Pith. sign in

REVIEW 3 major objections 5 minor 27 references

Expanders in Models of Social Networks

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read When "close enough" means matching on any one dimension, random geometric social networks become strong expanders, unlike Euclidean versions.

desk verdict The sublinear-scale expansion theorem is genuinely new and the appendix argument looks sound, but the advertised expander statement for sets up to |V'|/2 depends on an unproved adaptation of [23]. read the letter →

arxiv 2506.19485 v1 pith:6LJFVHOH submitted 2025-06-24 cs.SI math.COmath.MG

Marc Kaufmann , Johannes Lengler , Ulysse Schaller , Konstantin Sturm This is my paper · ORCID
classification cs.SImath.COmath.MG MSC 05C8005C4860C0568R10
keywords GeometricInhomogeneousRandomGraphsMinimum-ComponentDistanceexpanderpower-lawdegreedistributionsmallseparatorsgraphmodelssocialnetworks
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 Geometric Inhomogeneous Random Graphs in which two vertices are considered close if they agree along at least one latent dimension (Minimum-Component Distance, or MCD), rather than along all dimensions. It establishes that for dimension $d\ge 2$, the subgraph induced by vertices of polylogarithmic weight or degree is a strong vertex expander: every subset of size $o(|V'|)$ has external neighbourhood of size $\omega(1)$ times itself, and every subset up to a constant fraction has at least constant expansion. This is the opposite of the Euclidean version, which has small separators at every scale. The result matters because expanders support fast random walks, fast rumour and infection spread, robust routing, and fast distributed and approximation algorithms. The paper's quantitative statement gives expansion factor roughly $\min\{\log^{\gamma(3-\tau)} n,\ (|V'|/|S|)^{1-1/c_d}\}$.

What carries the argument

The load-bearing objects are the Minimum-Component Distance and a strip partition of the torus ground space. MCD makes two vertices close when $\min_i |x_{u,i}-x_{v,i}|$ is small, so edges form when the vertices match on any single coordinate; this is why the distance is not a metric and violates the triangle inequality. The proof divides each coordinate into strips of width roughly $n/\log^{2\gamma}n$, chosen so that two vertices of weight $\log^\gamma n$ in the same strip are connected with constant probability. Proposition 1 is a counting argument showing that, with high probability, any set $S$ of size $s$ spans many distinct strips in some coordinate, with the number $k$ of such strips at least of order $s\cdot\min\{1,\ (|V'|/s)^{1-1/c_d}/\log^{\gamma(3-\tau)}n\}$. Proposition 2 shows each vertex has $\Theta(\log^{\gamma(3-\tau)}n)$ neighbours of suitable weight inside its own strip. Adding these strip-neighbourhoods and subtracting $|S|$ gives the lower bound on the external neighbourhood.

What would settle it

One concrete check: in an MCD-GIRG with $d\ge 2$, $\tau\in(2,3)$, and $\gamma>1/(3-\tau)$, look for a sequence of vertex sets $S_n\subset V'$ with $\varepsilon|V'|\le |S_n|\le |V'|/2$ whose external neighbourhood in $G'$ is $o(|S_n|)$ with non-negligible probability. If such sets exist, the full expander statement fails in the constant-fraction regime, even if the $\omega(1)$ expansion for sublinear sets survives. The natural candidate geometry to inspect is the cross-shaped MCD ball, where strip-aligned cuts may behave differently from Euclidean cube separators.

Watch

Extended reading notes

Core claim

The central claim is Theorem 2: for each $d\ge 2$ there is $c_d>1$ such that, with high probability, the subgraph $G'$ of an MCD-GIRG induced by vertices of weight at least $c'\log^\gamma n$ has the property that every $S\subseteq V'$ with $|S|\le \varepsilon|V'|$ satisfies $|\mathrm{Next}(S)\cap V'| \ge \varepsilon\min\{\log^{\gamma(3-\tau)} n,\ (|V'|/|S|)^{1-1/c_d}\}|S|$. In particular the expansion factor is $\omega(1)$ whenever $|S|=o(|V'|)$. The same statement holds for subgraphs induced by degree thresholds and by degree intervals; the interval version is almost regular, with degrees within a factor $2+o(1)$. The paper notes that the regime of constant-fraction sets rests on adapting the earlier no-sublinear-separator result for the giant component to the induced subgraph $G'$, and the formal proof in the appendix covers only $|S|\le\varepsilon|V'|$.

