Pith. sign in

REVIEW 2 major objections 6 minor 48 references

Explosive opinion spreading with polarization and depolarization via asymmetric perception

T0 review · 2 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Asymmetric perception of others' opinions turns smooth opinion change into an abrupt, hysteresis-locked polarization transition.

desk verdict A plausible and interesting mechanism for explosive opinion transitions, but the mean-field derivation that carries the α>1 subcritical claim drops correlation terms that can dominate near the critical point. read the letter →

arxiv 2501.11863 v1 pith:F7QP6SKN submitted 2025-01-21 physics.soc-ph nlin.PS

Haoyang Qian , Malbor Asllani This is my paper · ORCID
classification physics.soc-phnlin.PS
keywords asymmetricperceptionexplosivepolarizationhysteresisnonlinearincidencemean-fieldopiniondynamicsmodularnetworksSISmodel
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 asks whether a minimal change in how people perceive the opinions of others can turn gradual opinion change into abrupt, explosive polarization. It builds an epidemic-style susceptible-infected-susceptible model in which the adoption rate of a new opinion depends nonlinearly on the number of neighbors already holding it (a term proportional to $1+\alpha x_j^{d-1}$), and it claims that when perception underestimates the true prevalence ($\alpha>1$), the model undergoes a first-order phase transition: the new opinion takes over discontinuously, and once established it stays until the control parameter is lowered past a second, smaller critical point, producing hysteresis. The authors derive a one-dimensional mean-field normal form that changes from transcritical/supercritical to subcritical exactly when $\alpha>1$, and they show that continuous transitions localize the new opinion inside one community while discontinuous transitions erase that pattern. If the claim is right, the model supplies a single mechanism for both explosive polarization and abrupt, hard-to-reverse depolarization.

What carries the argument

The load-bearing object is the one-dimensional mean-field equation (Eq. 3), $\langle \dot{x}\rangle = (\tilde{\beta}-\gamma)\langle x\rangle - \tilde{\beta}\langle x\rangle^2 + \alpha\tilde{\beta}(1-\langle x\rangle)\langle x\rangle^d$, with $\tilde{\beta}=\beta\langle k\rangle$. It is derived from the full network equations by assuming $x_i\sim k_i$ and a narrow, symmetric degree distribution, so that deviations $\delta x_i$ and $\delta k_i$ average out. For $d=2$ this becomes $\langle \dot{x}\rangle = (\tilde{\beta}-\gamma)\langle x\rangle + \tilde{\beta}(\alpha-1)\langle x\rangle^2 - \alpha\tilde{\beta}\langle x\rangle^3$, and the sign of the quadratic term at small $\langle x\rangle$ decides whether the bifurcation is transcritical ($\alpha=0$), supercritical ($\alpha<1$ and $\alpha=1$), or subcritical ($\alpha>1$). The second piece of machinery is spectral perturbation theory for modular networks: the eigenvectors belonging to the $M$ largest eigenvalues are localized on individual modules, which is what makes the opinion pattern polarize along community lines; the Perron-Frobenius theorem guarantees the principal eigenvector is the only positive one and therefore the relevant critical mode.

What would settle it

Run the individual-based model (Eq. 2) on a modular network with a broad, asymmetric degree distribution, with $d=2$ and $\alpha>1$, sweeping the reversion rate $\gamma$ upward and downward; if the equilibrium $\langle x\rangle$ shows no discontinuous jump and no bistable window, then the mean-field subcritical bifurcation is not the actual mechanism.

Watch

Extended reading notes

Core claim

The central claim is that asymmetric perception, encoded as nonlinear incidence in a metanode SIS model, is sufficient to produce explosive polarization and explosive depolarization. Each node contains $N$ opinion units that flip between the old opinion $S$ and the new opinion $I$; the standard linear contagion term $\beta(1-x_i)\sum_j A_{ij}x_j$ is modified to $\beta(1-x_i)\sum_j A_{ij}x_j(1+\alpha x_j^{d-1})$, where $d>1$ means underestimated and $0<d<1$ overestimated perception. Linearization shows the initial instability is set by $\beta\lambda_{\max}^A>\gamma$, independent of $\alpha$. For $d=2$, the mean-field equilibrium equation reduces to a cubic normal form whose quadratic coefficient is $\tilde{\beta}(\alpha-1)$; when $\alpha>1$ the bifurcation becomes subcritical, producing a bistable region, a discontinuous jump in the average opinion as $\gamma$ is lowered, hysteresis on the return path, and an abrupt switch back to the old opinion at a second critical point. For $0<d<1$, the new opinion is always embraced and abrupt depolarization is mitigated. Continuous transitions leave the spatial pattern close to the critical eigenvector, so the new opinion concentrates in one network community; discontinuous transitions deviate from that pattern and can produce a uniform opinion shift.

