p. 18 — every (real) Hilbert space has negative type, indeed strict negative type
ProvedDistCov.Hilbert.hilbert_strict_negTypehilbert-spacemetric-spacenegative-typep2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be a real Hilbert space with distance . Then:
- has negative type: for all , all and all real with ,
- has strict negative type: if moreover the points are distinct and the double sum equals , then .
Clause 2 is the restriction of strong negative type to measures of finite support (p. 11): for finitely supported , forces . The paper derives the statement from the Euclidean case; it supplies the negative-type half of Theorem 3.16.
Formalization Note. Hilbert spaces are real (Errata (viii)); a complex Hilbert space has the same distances as its realification. Strict negative type is stated directly on weights at distinct points, which is the finite-support case of the condition on p. 11.
Preamble
import Mathlib import Definitions.Def_DistCov_Hilbert_Setting
Formal statement
namespace DistCov.Hilbert
open MeasureTheory ProbabilityTheory
theorem hilbert_strict_negType (H : Type*) [NormedAddCommGroup H] [InnerProductSpace ℝ H]
[CompleteSpace H] :
DistCov.Indep.NegType H ∧
∀ (n : ℕ) (x : Fin n → H) (α : Fin n → ℝ), Function.Injective x → ∑ i, α i = 0 →
∑ i, ∑ j, α i * α j * dist (x i) (x j) = 0 → α = 0 := by sorry
end DistCov.Hilbert
Source
Lyons, Distance covariance in metric spaces, arXiv:1106.5758 (version of 19 Dec. 2020), p. 18, paragraph before Theorem 3.16; strict negative type defined on p. 11; Errata (viii), p. 27
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.