Load-bearing premise

The load-bearing premise is that the previously known result of [23] — that linear-size subsets of the giant component of an MCD-GIRG have linear external neighbourhoods — can be transferred without loss to the induced subgraph $G'$ of high-weight vertices; the paper states this transfer is straightforward but does not prove it, and the formal proof only handles sets of size at most $\varepsilon|V'|$.

Editorial extensions

If this is right

  • Random walks on the high-degree subgraph $G''$ mix rapidly, so sampling vertices and computing PageRank-style quantities from local access becomes fast.
  • Rumour-spreading and SI infection models reach essentially all large-degree vertices in polylogarithmic rounds, and infection spreads among them in $o(\log n)$ rounds.
  • Distributed consensus dynamics such as best-of-$k$ voting converge in $O(\log n)$ rounds on the almost-regular subgraph.
  • The induced subgraph remains well-connected after random removal of a small constant fraction of edges or vertices, supporting robust routing.
  • Approximation algorithms for balanced cuts, colouring, independent set, and related NP-hard problems inherit expander-based guarantees when run on $G''$.

Reading between the lines

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

  • If the same proof strategy extends to general Boolean Distance Functions that are not single-coordinate outer max, the expander-versus-separator split would become a full classification of all such geometries; the paper leaves this as an explicit open problem.
  • The strong expansion at sublinear scales suggests MCD-GIRGs may support sublinear-query estimation of global quantities, since a small random sample of high-degree vertices quickly reaches the rest of that subgraph.
  • The mechanism is essentially that 'OR' similarity — agreement on any one feature — makes neighbourhoods spread across many feature-strips, whereas 'AND' similarity concentrates them; this gives a design principle for synthetic networks that need fast mixing.
  • A direct empirical test: generate MCD-GIRG and Euclidean GIRG with the same degree sequence and measure the spectral gap of the normalized Laplacian of the high-degree subgraph; the dichotomy should be visible at moderate $n$.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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 paper studies Minimum-Component Distance Geometric Inhomogeneous Random Graphs (MCD-GIRGs) in dimension d≥2, with power-law exponent τ∈(2,3) and parameter α>1. It claims that the subgraph induced by vertices of weight at least c′log^γ n, for γ>1/(3−τ), is with high probability an expander: every set S of size at most |V′|/2 has external neighbourhood at least ε|S|, and the expansion factor is ω(1) for |S|=o(|V′|). The formal main theorem (Theorem 2) proves the expansion bound for subsets S with |S|≤ε|V′|, with an explicit factor ε·min{log^{γ(3−τ)}n, (|V′|/|S|)^{1−1/c_d}}, via two propositions: a strip-counting union bound (Proposition 1) and a lower bound on strip-local neighbourhoods (Proposition 2). The linear-size case ε|V′|<|S|≤|V′|/2 is deferred to an unproved adaptation of a result from [23]. The proof is presented in Appendix A.2 and relies on earlier lemmas from [18] and [4]. The paper also claims the result transfers to subgraphs induced by degree thresholds, and discusses algorithmic implications of expander structure.

Significance. If completed, the result is significant: it would provide the first demonstration that a non-metric, 'match in some dimension' geometry can produce strong expander behaviour in the high-degree core, in sharp contrast to Euclidean GIRGs, which have small separators at all scales. The strip-based mechanism is natural, and the explicit trade-off between γ and the expansion factor is informative. The paper is careful about the tightness of γ and about transferring the result from weights to degrees, and it contains no fitted parameters or ad-hoc axioms; the sublinear case is supported by a detailed counting argument that appears coherent. The main open direction (generalization to all BDF-GIRGs) is honestly stated. The value of the paper is currently reduced by the missing proof for linear-size sets, because the full expander property is what the abstract, Theorem 1, and the applications require.

