2201.05116
POISSON APPROXIMATION AND WEIBULL ASYMPTOTICS IN THE GEOMETRY OF NUMBERS
Michael Björklund, Alexander Gorodnik
correctmedium confidence
- Category
- Not specified
- Journal tier
- Strong Field
- Processed
- Sep 28, 2025, 12:56 AM
- arXiv Links
- Abstract ↗PDF ↗
Audit review
Both the paper and the model prove Poisson convergence of exceedances and Weibull limits under the same separation hypothesis and (a,b,c)-regularity. However, the model’s evaluation of the limiting constant m0 conflicts with the paper’s. With the scaling δn=|Fn|^{-1/a}(log|Fn|)^{-b/a}, the paper derives |Fn|·Vold(C(δn u))→c·u^a·a^b, hence Nn(·;δn u)⇒ Poi((c a^b/(2ζ(d))) u^a) and η̃Fn/δn⇒Wei((2ζ(d))^{1/a}/(c a^b)^{1/a},a) under µd, for d≥3. The model instead claims m0=c(1/a)^b/(2ζ(d)), attributing the discrepancy to “normalization,” but this would contradict the paper’s explicit computation and statements. Apart from this constant, the model’s proof sketch (small-target asymptotics, mixing, factorial moments) aligns conceptually with the paper’s framework, though the paper implements a more robust W-equidistribution criterion and precise second-moment bounds via Siegel–Rogers.
Referee report (LaTeX)
\textbf{Recommendation:} minor revisions \textbf{Journal Tier:} strong field \textbf{Justification:} The paper offers a strong and general method for Poisson approximation in shrinking-target settings on homogeneous spaces, with clear applications to classic problems in the geometry of numbers. The arguments are technically solid and well-structured. A minor clarification concerning the explicit constant in the limiting mean would enhance clarity but does not affect the overall validity.