Load-bearing premise

The analytic story rests on a simplification that assumes each person's opinion is proportional to how connected they are and that random fluctuations cancel out, so the one-dimensional equation, not the full network, is what predicts the explosive transition.

Editorial extensions

If this is right

  • For $\alpha>1$ (underestimated perception), adoption of a new opinion is discontinuous: once the reversion rate $\gamma$ passes the critical value, the average opinion jumps to a higher level, with the jump size growing with $\alpha$.
  • The backward path is hysteretic: lowering $\gamma$ back to its original value does not undo the shift; the new opinion persists until a second, smaller critical point, so depolarization happens suddenly and the new opinion is resistant to reversal.
  • For $0<d<1$ (overestimated perception), the new opinion is always embraced and the abrupt depolarization is suppressed, so perception bias determines not just whether but how abruptly opinions change.
  • In the continuous regime ($\alpha\le 1$), the final opinion pattern near criticality mirrors the critical eigenvector and thus localizes in a single community; in the discontinuous regime the pattern decouples from that eigenvector and can become uniform, meaning explosive transitions also change who ends up holding the new opinion.
  • The condition for the new opinion to start spreading, $\beta\lambda_{\max}^A>\gamma$, is independent of $\alpha$, so network structure alone sets the initial threshold while perception asymmetry controls the order of the transition.

Reading between the lines

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

  • Beyond the paper: because nonlinear incidence is a generic feature of binary-state dynamics, the same $\alpha>1$ condition may produce first-order cascades in rumor spreading, behavioral adoption, and financial herding, where a threshold-curve shape similar to the opinion model appears.
  • Beyond the paper: the mean-field proxy is derived under a narrow symmetric degree distribution and by cancelling fluctuation products; on empirical networks with hubs, quantitative discrepancies are expected, and measuring the actual equilibrium curve would separate perception-driven bistability from purely structural effects.
  • Beyond the paper: a testable intervention follows directly: if platforms make perceived consensus more proportional to true prevalence (pushing $\alpha$ toward $1$ or below), the model predicts the explosive jump and hysteresis should disappear; this could be probed with online experiments that modulate exposure salience.
  • Beyond the paper: fitting the effective exponent $d$ to adoption time series would place a community in the underestimation ($d>1$) or overestimation ($0<d<1$) regime, and the two regimes make opposite predictions about depolarization that observational data could discriminate.
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

2 major / 6 minor

Summary. The paper proposes an SIS-type opinion dynamics model in which asymmetric perception is encoded by a nonlinear incidence term, β(1−xi)Σ_j A_ij x_j (1 + α x_j^{d−1}). The central claim is that for underestimating perception (d>1) with α>1, the model exhibits explosive, first-order polarization transitions and hysteresis, while overestimation (0<d<1) mitigates abrupt depolarization. To explain this analytically, the authors reduce the network dynamics to a one-dimensional mean-field equation, Eq. (3), which takes the form of a cubic normal form whose bifurcation type is controlled by α. Numerical simulations on a small modular network of three Erdős–Rényi modules show the predicted continuous-to-discontinuous transition and hysteresis, and the pattern of the polarized state is connected to the spectral properties of modular networks via perturbation theory. The Supplemental Material extends the comparison to scale-free and small-world networks, reporting qualitative agreement for scale-free networks and a quantitative error analysis for Watts–Strogatz networks.