major comments (3)
  1. [Section 2.2, paragraph after Theorem 2] The abstract and Theorem 1 claim that the induced subgraph G′ is an expander, i.e. that every S⊆V′ with |S|≤|V′|/2 has external neighbourhood at least ε|S|. Theorem 2, however, only proves this for |S|≤ε|V′|; the proof in Appendix A.2 (via Propositions 1 and 2) does not cover the range ε|V′|<|S|≤|V′|/2. The paragraph after Theorem 2 dismisses this range with the assertions that the linear-size result of [23] 'can readily be adapted' to G′ and that 'Since G′ is connected, we may also drop the condition on the giant component.' No proof of either assertion is given, and connectivity of G′ is not established by the paper; the condition γ>1/(3−τ) only rules out isolated vertices whp, not multiple components. This gap is load-bearing: constant expansion for constant-fraction sets is part of the headline expander claim, and the paper's own propositions cannot deliver it, since at |S|=|V′|/2 the bound k·log^{γ(3−τ)}n is only O(|V′|) with a hidden constant that may be far below 1, making the subtraction of |S| problematic. Please supply the full adaptation (stating precisely which theorem of [23] is transferred and how) or weaken the abstract, Theorem 1, and the applications accordingly.
  2. [Appendix A.2, Proposition 1 and proof of Theorem 2] Proposition 1 is stated pointwise: for each s=ω(1) it asserts the existence of k and a whp event, and the event may depend on s. In the proof of Theorem 2 the proposition is applied to every S with C≤|S|≤ε|V′| simultaneously, but no union bound over s is written. The proof of Proposition 1 actually gives pk,s≤exp(−Ω(s)) once the constants c1,c2,cd are fixed, so a union bound over s (with C chosen large enough) would close the gap; however, this is not stated in the manuscript. As written, the universal 'with high probability' conclusion for all S is not formally derived from the pointwise statement. Please state and prove a uniform version of Proposition 1 (with the union bound over s) or adjust the proof of Theorem 2 accordingly.
  3. [Appendix A.2, proof of Theorem 2, degree-threshold part] In the proof of the degree-threshold version, the inclusions are reversed: from the two preceding sentences one obtains V2⊆VH′⊆V1, but the text states 'V1⊆VH′⊆V2'. Moreover, the sentence 'it suffices to show that every subset S⊆V2 has at least f·|S| neighbours in V1' does not by itself imply the claimed external-neighbourhood bound for H′: neighbours counted in V1 may lie inside S, and the subtraction of |S| (used in the weight-based case) is not accounted for here. These points need to be fixed; with the correct inclusions and a suitably enlarged neighbour-count bound before subtracting |S|, the argument should go through.
minor comments (5)
  1. [Proposition 1, statement] Typo: 'c′ > 0 be consatnts' should read 'c′ > 0 be constants'.
  2. [Proposition 2, statement] The phrase 'let γ > 1/(3−τ) and c′ > 0 c′2 > c′1 > 0 be constants' is malformed; it should introduce the constants c′1, c′2 separately.
  3. [Proposition 1 proof] The expression for the probability that a fixed S lies in the union of k strips is written as '(k⌊n/ log2γn⌋−1)s'; it should be (k/⌊n/log^{2γ}n⌋)^s, with the exponent and division displayed correctly.
  4. [Section 2.2, paragraph after Theorem 2] The sentence 'The larger the weights we consider, the better the expansion factor log^{γ(3−τ)}n becomes' is somewhat loose, since the theorem's stated expansion factor is the same for the interval-restricted subgraph; consider rephrasing.
  5. [Theorem 2 proof] The phrase 'The proof for vertices of degrees in the interval [...] is analogously' should be 'is analogous'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's novel sublinear expansion result is derived self-containedly from Propositions 1 and 2; the linear-size case is inherited from prior work, not re-derived by definition.

full rationale

The derivation chain of the paper's novel contribution—expansion for |S|=o(|V'|) and |S|≤ε|V'|—is self-contained: Theorem 2 is proved from Propositions 1 and 2, whose proofs are counting arguments and local degree-concentration arguments using only the model definition, Chernoff bounds, and the standard volume scaling V_min(r)=Θ(r) of MCD-GIRGs. The expectation E[deg_{G'}(v)] = Θ(w_v log^{γ(2−τ)}n) is quoted from [18, Lemma 10], but this is a prior published technical lemma (with partially overlapping authorship) that is not the target result; it is independently proved relative to the same model and is used as a standard input. The only caveat is that the formal statement of Theorem 2 covers |S|≤ε|V'|, while the full expander claim for |S|∈[ε|V'|, |V'|/2] is delegated to [23] with the unproved assertion that the linear-size separator result 'can readily be adapted' from the giant component of G to G'. This is a correctness/completeness gap, not circularity: [23] is prior independent work, and the novel sublinear result does not reduce to it. No equation is defined in terms of the target result, no fitted parameter is renamed a prediction, and no load-bearing claim is forced by a self-citation chain.

Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

No empirical free parameters appear in the proof; model parameters tau, d, alpha, gamma, and c' are part of the model definition, and the proof introduces only existential constants epsilon, c1, c2, and c_d whose values are constrained by inequalities, not fitted to data. The main external inputs are prior theorems on GIRG degree expectations and separators, listed above.

assumptions (7)
  • domain assumption MCD-GIRG generative model: i.i.d. power-law weights, uniform torus positions, independent edges with probability Theta(min{w_u w_v/(n Vmin(distance)),1}^alpha) (Definition 2).
    The theorem is conditional on this model, which is proposed in prior work [4] and not derived here.
  • domain assumption Expected degree of a weight-w vertex in the induced subgraph on vertices of weight at least c log^gamma n is Theta(w log^{gamma(2-tau)} n) ([18, Lemma 10]).
    Invoked in Section 2.2 and Lemma 3; it determines the sharp threshold gamma > 1/(3-tau) and the polylogarithmic expansion factor. Not proved in this paper.
  • domain assumption Linear-sized vertex subsets of the MCD-GIRG giant component have external neighbourhoods of size Theta(n) ([23]).
    Used after Theorem 2 to extend expansion to sets with epsilon |V'| <= |S| <= |V'|/2; the paper says the result can be readily adapted but gives no proof.
  • domain assumption The MCD volume function satisfies Vmin(r) = Theta(r) as r -> 0 in dimension d >= 2.
    Used in Proposition 2 to obtain Omega(1) edge probability for vertices in the same strip; follows from the cross-shaped geometry and is cited from [23].
  • domain assumption Expected degree of a vertex of weight w in the full MCD-GIRG is Theta(w) ([4, Lemma 4.3]).
    Used in Lemma 3 to relate degree thresholds to weight thresholds so the theorem transfers from weight-defined to degree-defined subgraphs.
  • standard math Chernoff bound for sums of independent indicator variables.
    Used throughout Lemma 3 and Proposition 2 for concentration of degrees and neighbour counts.
  • standard math Stirling approximation for binomial coefficients.
    Used in Proposition 1 to bound the union bound probability over strip choices and vertex sets.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Expanders in Models of Social Networks." pith.science (2026). https://pith.science/paper/6LJFVHOH

@misc{pith2026250619485,
  author       = {Pith},
  title        = {Pith review of: Expanders in Models of Social Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6LJFVHOH}},
  note         = {Machine review of arXiv:2506.19485}
}
abstract

A common model for social networks are Geometric Inhomogeneous Random Graphs (GIRGs), in which vertices draw a random position in some latent geometric space, and the probability of two vertices forming an edge depends on their geometric distance. The geometry may be modelled in two ways: either two points are defined as close if they are similar in all dimensions, or they are defined as close if they are similar in some dimensions. The first option is mathematically more natural since it can be described by metrics. However, the second option is arguably the better model for social networks if the different dimensions represent features like profession, kinship, or interests. In such cases, nodes already form bonds if they align in some, but not all dimensions. For the first option, it is known that the resulting networks are poor expanders. We study the second option in the form of Minimum-Component Distance GIRGs, and find that those behave the opposite way for dimension $d\ge 2$, and that they have strong expanding properties. More precisely, for a suitable constant $C>0$, the subgraph induced by vertices of (expected) degree at least $(\log n)^C$ forms an expander. Moreover, we study how the expansion factor of the resulting subgraph depends on the choice of $C$, and show that this expansion factor is $\omega(1)$ except for sets that already take up a constant fraction of the vertices. This has far-reaching consequences, since many algorithms and mixing processes are fast on expander graphs.

Figures

Figures reproduced from arXiv: 2506.19485 by the authors.

