p. 18 — Z(u) = Σ uₙZₙ converges ρ-a.s., is N(0, ‖u‖₂²), and E|Z(u)| = c‖u‖₂
ProvedDistCov.Hilbert.gaussian_seriesgaussianl2p2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1random-series
Let be the law on of an IID sequence of standard normal variables, , and for write . For every :
- for -a.e. the series converges, to ;
- the law of under is the centered normal law with variance ;
- is -integrable, and
In the language of the paper: is a centered normal random variable with standard deviation , so . This is the probabilistic input of the Gaussian Crofton embedding used to prove Theorem 3.16.
Formalization Note. The index is shifted to start at . is computed in ; clause 1 shows that its junk value on divergent sequences occurs only on a -null set. The normal law of variance is Mathlib's gaussianReal 0 σ²; for it is the Dirac mass at .
Preamble
import Mathlib import Definitions.Def_DistCov_Hilbert_Setting
Formal statement
namespace DistCov.Hilbert
open MeasureTheory ProbabilityTheory
theorem gaussian_series (u : DistCov.Indep.L2N) :
(∀ᵐ w ∂rho, Filter.Tendsto (fun N : ℕ => ∑ n ∈ Finset.range N, u n * w n) Filter.atTop
(nhds (wEval w u))) ∧
rho.map (fun w => wEval w u) = gaussianReal 0 (‖u‖₊ ^ 2) ∧
Integrable (fun w => wEval w u) rho ∧
∫ w, |wEval w u| ∂rho = cGauss * ‖u‖ := by sorry
end DistCov.Hilbert
Source
Lyons, Distance covariance in metric spaces, arXiv:1106.5758 (version of 19 Dec. 2020), p. 18, proof of Theorem 3.16, second paragraph, and 'the series w(u) converges ρ-a.s.' (p. 18, after the second display)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.