Significance. If the central claim were rigorously established, the model would provide a simple and appealing mechanism for explosive opinion shifts and for the difficulty of reversing them, with falsifiable predictions (a first-order transition when α>1 and d>1). The paper includes concrete numerical simulations on modular, scale-free, and small-world networks, and it makes an honest attempt to connect the network-level phenomenology to a mean-field normal form. However, the analytical bridge—the derivation of Eq. (3)—rests on uncontrolled approximations that, under the model's own ansatz, are not guaranteed to be small. Because the sign of the quadratic coefficient in the normal form is precisely what determines whether the bifurcation is subcritical or supercritical, the paper's strongest analytical claim (that nonlinear perception is a minimal universal mechanism) is not yet supported. The numerical evidence for the phenomenon on specific networks is credible, but the general mechanism remains to be demonstrated.

major comments (2)
  1. [Supplemental Material, Section I, Eq. (6) and the paragraph that follows] The reduction to Eq. (7) drops the correlation terms (β/N)[(1−2⟨x⟩+αd⟨x⟩^{d−1}(1−⟨x⟩)) Σ_j δx_j δk_j − (1+αd⟨x⟩^{d−1}) Σ_{i,j} A_ij δx_i δx_j] with the assertion that deviations are small and the degree distribution is narrow and symmetric. This assertion is not checked and is in fact inconsistent with the model's own ansatz x_i ∼ k_i: under that ansatz δx_i ≈ (⟨x⟩/⟨k⟩)δk_i, so Σ_j δx_j δk_j ≈ N (⟨x⟩/⟨k⟩) Var(k), which is first order in ⟨x⟩ and does not vanish for a narrow distribution. Near the mean-field critical point γ = β⟨k⟩, the retained linear term (β⟨k⟩ − γ)⟨x⟩ vanishes, so this dropped term can dominate and shift the critical point; the SM itself concedes that the difference in critical points depends on Σ_j δx_j δk_j. Moreover, the term Σ_{i,j} A_ij δx_i δx_j renormalizes the coefficient of ⟨x⟩² in the normal form, which is exactly the coefficient that decides whether the bifurcation is subcritical or supercritical. Therefore Eq. (3) and the claim that α>1 implies a first-order transition are not a controlled consequence of the individual-based model, Eq. (2). The numerical simulations on the modular network support the phenomenon, but the analytical mechanism is not established.
  2. [Supplemental Material, Section IV.B (Fig. 7)] The paper's own quantitative comparison on small-world networks shows that the error between the individual-based simulation and the degree-based mean-field proxy can be large and generally grows with the rewiring probability p. For example, at k=4 the error is large across all p, and at moderate k the error increases with p before declining near p=1. This contradicts the impression given in the main text that Eq. (3) provides a general explanation for the observed transitions. The paper needs to either characterize the regime in which the approximation is quantitatively faithful (e.g., by verifying that the dropped correlation terms are indeed negligible) or substantially weaken the claim that the mean-field proxy explains the mechanism in a universal way.
minor comments (6)
  1. [Fig. 1 caption] The word 'Understimate' should be 'Underestimate'.
  2. [Supplemental Material, Section I, after Eq. (6)] The sentence beginning 'When all terms ⟨x⟩ are considered' should be reworded to 'When all terms involving ⟨x⟩ are considered', since the current phrasing is unclear.
  3. [Main text, after Eq. (2)] The phrase 'for simplicity of representation α → αβ' is confusing; please clarify that α has been rescaled by absorbing β, and state the resulting dimensions of α.
  4. [References] Reference [33] (Granovetter 1978) is a duplicate of Reference [17]; please remove or distinguish them.
  5. [Supplemental Material, Section IV.A and footnote [44]] Footnote [44] calls Eq. (4) a mean-field approximation, but Eq. (4) is the individual-based model; the terminology is confusing and should be clarified.
  6. [Fig. 1(b) inset] The inset shows the degree distribution but the caption does not state the mean and variance of the degree distribution; adding these values would help the reader judge the validity of the 'narrow and symmetric' assumption.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the explosive-polarization result follows algebraically from the stated model, and the only self-citations are supporting, not load-bearing.

