Pith. sign in

REVIEW 3 major objections 5 minor 29 references

Emergence of continuously varying critical exponents in coupled map lattice as an effect of quenched disorder

T0 review · 3 major / 5 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read A coupled map lattice with quenched asymmetric couplings shows a phase transition whose critical exponents change continuously with the disorder fraction p, instead of staying in the directed percolation class.

desk verdict A plausible but under-supported claim of continuously varying critical exponents; the central issue is whether the chosen ϵ_c is a true critical point or just a point inside a Griffiths phase. read the letter →

arxiv 2509.07529 v1 pith:3TON7X4J submitted 2025-09-09 cond-mat.stat-mech physics.comp-ph

Priyanka D. Bhoyar , Govindan Rangarajan , Prashant M. gade This is my paper · ORCID
classification cond-mat.stat-mechphysics.comp-ph PACS 05.45.Ra05.70.Jk05.90.+m
keywords coupledmaplatticeabsorbingphasetransitiondirectedpercolationquencheddisordercriticalexponentsuniversalityclassasymmetriccouplingeigenvaluespectrum
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 studies a one-dimensional lattice of coupled circle maps at the onset of spatiotemporal intermittency, a system known to sit in the directed percolation universality class. It adds quenched spatial disorder by making each directed coupling asymmetric: a fraction p of sites couple more strongly to the right neighbor, the rest more strongly to the left. For any p > 0, the transition to the absorbing laminar state still shows power-law decay of the turbulent-site fraction at a critical coupling, but the decay exponent δ, the order-parameter exponent β, and the correlation-time exponent ν∥ change continuously with p. The paper's claim is that this is a genuinely new, continuously parametrized set of critical exponents—not a known universality class, not weak universality, and not a Griffiths phase—and it suggests the eigenvalue spectrum of the random connectivity matrix carries the mechanism.

What carries the argument

The central object is a random connectivity matrix that encodes which neighbor receives the stronger coupling. At each site, the map's output goes to neighbors with weights ε1=ε/2+0.1 and ε2=ε/2−0.1; a site is type A (stronger to the left) with probability p, otherwise type B (stronger to the right), and the assignment is frozen in time. The order parameter is m(t), the fraction of sites with x_i(t)>0.5. Critical exponents are read from the power-law decay m(t)∼t^{−δ} at the critical coupling ε_c, the steady-state scaling m(∞)∼(ε−ε_c)^β, and data-collapse off-critical/finite-size scalings that yield ν∥ and z. The paper uses the eigenvalue spectrum of the connectivity matrix as the diagnostic

What would settle it

Perform a seed-spreading analysis for fixed p (e.g., p=0.2 and p=0.4) on much larger lattices and longer times, and locate the absorbing transition independently from the long-time exponential decay rate or from the crossing of a cumulant. Then re-extract δ, β, and ν∥ at that independently located ε_c. If the exponents come out p-independent, or if the independently located ε_c systematically differs from the power-law-fit value and grows with p, the continuous exponent variation is a finite-window artifact.

Watch

Extended reading notes

Core claim

The central claim: frozen, randomly assigned asymmetric couplings replace the directed-percolation universality of an absorbing phase transition with a one-parameter family of critical exponents. At p=0 the model gives DP values; for p=0.1–0.5, δ runs from 0.034 to 0.158, β from 0.125 to 0.607, ν∥ from 2.47 to 3.84, and z=3.90 at p=0.5. All change continuously with p; except δ at p=0.5, none match a known class. The paper excludes a Griffiths phase (the power law is at the critical point) and weak universality (all exponents vary). It attributes the effect to the connectivity matrix's eigenvalue spectrum: elliptic at p=0, holey and spread for 0<p<0.5, real at p=0.5.

Load-bearing premise

The whole claim rests on assigning each p the true critical coupling ε_c; if the chosen ε_c lies just below the actual transition, inside the region where generic power-law decay already appears (the paper's Fig. 6 shows this region for p=0.4), then the smoothly changing δ is an artifact of curve fitting rather than a critical exponent.

