whose universal properties are strongly determined by the dimensionality of the system.
Figures
Reference graph
Works this paper leans on
-
[1]
P . Erd˝ os, T. Gallai, ”Gr´ afokel˝ o´ ırt foksz´ am´ u pont okkal”, Matematikai Lapok, 11: 264-274 (1960)
work page 1960
-
[2]
Aigner, E
M. Aigner, E. Triesch, Discrete Math. 136, 3-20 (1994)
1994
-
[3]
I. J. Farkas, I. Der´ enyi, A.-L. Barab´ asi, and T. Vicsek , Phys. Rev. E 64, 026704 (2001)
work page 2001
-
[4]
M. E. J. Newman Networks: An Introduction Oxford Univer- sity Press, Oxford (2010)
2010
- [5]
-
[6]
Berg and M
J. Berg and M. Lassig., Phys. Rev. Lett. 89, 228701 (2002)
2002
-
[7]
S. Mizutaka and T. Hasegawa, Journal of Physics: Complex ity 1, 035007 (2020)
work page 2020
-
[8]
Kleftogiannis and I
I. Kleftogiannis and I. Amanatidis, Phys. Rev. E 105, 024141 (2022)
2022
Show all 26 references
- [9]
-
[10]
S N Evangelou and T Ziman, J. Phys. C: Solid State Phys. 20 L235 (1987)
1987
-
[11]
Evangelou, Phys
S.N. Evangelou, Phys. Rev. B 49, 16805 (1994)
1994
-
[12]
S. N. Evangelou and D. E. Katsanos, J. Stat. Phys. 85, 525-550 (1996)
1996
-
[13]
Mirlin Phys
A.D. Mirlin Phys. Rep. 326, 260 (2000)
2000
-
[14]
Amanatidis, I
E. Amanatidis, I. Kleftogiannis, D. E. Katsanos, and S. N. Evangelou, J. Phys. Condens. Matter 26, 15 (2014)
2014
-
[15]
Amanatidis and S
I. Amanatidis and S. N. Evangelou Phys. Rev. B 79, 205420 (2009)
2009
-
[16]
M. V . Berry, Proc. R. Soc. A 413, 183 (1987)
1987
-
[17]
Berry, New Scientist 44, 47 (1987)
M. Berry, New Scientist 44, 47 (1987)
1987
-
[18]
St¨ ockmann and J
H.J. St¨ ockmann and J. Stein, Phys. Rev. Lett. 64, 2215 (1990)
1990
-
[19]
Courtney, H
M. Courtney, H. Jiao, N. Spellmeyer, D. Kleppner, J. Gao , and J. B. Delos, Phys. Rev. Lett. 74, 1538 (1995)
1995
-
[20]
Nakamura and T
K. Nakamura and T. Harayama, Quantum Chaos and Quantum Dots (Oxford University Press, Oxford, UK) (2004)
2004
-
[21]
Kottos and U
T. Kottos and U. Smilansky Phys. Rev. Lett. 79, 4794 (1997)
1997
-
[22]
Kottos, U
T. Kottos, U. Smilansky, Phys. Rev. Lett. 85, 968 (2000)
2000
-
[23]
Haake, Quantum Signatures of Chaos (Springer), (201 0)
F. Haake, Quantum Signatures of Chaos (Springer), (201 0)
-
[24]
Smilansky, Chaos
U. Smilansky, Chaos. Prog. Math. Phys. Birkh¨ auser, Ba sel 66, 97-124 (2013)
2013
-
[25]
Grabarits, K
A. Grabarits, K. R. Swain, M. S. Heydari, P . Chandarana, F. J. G´ omez-Ruiz, and A. del Campo .arXiv:2405.14376v1 (2024)
2024 arXiv
-
[26]
Pournaki, E
A. Pournaki, E. Olbrich, S. Banisch and K. Klemm, Phys. R ev. E 107, 054112 (2023)
2023
Reviewed August 11, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.