full rationale

The central derivation is self-contained. Starting from Eq. (2), the paper reduces the network dynamics to the one-dimensional proxy Eq. (3) under stated assumptions (xi ~ ki and a narrow, symmetric degree distribution). For d = 2, the normal form Eq. (9) has quadratic coefficient beta<k>(alpha - 1), so the condition alpha > 1 for a subcritical bifurcation is a direct algebraic consequence of the model equations, not a fitted parameter renamed as a prediction. The parameters alpha and d are free model parameters explored in a bifurcation analysis, not calibrated to external data. The paper compares the mean-field proxy with direct numerical integration of Eq. (2), providing an independent check. The only self-citations, Refs. [41,42], support the statement that modular-network eigenvectors localize within communities; however, the Supplemental Material supplies the block-perturbation and Perron-Frobenius argument, so those citations are not the sole or load-bearing justification. The SM explicitly acknowledges the gap between individual-based and degree-based mean-field critical points and documents quantitative deviations on scale-free and small-world networks; this is an approximation-accuracy limitation, not circularity. Overall, no claimed result reduces to its own input by construction.

Assumptions & free parameters 3 free parameters · 6 assumptions · 2 invented entities

The central claims rest on the assumed nonlinear-incidence model and on the mean-field reduction. The free parameters α and d are central to the predicted regimes, and the mean-field assumptions, especially the neglect of correlation terms, are the most fragile parts of the analytical derivation. No new empirical entities beyond model abstractions are introduced, but the D compartment and the metanode particles are invented constructs without external validation.

free parameters (3)
  • α = 0, 0.8, 1, 1.1, 1.2, 2 (chosen by hand)
    Coefficient of the nonlinear incidence term. The condition α > 1 is required for the subcritical bifurcation and explosive transition in the d=2 case, so the main qualitative result is parameter-dependent.
  • d = 2 for underestimation; 0 < d < 1 for overestimation
    Exponent of nonlinear incidence. The paper states it is 'typically estimated from experimental data' but does not estimate it or provide a citation, leaving it as a hand-chosen parameter.
  • β = 1 (dimensionless)
    Adoption rate, set to 1 in all simulations. It sets the scale of the critical reversion rate but is not fitted to data.
assumptions (6)
  • domain assumption Individuals are modeled as metanodes containing a fixed number N of noninteractive opinion units; the continuous state variable xi is the fraction of units holding the new opinion.
    Introduced in the section 'Explosive opinion spreading...' to justify mass-action kinetics and the SIS-like equations. It is a standard compartmental abstraction, not derived from microscopic social mechanisms.
  • domain assumption The dynamics follow an SIS structure with spontaneous loss of interest at rate γ and adoption at rate β via contacts.
    Borrowed from epidemic modeling and standard in opinion-dynamics literature; the paper cites Refs. [5,31] for this framework.
  • ad hoc to paper The specific nonlinear incidence functional form β(1 − xi) Σ_j A_ij x_j (1 + α x_j^{d−1}) captures asymmetric perception.
    The form is introduced by analogy to nonlinear incidence in epidemiology (Refs. [20,21]) to represent overestimation or underestimation, but no microfoundation or empirical validation is provided for this exact expression.
  • domain assumption Degree-based mean-field assumption: xi(t) ∼ ki, i.e., the opinion state of a node scales with its degree.
    Stated before Eq. (3) and used to reduce the network system to one dimension; this is a standard but uncontrolled approximation in network mean-field theories.
  • domain assumption The degree distribution P(k) is narrow and symmetric, so deviation sums and products (Σ δxi, Σ δki, δxi δkj, Σ A_ij δxi δxj) can be neglected.
    Used in the Supplemental Material to derive Eq. (7). The narrow-symmetric assumption is explicitly violated for scale-free networks, and the SM shows only qualitative agreement there.
  • standard math For weakly perturbed M-modular networks, the eigenvectors corresponding to the M largest eigenvalues are localized on individual modules.
    Derived in the Supplemental Material via spectral perturbation theory and Perron-Frobenius; the paper also cites Refs. [41,42] for this property.
invented entities (2)
  • Metanode with N noninteractive opinion particles
    purpose: Provides a continuous fraction variable xi per individual and enables the mass-action reaction scheme in Eq. (1).
    A modeling abstraction with no direct empirical counterpart described in the paper.
  • D compartment (indifferent individuals) in the supplemental purely nonlinear model
    purpose: Introduced to bypass the technical limitation that the non-null fixed point disappears when D=1 and −γ=0 in Eq. (12).
    An extra model compartment added in the SM solely to fix the bifurcation structure; it has no independent evidence or empirical justification.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Explosive opinion spreading with polarization and depolarization via asymmetric perception." pith.science (2026). https://pith.science/paper/F7QP6SKN