Editorial extensions

If this is right

  • If correct, the directed percolation universality class is not stable against this kind of frozen, locally asymmetric coupling: a line of critical points with p-dependent exponents replaces a single class.
  • The measured exponents satisfy the hyperscaling relation ν∥=β/δ, so the continuously varying numbers are internally consistent critical exponents, not arbitrary fit parameters.
  • Matching a single exponent (δ at p=0.5) is not enough to identify the universality class; the paper's p=0.5 case has DP-like δ but z and ν∥ several times the DP values.
  • The generic power-law range below ε_c means that standard off-critical scaling collapses cannot be performed in the subcritical regime for p>0; analyses must be restricted to the supercritical side or use other methods.
  • The eigenvalue-spectrum picture suggests that the connectivity matrix's spectral structure, not the local dynamics, controls the critical behavior; spectral observables could become a diagnostic for such transitions.

Reading between the lines

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

  • I infer that the clearest way to separate a genuinely new class from a Griffiths-like effective scaling is to measure the survival probability from a single seed at the same ε_c; if the survival exponent differs from the bulk decay exponent, the transition is not a conventional critical point.
  • The real-axis collapse of the eigenvalue spectrum at p=0.5 suggests a testable prediction: activity fronts should have zero mean drift at p=0.5 and drift left or right at other p; measuring front velocity as a function of p would tie the spectral change to the dynamics.
  • If the exponent family is real, a natural next step is to look for a p-independent master scaling function after rescaling time by the p-dependent ν∥; the authors do not attempt this, and I infer that its existence would place the model under the superuniversality umbrella rather than a wholly new class.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The manuscript numerically studies a one-dimensional coupled map lattice with quenched asymmetric couplings. It claims that for p=0 the model reproduces the directed percolation (DP) universality class, while for p>0 the critical point epsilon_c and the critical exponents delta, beta, nu_parallel, and z vary continuously with p, defining a new universality class. The order parameter is the fraction of turbulent sites m(t), and the paper reports power-law decay at epsilon_c, off-critical scaling in the supercritical regime, and a range of power-law decays in the absorbing phase for p=0.4. The variation of the exponents is conjectured to be related to changes in the eigenvalue spectrum of the connectivity matrix.

Significance. If the central claim is correct, this is a notable example of continuously varying critical exponents in a non-equilibrium absorbing phase transition, which would go beyond the usual few universality classes. The p=0 baseline reproduces the DP exponents, which is a useful internal check, and the model is simple enough to be studied further. However, the central result rests on identifying epsilon_c(p) in a setting where the order parameter also shows generic power-law decay below the apparent critical point, and the dynamical exponent z is measured only for p=0.5. The claim that all critical exponents vary continuously with p is therefore not fully established by the present data.

