p. 18 — Gaussian Crofton embedding: ‖ϕ(u) − ϕ(u′)‖₂² = ‖ϕ(u) − ϕ(u′)‖₁ = ‖u − u′‖₂
ProvedDistCov.Hilbert.gaussEmb_isometrycroftongaussianisometric-embeddingnegative-typep2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be the IID standard normal law on , Lebesgue measure on , , and for let
Then for all the functions and are integrable with respect to , and
Thus embeds isometrically into : it is the embedding through which the paper computes on .
Formalization Note. is a plain real function on ; the squared norm and the norm of the paper are the two integrals. The integrability clauses are added so the integrals are not Lean's junk value . Indices start at .
Preamble
import Mathlib import Definitions.Def_DistCov_Hilbert_Setting
Formal statement
namespace DistCov.Hilbert
open MeasureTheory ProbabilityTheory
theorem gaussEmb_isometry (u u' : DistCov.Indep.L2N) :
Integrable (fun p => (gaussEmb u p - gaussEmb u' p) ^ 2) (rho.prod volume) ∧
∫ p, (gaussEmb u p - gaussEmb u' p) ^ 2 ∂(rho.prod volume) = ‖u - u'‖ ∧
Integrable (fun p => |gaussEmb u p - gaussEmb u' p|) (rho.prod volume) ∧
∫ p, |gaussEmb u p - gaussEmb u' p| ∂(rho.prod volume) = ‖u - 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, display defining ϕ(u) and the sentence after it
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.