Figure 1
Figure 1. MCD GIRG (left), with parameters d = 2, n = 500, τ = 2.1, α = 1.5, and its induced subgraph of weights larger than log2 n (right). Larger (and lighter-coloured) disks indicate nodes of larger weight/degree. Disks are only scaled on the left illustration to increase visibility. 1.2 Implications Expander graphs have an astonishing wealth of applications across mathematics and computer science – and consequently, MCD-G… view at source ↗
Figure 2
Figure 2. Left: Balls in the Minimum-Component Distance are cross-shaped. Right: Illustration of the proof technique in two dimensions. The vertices marked by triangles form the adversarially chosen set S for s = 3. Although the adversary has managed to have S only intersect with one vertical strip, it does not achieve k = 1 because S intersects with two horizontal strips. In fact, for s = 3 and k = 1 the claim holds as there… view at source ↗

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 25 canonical work pages

  1. [23]

    Existence of Small Separators Depends on Geometry for Geometric Inhomogeneous Random Graphs

    Lengler, J., Todorovic, L.: Existence of small separators depends on geometry for geometric inhomogeneous random graphs. arXiv preprint arXiv:1711.03814 (2017)

    work page Pith review arXiv 2017

  2. [18]

    Internet Mathematics p

    Koch, C., Lengler, J.: Bootstrap percolation on geometric inhomogeneous random graphs. Internet Mathematics p. 05c80 (2021)

    work page 2021

  3. [4]

    Advances in Applied Probability p

    Bringmann, K., Keusch, R., Lengler, J.: Average distance in a general class of scale-free networks. Advances in Applied Probability p. 1–36 (2024)

    work page 2024

  4. [1]

    ACM Transactions on Algorithms20(1) (Jan 2024)

    Bläsius, T., Fischbeck, P.: On the external validity of average-case analyses of graph algorithms. ACM Transactions on Algorithms20(1) (Jan 2024)

    work page 2024

  5. [2]

    Network Science 10(4), 361–380 (2022)

    Bläsius, T., Friedrich, T., Katzmann, M., Meyer, U., Penschuck, M., Weyand, C.: Efficiently generating geometric inhomogeneous and hyperbolic random graphs. Network Science 10(4), 361–380 (2022)

    work page 2022

  6. [3]

    Theoretical Computer Science760, 35–54 (2019)

    Bringmann, K., Keusch, R., Lengler, J.: Geometric inhomogeneous random graphs. Theoretical Computer Science760, 35–54 (2019)

    work page 2019

  7. [5]

    Journal of Computer and System Sciences 125, 59–105 (2022)

    Bringmann, K., Keusch, R., Lengler, J., Maus, Y., Molla, A.R.: Greedy routing and the algorithmic small-world phenomenon. Journal of Computer and System Sciences 125, 59–105 (2022)

    work page 2022

  8. [6]

    2020 IEEE 61st Annual Symposium on Foundations of Computer Science (FOCS) pp

    Chuzhoy, J., Gao, Y., Li, J., Nanongkai, D., Peng, R., Saranurak, T.: A determin- istic algorithm for balanced cut with applications to dynamic connectivity, flows, and beyond. 2020 IEEE 61st Annual Symposium on Foundations of Computer Science (FOCS) pp. 1158–1167 (2019)

    work page 2019

Show all 27 references
  1. [7]

    In: International Conference on Complex Networks

    Dayan, B., Kaufmann, M., Schaller, U.: Expressivity of geometric inhomogeneous random graphs—metric and non-metric. In: International Conference on Complex Networks. pp. 85–100. Springer (2024)

    2024

  2. [8]

    In: Annales de l’IHP Probabilités et statistiques

    Deijfen, M., Van der Hofstad, R., Hooghiemstra, G.: Scale-free percolation. In: Annales de l’IHP Probabilités et statistiques. vol. 49, pp. 817–838 (2013)

    2013

  3. [9]

    In: Algorithms and Models for the Web Graph: 14th International Workshop, WAW 2017, Toronto, ON, Canada, June 15–16, 2017, Revised Selected Papers 14

    Feldman, J., Janssen, J.: High degree vertices and spread of infections in spatially modelled social networks. In: Algorithms and Models for the Web Graph: 14th International Workshop, WAW 2017, Toronto, ON, Canada, June 15–16, 2017, Revised Selected Papers 14. pp. 60–74. Spri...