major comments (3)
  1. [Section 2, Figs. 3 and 6, Table 1] The critical point epsilon_c(p) is identified as the coupling at which m(t) shows a 'clean power law' on a log-log plot. Figure 6 shows, however, that for p=0.4 the order parameter decays as a power law over a range of epsilon values in the absorbing phase, with delta varying continuously with epsilon. Thus the power-law criterion does not discriminate between a true critical point and a Griffiths-type generic power-law region. Since the reported delta(p) in Table 1 is extracted at the visually selected epsilon_c, the continuous variation of delta with p could be an artifact of selecting epsilon_c inside a subcritical region where delta already varies with epsilon at fixed p. The manuscript itself states that subcritical data do not collapse and that z cannot be obtained for p<0.5 because of ultra-slow decay, which is consistent with a Griffiths-phase scenario. The central claim requires
  2. [Section 2, Table 1, hyperscaling relation] The nu_parallel values in Table 1 are numerically close to beta/delta for every p, e.g. p=0.1: 3.67 vs 3.68; p=0.2: 2.86 vs 2.84; p=0.4: 3.71 vs 3.72. The text says 'We also expect a similar value of nu_parallel from the hyperscaling relation nu_parallel = beta/delta' and then concludes that the hyperscaling relation is valid. Because beta and delta have no reported uncertainties, this check is not independent. If nu_parallel was obtained from the off-critical collapse, the near-exact agreement is suspicious and needs explanation; if nu_parallel was instead set equal to beta/delta, then the table does not provide an independent measurement of nu_parallel. Please report the fitted values with error bars and clarify the fitting procedure.
  3. [Section 2, Fig. 8, Table 1] The dynamical exponent z is measured only for p=0.5 (z=3.90). For p=0.1-0.4, z is not measured, and the statement that z would be even larger for 0<p<0.5 is an extrapolation. The abstract and summary nevertheless claim that 'these exponents change continuously' and that all critical exponents vary. With a single value of z, the continuous variation of z with p is not supported. Either provide z for at least a few values of p<0.5, or restrict the claim to the exponents that are actually measured, namely delta, beta, and nu_parallel.
minor comments (5)
  1. [Section 2, text below Eq. (3)] The list of epsilon_c and delta values contains a duplicated label: 'for p=0.1, epsilon_c=0.6625' should presumably be p=0.2, and the sequence skips p=0.4. Compare with Table 1, where p=0.4 has epsilon_c=0.708 or 0.709 (the text gives 0.709 in one place and 0.708 in the table).
  2. [Fig. 3 caption] The caption says 'for epsilon > epsilon_c, epsilon and epsilon_c (top to bottom)' but the figure does not show the three curves in a way that makes the ordering clear. Please label each curve directly with the value of epsilon relative to epsilon_c.
  3. [Figs. 5 and 7] The beta fits report errors, but epsilon_c and delta are quoted without errors. State the range of epsilon used for each fit and the number of data points. The same applies to the off-critical collapse in Fig. 7, where the quality of collapse is only qualitative.
  4. [Introduction, reference [25]] The text says Saha and Mohanty found that for q=3 and q=4 the critical exponents vary continuously, but the cited title is 'Non-reciprocal interactions preserve the universality class of Potts model'. Please reconcile the text with the cited result, or rephrase to avoid apparent contradiction.
  5. [Section 2, Fig. 9] The spectral analysis is suggestive but speculative. State explicitly that it is not used to derive any of the critical exponents, and consider moving it to a separate discussion section.

Circularity Check

1 steps flagged · score 4.0 of 10

Central δ(p) variation is conditioned on ϵ_c being selected by the same power-law criterion that defines δ, while the paper's own Fig. 6 shows a whole range of power laws with varying δ in the absorbing phase.

  1. fitted input called prediction [Section 2, Fig. 3 caption and Fig. 6 text]
    "For all values of p > 0, the order parameter undergoes a power-law decay m(t)∼t −δ at ϵ=ϵ c, similar to DP. ... Fig.6 shows several power-laws over a range of ϵ values in the absorbing phase for p=0.4. The decay exponent δ decreases with an increase in ϵ."

    ϵ_c is not determined by an independent method; it is selected as the coupling where m(t) shows a power law, and δ is then fitted from that same curve and presented as a critical exponent. Fig. 6 shows that, for p=0.4, power-law decay with continuously varying δ occurs over a range of ϵ below the transition — a Griffiths-type generic power-law region. Therefore the criterion 'power law at ϵ_c' does not uniquely identify the true critical point. For each p, the reported δ(p) is δ(ϵ_c(p)) under a non-quantified selection rule, and because δ already varies with ϵ inside the absorbing phase, the observed continuous variation of δ with p can be produced by the fitting procedure alone. The paper offers no independent check: it states that subcritical data do not collapse and that z cannot be mea

full rationale

This is not a paper that derives exponents from an input theory, so there is no definitional or self-citation circularity in the usual sense; the self-citations to the authors' earlier work are contextual and not load-bearing. However, the central claim — continuously varying critical exponents — rests on the identification of ϵ_c(p). The text shows that ϵ_c is tuned until a power law is seen, and δ is measured from that same curve. The paper's own Fig. 6 demonstrates that power laws with varying δ are generic in the absorbing phase for p=0.4, so the power-law criterion is not sufficient to locate the critical point. In the absence of an independent determination of ϵ_c (e.g., a collapse of subcritical data or a separate estimate from steady-state saturation), the continuously varying δ(p) is at least partially an artifact of the fitting procedure. This is a partial, construction-level circularity rather than a full reduction of the result to its input, and no external uniqueness theorem or self-citation chain is invoked. Score 4.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central claim rests on fitted critical points and several numerical-fitting assumptions. No code, data, or certificates are shipped. The only new entity is the claimed continuously-varying-exponent universality class, which is an interpretation of the fits rather than a new physical object.