@misc{pith2026250111863,
  author       = {Pith},
  title        = {Pith review of: Explosive opinion spreading with polarization and depolarization via asymmetric perception},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F7QP6SKN}},
  note         = {Machine review of arXiv:2501.11863}
}
read the original abstract

Polarization significantly influences societal divisions across economic, political, religious, and ideological lines. Understanding these mechanisms is key to devising strategies to mitigate such divisions and promote depolarization. Our study examines how asymmetric opinion perception, modeled through nonlinear incidence terms, affects polarization and depolarization within structured communities. We demonstrate that such asymmetry leads to explosive polarization and causes a hysteresis effect responsible for abrupt depolarization. We develop a mean-field approximation to explain how nonlinear incidence results in first-order phase transitions and the nature of bifurcations. This approach also helps in understanding how opinions polarize according to underlying social network communities and how these phenomena intertwine with the nature of such transitions. Numerical simulations corroborate the analytical findings.

Figures

Figures reproduced from arXiv: 2501.11863 by the authors.

Figure 1
Figure 1. (b) demonstrates that Eq. (3) exhibits quali￾tatively similar behavior to that observed earlier for the full system (2), thereby confirming the ansatz introduced earlier when applied to a symmetric narrow degree dis￾tribution (shown by the histogram in the inset). In par￾ticular, considering d = 2 simplifies Eq. (3) to ⟨x˙⟩ =  β˜ − γ  ⟨x⟩ + β˜(α − 1)⟨x⟩ 2 − αβ˜⟨x⟩ 3 allowing a direct comparison between the linear … view at source ↗
Figure 2
Figure 2. FIG. 2. (a) The equilibrium state for each node [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. The mean equilibrium state, [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Comparison between the two principal eigenvectors of [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. (a) The red point represents Λ [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. (a) [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]
Figure 7
Figure 7. Figure 7: FIG. 7. (a) In the main panel, we set [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]
Figure 8
Figure 8. Figure 8: FIG. 8 [PITH_FULL_IMAGE:figures/full_fig_p016_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 45 canonical work pages

  1. [1]

    M. O. Jackson et al., Social and economic networks, Vol. 3 (Princeton university press Princeton, 2008)

    work page 2008

  2. [2]

    Lazer, A

    D. Lazer, A. Pentland, L. Adamic, S. Aral, A.-L. Barab´ asi, D. Brewer, N. Christakis, N. Contractor, J. Fowler, M. Gutmann, et al. , Science 323, 721 (2009)

    work page 2009

  3. [3]

    Centola, Science 329, 1194 (2010)

    D. Centola, Science 329, 1194 (2010)

    work page 2010

  4. [4]

    Cinelli, G

    M. Cinelli, G. De Francisci Morales, A. Galeazzi, W. Quattrociocchi, and M. Starnini, Proceedings of the National Academy of Sciences 118, e2023301118 (2021)

    work page 2021

  5. [5]

    Castellano, S

    C. Castellano, S. Fortunato, and V. Loreto, Reviews of Modern Physics 81, 591 (2009)

    work page 2009

  6. [6]

    Easley and J

    D. Easley and J. Kleinberg, Networks, crowds, and mar- kets: Reasoning about a highly connected world , Vol. 1 (Cambridge university press Cambridge, 2010)

    work page 2010

  7. [7]

    M. E. J. Newman, Networks, second edition ed. (Oxford University Press, Oxford, United Kingdom ; New York, NY, United States of America, 2018)

    work page 2018

  8. [8]

    Axelrod, Journal of conflict resolution 41, 203 (1997)

    R. Axelrod, Journal of conflict resolution 41, 203 (1997)

    work page 1997

Show all 48 references
  1. [9]

    M. A. Nowak, N. L. Komarova, and P. Niyogi, Nature 417, 611 (2002)

    2002

  2. [10]

    Helbing, I

    D. Helbing, I. Farkas, and T. Vicsek, Nature 407, 487 (2000)

    2000

  3. [11]

