← All problems

Quartic fluctuation conjecture for edge–triangle exponential random graph models

Let QnQ_n denote the edge density in the dense edge–triangle exponential random graph model at its critical parameters (αc,hc)(\alpha_c,h_c). The conjecture asserts that n(Qn−2/3)\sqrt{n}(Q_n-2/3) converges in distribution to a random variable ZZ with density f(z)=Cexp⁡{−81z4/64}f(z)=C\exp\{-81z^4/64\} on R\mathbb{R}, where CC is the normalizing constant.

References

Source: arXiv

Progress summary

Refreshed
Open · solution reported

An unrefereed preprint claims the conjecture is false for the original graph model, with Gaussian behavior at the fixed critical point and quartic behavior only under a tuned perturbation.

The conjecture, attributed here to Bianchi, Collet, and Magnanini (2024), predicted quartic fluctuations for the edge–triangle model at its critical parameters. The new work argues that this prediction applies only to a mean-field approximation, not to the original graph law.

Known results

  • Bianchi, Collet, and Magnanini (2024): conjectured a quartic limit for the mean-field approximation, with fluctuations on scale n\sqrt{n} and density proportional to exp⁡{−81z4/64}\exp\{-81z^4/64\}.

October 2026 preprint

Liu, Shao, Shi, and Xu claim that at fixed critical parameters, n2/3(Q−EQ)n^{2/3}(Q-\mathbb{E}Q) converges to a Gaussian law, while quartic limits occur only on a tuned perturbation line. They also claim quantitative error bounds, but the preprint has no independent mathematical assessment in the retrieved sources.

Current status (as of November 2026): The quartic conjecture for the original model is claimed disproved, but the new Gaussian and tuned-quartic asymptotics remain unverified; the mean-field conjecture is distinct.

Sources

Solutions 0

No solutions have been posted yet.