free parameters (2)
  • critical coupling epsilon_c(p) = 0.633 (p=0.1), 0.6625 (p=0.2), 0.69 (p=0.3), 0.708/0.709 (p=0.4), 0.713 (p=0.5)
    Chosen so that m(t) appears to follow a power law at each p; no independent criterion or uncertainty is given. All extracted exponents are conditioned on this choice.
  • asymmetry strength 0.1 in epsilon1=epsilon/2+0.1 and epsilon2=epsilon/2-0.1 = 0.1
    Hand-set disorder magnitude. The claim of continuously varying exponents is only demonstrated at this one disorder strength, so the generality of the effect is untested.
assumptions (5)
  • domain assumption The p=0 limit of this CML belongs to the directed percolation universality class.
    Inherited from Janaki and Sinha [26]; the violation-of-universality claim is defined against this baseline.
  • domain assumption Averaging over 500 disorder configurations and initial conditions is sufficient to represent quenched disorder.
    No convergence checks or error estimates for the disorder average are shown.
  • domain assumption The asymptotic regime is reached with N=2e5 and the simulated times, so a single system size gives thermodynamic-limit exponents for delta.
    delta is extracted without finite-size scaling; beta and nu_parallel use scaling collapses, but delta does not.
  • domain assumption Standard scaling forms and the hyperscaling relation nu_parallel = beta/delta apply at the critical point.
    Used to obtain nu_parallel from collapse and checked for consistency, but not derived for this model.
  • domain assumption The eigenvalue spectrum of the connectivity matrix is epsilon-independent and relevant to the critical exponents.
    This is a speculative explanatory remark in Section 2, not used to derive the exponents.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Emergence of continuously varying critical exponents in coupled map lattice as an effect of quenched disorder." pith.science (2026). https://pith.science/paper/3TON7X4J

@misc{pith2026250907529,
  author       = {Pith},
  title        = {Pith review of: Emergence of continuously varying critical exponents in coupled map lattice as an effect of quenched disorder},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3TON7X4J}},
  note         = {Machine review of arXiv:2509.07529}
}
abstract

The transition to an absorbing phase in a spatiotemporal system is a well-investigated nonequilibrium dynamic transition. The absorbing phase transitions fall into a few universality classes, defined by the critical exponents observed at the critical point. We present a coupled map lattice (CML) model with quenched disorder in the couplings. In this model, spatial disorders are introduced in the form of asymmetric coupling with a larger coupling ($p$) to a neighbor on the right and a smaller coupling ($1-p$) to the neighbor on the left, for $0 \le p \le0.5$. For $p=0$, the system belongs to the directed percolation universality class. For $p>0$, we observe continuously changing critical exponents at the critical point. The order parameter is the fraction of turbulent sites $m(t)$. %sites that are not in the laminar region. We observe a power-law decay, $m(t) \sim t^{-\delta}$, at the critical point $\epsilon_c$, where $\epsilon$ is the diffusive coupling parameter. These exponents change continuously and do not match any known universality class in any limit. This could be related to changes in the eigenvalue spectrum of the connectivity matrix as the disorder is introduced.

Figures

Figures reproduced from arXiv: 2509.07529 by the authors.