    C. A. Bail, L. P. Argyle, T. W. Brown, J. P. Bumpus, H. Chen, M. B. Hunzaker, J. Lee, M. Mann, F. Merhout, and A. Volfovsky, Proceedings of the National Academy of Sciences 115, 9216 (2018)

    2018

  4. [12]

    Baumann, P

    F. Baumann, P. Lorenz-Spreen, I. M. Sokolov, and M. Starnini, Physical Review Letters124, 048301 (2020)

    2020

  5. [13]

    Baumann, P

    F. Baumann, P. Lorenz-Spreen, I. M. Sokolov, and M. Starnini, Physical Review X 11, 011012 (2021)

    2021

  6. [14]

    Understanding the transition from consensus to polarization necessitates studying the nature of bifurcations, which in turn re- quires continuous variable models [30]

    uses a phase-coupled model to explain explosive de- polarization, analogous to the abrupt synchronization observed in the Kuramoto model [29]. Understanding the transition from consensus to polarization necessitates studying the nature of bifurcations, which in turn re- quires...

  7. [15]

    J. Ojer, M. Starnini, and R. Pastor-Satorras, Physical Review Letters 130, 207401 (2023)

    2023

  8. [16]

    Balietti, L

    S. Balietti, L. Getoor, D. G. Goldstein, and D. J. Watts, Proceedings of the National Academy of Sciences 118, e2112552118 (2021)

    2021

  9. [17]

    Sornette, Why stock markets crash: Critical events in complex financial systems (Princeton University Press, 2003)

    D. Sornette, Why stock markets crash: Critical events in complex financial systems (Princeton University Press, 2003)

    2003

  10. [19]

    D. J. Watts, Proceedings of the National Academy of Sciences 99, 5766 (2002)

    2002

  11. [20]

    Milli, Applied Network Science 6, 76 (2021)

    L. Milli, Applied Network Science 6, 76 (2021)

    2021

  12. [21]

    W.-m. Liu, S. A. Levin, and Y. Iwasa, Journal of Math- ematical Biology 23, 187 (1986)

    1986

  13. [22]

    W.-m. Liu, H. W. Hethcote, and S. A. Levin, Journal of Mathematical Biology 25, 359 (1987)

    1987

  14. [23]

    Carletti, M

    T. Carletti, M. Asllani, D. Fanelli, and V. Latora, Phys- ical Review Research 2, 033012 (2020)

    2020

  15. [24]

    B. A. Siebert, J. P. Gleeson, and M. Asllani, Chaos, Soli- tons & Fractals 161, 112322 (2022)

    2022

  16. [25]

    de Kemmeter, A

    J.-F. de Kemmeter, A. Byrne, A. Dunne, T. Carletti, and M. Asllani, Physical Review E 110, L012201 (2024)

    2024

  17. [26]

    Cross and H

    M. Cross and H. Greenside, Pattern formation and dy- namics in nonequilibrium systems (Cambridge University Press, Cambridge, UK ; New York, 2009)

    2009

  18. [27]

    Dandekar, A

    P. Dandekar, A. Goel, and D. T. Lee, Proceedings of the National Academy of Sciences 110, 5791 (2013)

    2013

  19. [28]

    Deffuant, D

    G. Deffuant, D. Neau, F. Amblard, and G. Weisbuch, Advances in Complex Systems 3, 87 (2000)

    2000

  20. [29]

    Rainer and U

    H. Rainer and U. Krause, Journal of Artificial Societies and Social Simulation 5 (2002)

    2002

  21. [30]

    Kuramoto, in International Symposium on Mathemat- ical Problems in Theoretical Physics: January 23–29, 1975, Kyoto University, Kyoto/Japan (Springer, 1975) pp

    Y. Kuramoto, in International Symposium on Mathemat- ical Problems in Theoretical Physics: January 23–29, 1975, Kyoto University, Kyoto/Japan (Springer, 1975) pp. 420–422

    1975

  22. [31]

    Strogatz, Nonlinear dynamics and chaos: with appli- cations to physics, biology, chemistry, and engineering , repr

    S. Strogatz, Nonlinear dynamics and chaos: with appli- cations to physics, biology, chemistry, and engineering , repr. ed., Studies in nonlinearity (Westview Press, Cam- bridge, Mass, 2007)

    2007

  23. [32]

