Tropp–Webber inverse spectral moment: for ,
ProvedGaussianMatrix.inverse_wishart_spectral_momentLet be a standard Gaussian matrix, let be an integer with , and assume . Then is integrable and
where is the spectral norm. Since , this bounds the -th inverse moment of the smallest singular value of a Gaussian matrix.
The case , , is the spectral-norm input for expected spectral-norm error bounds of the randomized SVD and of truncated sketching algorithms, where it is combined with the Frobenius inverse moment . The proof in the source starts from Edelman's bound on the density of .
Formalization Note. The bound is stated in the equivalent form , together with integrability of the integrand. The source writes the hypothesis as with the moment order (a typographical variant of ); its proof uses exactly .
import Definitions.Def_GaussianMatrix_basic open MeasureTheory ProbabilityTheory open scoped Matrix
namespace GaussianMatrix
theorem inverse_wishart_spectral_moment {r k p : ℕ} (hp : 1 ≤ p) (hp18 : p ≤ 18)
(hrk : r + 2 * p ≤ k) :
Integrable (fun G : Fin r → Fin k → ℝ => specNorm (Matrix.of G * (Matrix.of G)ᵀ)⁻¹ ^ p)
(gaussianMatrix r k) ∧
∫ G, specNorm (Matrix.of G * (Matrix.of G)ᵀ)⁻¹ ^ p ∂(gaussianMatrix r k)
≤ (Real.exp 1 ^ 2 * ((k : ℝ) + r) / (2 * ((k : ℝ) - r) ^ 2)) ^ p := by sorry
end GaussianMatrixRead-back
What the Lean code literally says, in plain math · claude-fable-5-1
Throughout, denotes the measure called gaussianMatrix r k: the product measure on the function space in which every one of the coordinates is an independent standard normal ; it is a probability measure, and a sample is viewed as the real matrix with entries . Matrix inversion is Mathlib's total inverse on square real matrices: it is the usual inverse when the determinant is nonzero and is the zero matrix otherwise. The bundle's specNorm of a real matrix is Mathlib's operator norm
i.e. the largest singular value of (equal to for the zero matrix and for any matrix with an empty index set).
Statement. Let be natural numbers with , , and . Let be the Gram matrix of a sample . The theorem asserts the conjunction of:
-
Integrability. The function (natural-number power of the spectral norm of the inverse Gram matrix) is integrable with respect to .
-
Upper bound.
where , are cast to , and the powers are natural-number powers. Under , , so the division is ordinary.
Edge cases. This is a one-sided inequality (), not an equality. The upper bound is an explicit hypothesis: nothing is claimed for , and nothing is claimed for . The integral is Lean's real Bochner integral (which is for non-integrable integrands; part 1 asserts integrability). On the null set where is singular, and the integrand is . When , is the matrix, its spectral norm is , so the left side is (as ) and the right side is ; the hypothesis then reads .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.