Figure 1
Figure 1. (a) shows the spatiotemporal behavior of the lattice of size [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. (a) shows the plot of |ϵ-ϵc| vs m(∞) on log-log scale for p = 0. We obtain m(∞) ∼ |ϵ − ϵc| β where β=0.253. (b) Shows the finite-size scaling, t Nz vs m(t)Nδz for various sizes of lattice N = 20, 25, 50, 100, 200. The best collapse is obtained for z=1.58. and δ=0.159. (c) Shows the off-critical scaling, m(t)∆−δν∥ vs t∆ν∥ for various values of ∆=ϵ − ϵc in the range [-0.002, 0.002]. The best collapse is obtained for ν… view at source ↗
Figure 3
Figure 3. We plot the decay of the order parameter [PITH_FULL_IMAGE:figures/full_fig_p016_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: (a) Shows the decay of order parameter for various values of [PITH_FULL_IMAGE:figures/full_fig_p017_4.png]
Figure 5
Figure 5. Figure 5: Shows the scaling of asymptotic value of order parameter as [PITH_FULL_IMAGE:figures/full_fig_p018_5.png]
Figure 6
Figure 6. Figure 6: Shows the algebraic decay of order parameter for several values of [PITH_FULL_IMAGE:figures/full_fig_p019_6.png]
Figure 7
Figure 7. Figure 7: Shows the off-critical scaling, t∆ν∥ vs m(t)∆δν∥ on log-log scale for several values of ϵ > ϵc. Here, ∆ = |ϵ − ϵc|. The system size is 2 × 105 . (a) for p = 0.1, ϵc = 0.633, ν∥ = 3.67 (b) for p = 0.2, ϵc = 0.6625, ν∥ = 2.86, (c) for p = 0.3, ϵc = 0.69, ν∥ = 2.47, (d) f…
Figure 8
Figure 8. Figure 8: Shows the plot of the average time τ taken by system of size N = 20, 40, 60 . . . to reach the absorbing state for p = 0.5. The power-law is obtained at the critical point with dynamical exponent z = 3.90. 19 [PITH_FULL_IMAGE:figures/full_fig_p021_8.png]
Figure 9
Figure 9. Figure 9: (a) Shows the distribution of eigenvalues in the complex plane for [PITH_FULL_IMAGE:figures/full_fig_p022_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 29 canonical work pages

  1. [1]

    Henkel, H

    M. Henkel, H. Hinrichsen, S. Lübeck, Scaling properties of absorbing phase transitions, Non-Equilibrium Phase Transitions: Volume I: Ab- sorbing Phase Transitions (2008) 101–196

    work page 2008

  2. [2]

    D. D. Joshi, P. M. Gade, Cellular automata model for period-n synchro- nization: a new universality class, J. Phys. A 58 (2) (2024) 02LT01

    work page 2024

  3. [3]

    Vojta, Rare region effects at classical, quantum and nonequilibrium phase transitions, J

    T. Vojta, Rare region effects at classical, quantum and nonequilibrium phase transitions, J. Phys. A 39 (22) (2006) R143

    work page 2006

  4. [4]

    N. R. Sabe, P. D. Bhoyar, P. M. Gade, Synchronization transition in space–time chaos in the presence of quenched disorder, Commun. Non- linear Sci. Numer. Simul. 138 (2024) 108182

    work page 2024

  5. [5]

    R. J. Baxter, Partition function of the eight-vertex lattice model, Ann. Phys. 70 (1) (1972) 193–228

    work page 1972

  6. [6]

    L. P. Kadanoff, Connections between the critical behavior of the planar model and that of the eight-vertex model, Phys. Rev. Lett. 39 (15) (1977) 903

    work page 1977

  7. [7]

    N. P. Butch, M. B. Maple, Evolution of critical scaling behavior near a ferromagnetic quantum phase transition, Phys. Rev. Lett. 103 (7) (2009) 076404

    work page 2009

  8. [8]

    M.Corti, V.Degiorgio, Criticalexponentsnearthelowerconsolutepoint of nonionic micellar solutions, Phys. Rev. Lett. 55 (19) (1985) 2005. 11

    work page 1985

Show all 29 references
  1. [9]

    Inayat-Hussain, M

    A. Inayat-Hussain, M. Buckingham, Continuously varying critical expo- nents to o (1/n), Phys. Rev. A 41 (10) (1990) 5394

    1990

  2. [10]

    Roberts, O

    B. Roberts, O. I. Motrunich, Infinite randomness with continuously varying critical exponents in the random xyz spin chain, Phys. Rev. B 104 (21) (2021) 214208

    2021

  3. [11]

    Suzuki, New universality of critical exponents, Prog

    M. Suzuki, New universality of critical exponents, Prog. Theor. Phys. 51 (6) (1974) 1992–1993

    1974

  4. [12]

    Bernardi, I

    L. Bernardi, I. Campbell, Violation of universality for ising spin-glass transitions, Phys. Rev. B 52 (17) (1995) 12501

    1995

  5. [13]

