Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Eq. (6.26) — the nnn-dimensional Hermite kernel as one exponential (as printed)

Disproved
NatureSLT.Kernels.hermite_kernel_multidim

by Shuze Chen · Oct 20, 2026 · Mathlib 0df444a (Lean v4.33.1)

hermite-polynomialskernelsstatistical-learning-theorytensor-product

Let n≥0n \ge 0n≥0, 0≤q<10 \le q < 10≤q<1, and let x=(x1,…,xn)x = (x^1, \dots, x^n)x=(x1,…,xn) and y=(y1,…,yn)y = (y^1, \dots, y^n)y=(y1,…,yn) be vectors of Rn\mathbb R^nRn, with inner product (x∗y)=∑ixiyi(x * y) = \sum_i x^i y^i(x∗y)=∑i​xiyi and Euclidean norm ∣x−y∣|x - y|∣x−y∣. The second equality of (6.26) claims

∏i=1n1π(1−q2)exp⁡{2xiyiq1+q−(xi−yi)2q21−q2}=1(1−q2)n/2exp⁡{2(x∗y)q1+q−∣x−y∣2q21−q2}.\prod_{i=1}^{n} \frac{1}{\sqrt{\pi(1 - q^2)}} \exp\left\{\frac{2x^iy^iq}{1 + q} - \frac{(x^i - y^i)^2 q^2}{1 - q^2}\right\} = \frac{1}{(1 - q^2)^{n/2}} \exp\left\{\frac{2(x * y)q}{1 + q} - \frac{|x - y|^2 q^2}{1 - q^2}\right\}.i=1∏n​π(1−q2)​1​exp{1+q2xiyiq​−1−q2(xi−yi)2q2​}=(1−q2)n/21​exp{1+q2(x∗y)q​−1−q2∣x−y∣2q2​}.

The left side is the product of nnn one-dimensional Hermite kernels (6.24), which by Theorem 6.1 is the kernel of the tensor-product Hermite basis; the right side is the form the book uses to motivate the semilocal kernels (6.27).

Formalization Note Only the second "=" of (6.26) is stated; the first is Theorem 6.1 applied to (6.24). The range 0≤q<10 \le q < 10≤q<1 is that of (6.24). This statement is believed to be false as printed for n≥1n \ge 1n≥1: the product of the prefactors is (π(1−q2))−n/2(\pi(1-q^2))^{-n/2}(π(1−q2))−n/2, and the factor π−n/2\pi^{-n/2}π−n/2 is missing on the right (see the corrected variant). The power n/2n/2n/2 is a real power of a positive number.

Preamble
import Mathlib
Formal statement
namespace NatureSLT.Kernels

/-- Eq. (6.26), second equality (Vapnik, *The Nature of Statistical Learning Theory*, 2nd ed.,
§6.3.2, p. 193), as printed: for `x, y ∈ ℝⁿ` and `0 ≤ q < 1`,
`Π_{i=1}^n (π(1 - q²))^{-1/2} exp{2x^i y^i q/(1 + q) - (x^i - y^i)² q²/(1 - q²)}
  = (1 - q²)^{-n/2} exp{2(x * y) q/(1 + q) - |x - y|² q²/(1 - q²)}`,
where `x * y` is the inner product and `|x - y|` the Euclidean norm. -/
theorem hermite_kernel_multidim (n : ℕ) (q : ℝ) (hq0 : 0 ≤ q) (hq1 : q < 1)
    (x y : EuclideanSpace ℝ (Fin n)) :
    ∏ i : Fin n, (1 / Real.sqrt (Real.pi * (1 - q ^ 2)) *
        Real.exp (2 * x i * y i * q / (1 + q) - (x i - y i) ^ 2 * q ^ 2 / (1 - q ^ 2))) =
      1 / (1 - q ^ 2) ^ ((n : ℝ) / 2) *
        Real.exp (2 * inner ℝ x y * q / (1 + q) - ‖x - y‖ ^ 2 * q ^ 2 / (1 - q ^ 2)) := by sorry

end NatureSLT.Kernels
Source
Vapnik, The Nature of Statistical Learning Theory, 2nd ed., Springer 2000, p. 193, Eq. (6.26)
Read-back

What the Lean code literally says, in plain math · claude-opus-5-5

Data and hypotheses.

  • n≥0n \ge 0n≥0 is a natural number.
  • qqq is a real number with 0≤q<10 \le q < 10≤q<1.
  • x=(x1,…,xn)x = (x_1, \dots, x_n)x=(x1​,…,xn​) and y=(y1,…,yn)y = (y_1, \dots, y_n)y=(y1​,…,yn​) are points of Euclidean space Rn\mathbb{R}^nRn.
  • ⟨x,y⟩=∑ixiyi\langle x, y \rangle = \sum_i x_i y_i⟨x,y⟩=∑i​xi​yi​ is the standard inner product, and ∥x−y∥\|x - y\|∥x−y∥ is the Euclidean norm.

Statement.

∏i=1n1π(1−q2)exp⁡ ⁣(2xiyiq1+q−(xi−yi)2q21−q2)  =  1(1−q2)n/2exp⁡ ⁣(2⟨x,y⟩q1+q−∥x−y∥2q21−q2).\prod_{i=1}^{n} \frac{1}{\sqrt{\pi(1-q^2)}}\exp\!\Big(\frac{2x_i y_i q}{1+q} - \frac{(x_i-y_i)^2 q^2}{1-q^2}\Big) \;=\; \frac{1}{(1-q^2)^{n/2}}\exp\!\Big(\frac{2\langle x,y\rangle q}{1+q} - \frac{\|x-y\|^2 q^2}{1-q^2}\Big).i=1∏n​π(1−q2)​1​exp(1+q2xi​yi​q​−1−q2(xi​−yi​)2q2​)=(1−q2)n/21​exp(1+q2⟨x,y⟩q​−1−q2∥x−y∥2q2​).

The power (1−q2)n/2(1-q^2)^{n/2}(1−q2)n/2 is a real power with the positive base 1−q2∈(0,1]1 - q^2 \in (0, 1]1−q2∈(0,1]. No factor of π\piπ appears on the right.

Equivalent form. The exponents on the left add up to the exponent on the right, because ⟨x,y⟩=∑ixiyi\langle x, y\rangle = \sum_i x_i y_i⟨x,y⟩=∑i​xi​yi​ and ∥x−y∥2=∑i(xi−yi)2\|x-y\|^2 = \sum_i (x_i - y_i)^2∥x−y∥2=∑i​(xi​−yi​)2. So the statement is equivalent to the equality of constant factors

π−n/2(1−q2)−n/2=(1−q2)−n/2.\pi^{-n/2}(1-q^2)^{-n/2} = (1-q^2)^{-n/2}.π−n/2(1−q2)−n/2=(1−q2)−n/2.

Degenerate cases.

  • Well-definedness. Because q<1q < 1q<1, both 1−q2>01 - q^2 > 01−q2>0 and 1+q>01 + q > 01+q>0, so every division and square root is well defined.
  • n=0n = 0n=0. The left side is the empty product 111. The right side is (1−q2)0⋅e0=1(1-q^2)^0 \cdot e^0 = 1(1−q2)0⋅e0=1. The statement holds.
  • q=0q = 0q=0. The statement reads π−n/2e0=1⋅e0\pi^{-n/2} e^{0} = 1 \cdot e^{0}π−n/2e0=1⋅e0, that is, π−n/2=1\pi^{-n/2} = 1π−n/2=1.
Human review
  • Endorsed by Community (Bot) · Oct 20, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Shuze Chen · Oct 20, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me