    2017

  4. [10]

    Fountoulakis, N., Panagiotou, K., Sauerwald, T.: Ultra-fast rumor spreading in so- cial networks. pp. 1642–1660 (01 2012). https://doi.org/10.1137/1.9781611973099. 130

    2012 doi

  5. [11]

    In: Proceedings of the 2018 ACM Symposium on Principles of Distributed Computing

    Ghaffari, M., Lengler, J.: Nearly-tight analysis for 2-choice and 3-majority con- sensus dynamics. In: Proceedings of the 2018 ACM Symposium on Principles of Distributed Computing. p. 305–313. PODC ’18, Association for Computing Ma- chinery, New York, NY, USA (2018)

    2018

  6. [12]

    American Journal of Sociology78(6), 1360–1380 (1973) 14 M

    Granovetter, M.S.: The strength of weak ties. American Journal of Sociology78(6), 1360–1380 (1973) 14 M. Kaufmann, J. Lengler, U. Schaller, K. Sturm

    1973

  7. [13]

    Hoory, S., Linial, N., Wigderson, A.: Expander graphs and their applications. Bull. Amer. Math. Soc.43, 439–561 (2006)

    2006

  8. [14]

    Chaos, Solitons & Fractals139, 109965 (2020)

    Jorritsma, J., Hulshof, T., Komjáthy, J.: Not all interventions are equal for the height of the second peak. Chaos, Solitons & Fractals139, 109965 (2020)

    2020

  9. [15]

    Jorritsma, J., Komjáthy, J., Mitsche, D.: Cluster-size decay in supercritical kernel- based spatial random graphs (2024), https://arxiv.org/abs/2303.00724

    2024 arXiv

  10. [16]

    arXiv preprint arXiv:2408.01268 (2024)

    Kaufmann, M., Lakis, K., Lengler, J., Ravi, R.R., Schaller, U., Sturm, K.: Rumour spreading depends on the latent geometry and degree distribution in social network models. arXiv preprint arXiv:2408.01268 (2024)

    2024 arXiv

  11. [17]

    In: The 13th International Conference on Complex Networks and their Applications (2024)

    Kaufmann, M., Ravi, R.R., Schaller, U.: Sublinear cuts are the exception in BDF- GIRGs. In: The 13th International Conference on Complex Networks and their Applications (2024)

    2024

  12. [19]

    Electronic Journal of Probability 29, 1 – 48 (2024)

    Komjáthy, J., Lapinskas, J., Lengler, J., Schaller, U.: Polynomial growth in degree- dependent first passage percolation on spatial random graphs. Electronic Journal of Probability 29, 1 – 48 (2024)

    2024

  13. [20]

    Stochastic Processes and their Applica- tions 130(3), 1309–1367 (2020)

    Komjáthy, J., Lodewijks, B.: Explosion in weighted hyperbolic random graphs and geometric inhomogeneous random graphs. Stochastic Processes and their Applica- tions 130(3), 1309–1367 (2020)

    2020

  14. [21]

    Komjáthy, J., Lapinskas, J., Lengler, J., Schaller, U.: Four universal growth regimes in degree-dependent first passage percolation on spatial random graphs i (2024)

    2024

  15. [22]

    Krioukov, D., Papadopoulos, F., Kitsak, M., Vahdat, A., Boguná, M.: Hyperbolic geometryofcomplexnetworks.PhysicalReviewE—Statistical,Nonlinear,andSoft Matter Physics 82(3), 036106 (2010)

    2010

  16. [24]

    Bulletin of the American Mathematical Society49 (05 2011)

    Lubotzky, A.: Expander graphs in pure and applied mathematics. Bulletin of the American Mathematical Society49 (05 2011)

    2011

  17. [25]

    In: The 43rd Annual IEEE Symposium on Foundations of Computer Science, 2002

    Racke, H.: Minimizing congestion in general networks. In: The 43rd Annual IEEE Symposium on Foundations of Computer Science, 2002. Proceedings. pp. 43–52 (2002)

    2002

  18. [26]

    Random Structures & Algorithms65(4), 613–643 (2024)

    Shimizu, N., Shiraga, T.: Quasi-majority functional voting on expander graphs. Random Structures & Algorithms65(4), 613–643 (2024)

    2024

  19. [27]

    −s d 1− k s log ⌊n/ log2γn⌋ k − log |V′| s − 2d # = exp

    Voitalov, I., van der Hoorn, P., van der Hofstad, R., Krioukov, D.: Scale-free net- works well done. Phys. Rev. Res.1, 033034 (Oct 2019) A Appendix A.1 Tools In this short section, we provide two technical tools that will come in handy during the proof, starting with the class...

    2019

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.