    Kondo, Critical exponents, scaling law, universality and renormal- ization group flow in strong coupling qed, Int

    K.-i. Kondo, Critical exponents, scaling law, universality and renormal- ization group flow in strong coupling qed, Int. J. Mod. Physs. A 6 (30) (1991) 5447–5466

    1991

  6. [14]

    N. Khan, P. Sarkar, A. Midya, P. Mandal, P. Mohanty, Continuously varying critical exponents beyond weak universality, Sci. Rep. 7 (1) (2017) 45004

    2017

  7. [15]

    Mukherjee, P

    I. Mukherjee, P. Mohanty, Hidden superuniversality in systems with continuous variation of critical exponents, Phys. Rev. B 108 (17) (2023) 174417

    2023

  8. [16]

    Lemaître, H

    A. Lemaître, H. Chaté, Phase ordering and onset of collective behavior in chaotic coupled map lattices, Phys. Rev. Lett. 82 (6) (1999) 1140

    1999

  9. [17]

    Janssen, K

    H. Janssen, K. Oerding, F. Van Wijland, H. Hilhorst, Lévy-flight spread- ing of epidemic processes leading to percolating clusters, Eur. Phys. J. B 7 (1) (1999) 137–145

    1999

  10. [18]

    M. Jo, J. Lee, K. Choi, B. Kahng, Absorbing phase transition with a continuously varying exponent in a quantum contact process: A neural network approach, Phys. Rev. Res. 3 (1) (2021) 013238

    2021

  11. [19]

    D. Dhar, P. Thomas, Self-organized criticality with continuously varying exponents, Europhys. Lett. 21 (9) (1993) 965

    1993

  12. [20]

    J. D. Noh, H. Park, Universality class of absorbing transitions with con- tinuously varying critical exponents, Phys. Rev. E 69 (1) (2004) 016122. 12

    2004

  13. [21]

    Smallenburg, G

    F. Smallenburg, G. Barkema, Universality class of the pair contact pro- cess with diffusion, Phys. Rev. E 78 (3) (2008) 031129

    2008

  14. [22]

    M. B. Matte, P. M. Gade, Persistence as the order parameter in a gener- alized pair-contact process with diffusion, J. Stat. Mech.: Theory Exp. 2016 (11) (2016) 113203

    2016

  15. [23]

    T.-C. Yi, C. Ding, M. Liu, L. Li, W.-L. You, Continuously varying critical exponents in an exactly solvable long-range cluster xy model, Phys. Rev. A 111 (2) (2025) 023307

    2025

  16. [24]

    Delfino, Nonuniversality in random criticality, J

    G. Delfino, Nonuniversality in random criticality, J. Stat. Mech.: Theory Exp. 2025 (1) (2025) 013211

    2025

  17. [25]

    S. K. Saha, P. K. Mohanty, Non-reciprocal interactions preserve the uni- versality class of potts model, arXiv preprint arXiv:2412.19664 (2024). arXiv:2412.19664. URLhttps://arxiv.org/abs/2412.19664

    2024 arXiv

  18. [26]

    Janaki, S

    T. Janaki, S. Sinha, N. Gupte, Evidence for directed percolation uni- versality at the onset of spatiotemporal intermittency in coupled circle maps, Phys. Rev. E 67 (5) (2003) 056218

    2003

  19. [27]

    P. D. Bhoyar, M. C. Warambhe, S. Belkhude, P. M. Gade, Robustness of directed percolation under relaxation of prerequisites: role of quenched disorder and memory, Eur. Phys. J. B 95 (4) (2022) 64

    2022

  20. [28]

    A. Y. Tretyakov, N. Inui, N. Konno, Phase transition for the one-sided contact process, J. Phys. Soc. Jpn. 66 (12) (1997) 3764–3769

    1997

  21. [29]

    Ashida, Z

    Y. Ashida, Z. Gong, M. Ueda, Non-hermitian physics, Adv. Phys. 69 (3) (2020) 249–435. 13 Figure 3: We plot the decay of the order parameterm(t)with time forϵ > ϵ c,ϵand ϵ < ϵc(top to bottom). The system size isN= 2×10 5. We averaged over 500 disorder configurations and initial...

    2020

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.