    I. Z. Kiss, J. C. Miller, and P. L. Simon, Mathematics of Epidemics on Networks: From Exact to Approximate Models, Interdisciplinary Applied Mathematics, Vol. 46 (Springer International Publishing, Cham, 2017)

    2017

  24. [33]

    Nekovee, Y

    M. Nekovee, Y. Moreno, G. Bianconi, and M. Marsili, Physica A: Statistical Mechanics and its Applications 374, 457 (2007)

    2007

  25. [34]

    Granovetter, American Journal of Sociology 83, 1420 (1978)

    M. Granovetter, American Journal of Sociology 83, 1420 (1978)

    1978

  26. [35]

    Horn and R

    F. Horn and R. Jackson, Archive for rational mechanics and analysis 47, 81 (1972)

    1972

  27. [36]

    Pastor-Satorras and A

    R. Pastor-Satorras and A. Vespignani, Physical Review E 63, 066117 (2001)

    2001

  28. [37]

    Pastor-Satorras and A

    R. Pastor-Satorras and A. Vespignani, Physical Review Letters 86, 3200 (2001)

    2001

  29. [38]

    Barth´ elemy, A

    M. Barth´ elemy, A. Barrat, R. Pastor-Satorras, and A. Vespignani, Physical Review Letters 92, 178701 (2004)

    2004

  30. [39]

    Barth´ elemy, A

    M. Barth´ elemy, A. Barrat, R. Pastor-Satorras, and A. Vespignani, Journal of Theoretical Biology 235, 275 (2005)

    2005

  31. [40]

    R. A. Horn and C. R. Johnson, Matrix analysis (Cam- bridge university press, 2012)

    2012

  32. [41]

    We have numerically shown that the fixed point xi = 0 ∀i is unstable for Eq. (2). For the analytical proxy, since Eq. (3) for 0 < d <1 is not analytical at the origin, one can use V (⟨x⟩) = 1 2 ⟨x⟩2 as a candidate Lyapunov function to analyze the stability of the fixed point ⟨...

  33. [42]

    B. A. Siebert, C. L. Hall, J. P. Gleeson, and M. Asllani, Physical Review E 102, 052306 (2020)

    2020

  34. [43]

    Asllani, B

    M. Asllani, B. A. Siebert, A. Arenas, and J. P. Gleeson, Chaos: An Interdisciplinary Journal of Nonlinear Science 32 (2022). 6

    2022

  35. [44]

    D. J. Watts and S. H. Strogatz, Nature 393, 440 (1998)

    1998

  36. [45]

    The reason for referring to Eq. (4) as a Mean-Field is that it represents a mean-field approximation, which, in principle, should be derived from the particle dynamics described by the reaction through an averaging process over the master equation, thereby neglecting any stoch...

  37. [46]

    xi(t) ∼ ki, where ki represents the degree of node i

  38. [47]

    The degree distribution P (k) is symmetric and narrow. Given the equation: ⟨ ˙x⟩ = −γ⟨x⟩ + β N X j xjkj − β N X i,j Aijxixj + α β N X i (1 − xi) X j Aijxd j (5) Substituting xi = ⟨x⟩ + δxi and ki = ⟨k⟩ + δki: ⟨ ˙x⟩ = −γ⟨x⟩+ β N X j (⟨x⟩+δxj)(⟨k⟩+δkj)− β N X i,j Aij(⟨x⟩+δxi)(⟨x...

  39. [48]

    Case 1: α = 0 When α = 0, the equation simplifies to: ⟨ ˙x⟩ = ( ˜β − γ)⟨x⟩ −˜β⟨x⟩2 (8) This is a quadratic equation, corresponding to the normal form of a transcritical bifurcation [30]

  40. [49]

    Case 2: α >0 and d = 2 When α >0 and d = 2, the equation becomes: ⟨ ˙x⟩ = −γ⟨x⟩ + ˜β⟨x⟩ −˜β⟨x⟩2 + α ˜β(1 − ⟨x⟩)⟨x⟩2 Factoring out ˜β only for the middle term, we get: ⟨ ˙x⟩ = ( ˜β − γ)⟨x⟩ + ˜β(α − 1)⟨x⟩2 − α ˜β⟨x⟩3 (9) This is a cubic equation, corresponding to the normal form...

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.