Edge universality for non-trivial random regular graphs

At least 4 years old · documented by

Let GG be a uniformly random dd-regular graph on NN vertices, with adjacency eigenvalues d(N−1)N\boldsymbol{\textstyle\frac{d(N-1)}{N}} excluded from consideration, and let λ2\boldsymbol{\lambda_2} denote its second eigenvalue. Assume 3⩽d⩽N−43\leqslant d\leqslant N-4. The notation TW1\mathrm{TW}_1 denotes the Tracy–Widom distribution for GOE.

Edge universality conjecture. We have

N2/3(λ2(d−1)(N−d−2)/N−2)⟶dTW1,N^{2/3}\bigg(\frac{\lambda_2}{\sqrt{(d-1)(N-d-2)/N}}-2\bigg) \overset{d}{\longrightarrow} \mathrm{TW}_1,

where analogous results also hold for other non-trivial extreme eigenvalues.

The conjecture proposes edge universality for all non-trivial random dd-regular graphs, extending the proved dense regime and respecting the degree symmetry d⟷N−d−1d\longleftrightarrow N-d-1. The authors indicate that the range d⩽N2/9d\leqslant N^{2/9}, especially fixed dd, is difficult, while suggesting that optimal edge rigidity and universality may be provable for N2/9≪d⩽N/2N^{2/9}\ll d\leqslant N/2.

References

Primary source

Yukun He, “Spectral gap and edge universality of dense random regular graphs”, arXiv:2203.07317 (2024).

Additional references

2 papers in this index state this conjecture (2021–2022). The statement above is taken from the most recent of them; the others are arXiv:2108.11634.

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.