Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Rotational invariance of the standard Gaussian matrix: UΓV∼ΓU\Gamma V\sim\GammaUΓV∼Γ for orthogonal U,VU,VU,V

Proved
GaussianMatrix.rotation_invariance

by tc · Oct 9, 2026 · Mathlib 0df444a (Lean v4.33.1)

gaussian-matrixprobabilityrandom-matrix-theoryrandomized-numerical-linear-algebra

Let Γ∈Rp×m\Gamma\in\mathbb R^{p\times m}Γ∈Rp×m be a standard Gaussian matrix (independent N(0,1)\mathcal N(0,1)N(0,1) entries, law γp,m\gamma_{p,m}γp,m​), and let U∈Rp×pU\in\mathbb R^{p\times p}U∈Rp×p and V∈Rm×mV\in\mathbb R^{m\times m}V∈Rm×m be orthogonal matrices, UTU=IpU^{\mathsf T}U=I_pUTU=Ip​ and VTV=ImV^{\mathsf T}V=I_mVTV=Im​. Then the matrix UΓVU\Gamma VUΓV is again a standard Gaussian matrix:

U Γ V ∼ γp,m.U\,\Gamma\,V\ \sim\ \gamma_{p,m}.UΓV ∼ γp,m​.

This is the basic symmetry of the Gaussian ensemble. It lets one reduce computations involving fixed matrices SΓTS\Gamma TSΓT to the case of diagonal SSS and TTT, and it is the engine behind every conditioning argument in the analysis of randomized sketching: multiplying a Gaussian sketch by an orthogonal factor of the input matrix does not change its law.

Formalization Note. The statement is an equality of measures: the pushforward of γp,m\gamma_{p,m}γp,m​ under G↦UGVG\mapsto UGVG↦UGV equals γp,m\gamma_{p,m}γp,m​.

Preamble
import Definitions.Def_GaussianMatrix_basic

open MeasureTheory ProbabilityTheory
open scoped Matrix
Formal statement
namespace GaussianMatrix
theorem rotation_invariance {p m : ℕ} (U : Matrix (Fin p) (Fin p) ℝ) (V : Matrix (Fin m) (Fin m) ℝ)
    (hU : Uᵀ * U = 1) (hV : Vᵀ * V = 1) :
    Measure.map (fun G : Fin p → Fin m → ℝ => Matrix.of.symm (U * Matrix.of G * V))
      (gaussianMatrix p m) = gaussianMatrix p m := by sorry
end GaussianMatrix
Source
N. Halko, P.-G. Martinsson, J. A. Tropp, *Finding structure with randomness: probabilistic algorithms for constructing approximate matrix decompositions*, SIAM Review 53(2), 2011, https://arxiv.org/abs/0909.4061 (v2), §10.1 p. 55: "The distribution of a standard Gaussian matrix is rotationally invariant: If U and V are orthonormal matrices, then U*GV also has the standard Gaussian distribution."
Read-back

What the Lean code literally says, in plain math · claude-fable-5-1

Let p,mp, mp,m be natural numbers (implicit; either may be 000), let U∈Rp×pU \in \mathbb{R}^{p \times p}U∈Rp×p and V∈Rm×mV \in \mathbb{R}^{m \times m}V∈Rm×m be real matrices, and assume

UTU=Ip,VTV=ImU^{\mathsf T} U = I_p, \qquad V^{\mathsf T} V = I_mUTU=Ip​,VTV=Im​

(each has orthonormal columns; since they are square this means each is an orthogonal matrix). Let γp,m\gamma_{p,m}γp,m​ denote the law of a p×mp \times mp×m array with independent N(0,1)\mathcal{N}(0,1)N(0,1) entries, i.e. the product measure ⨂i<p⨂j<mN(0,1)\bigotimes_{i<p}\bigotimes_{j<m}\mathcal{N}(0,1)⨂i<p​⨂j<m​N(0,1) on real p×mp \times mp×m arrays. The statement asserts the equality of measures on the space of p×mp \times mp×m real arrays

(G↦U G V)# γp,m  =  γp,m,\big(G \mapsto U\,G\,V\big)_{\#}\,\gamma_{p,m} \;=\; \gamma_{p,m},(G↦UGV)#​γp,m​=γp,m​,

i.e. the pushforward (image measure) of γp,m\gamma_{p,m}γp,m​ under the map sending the array GGG (viewed as a matrix) to the matrix product UGVU G VUGV (viewed again as an array) is γp,m\gamma_{p,m}γp,m​ itself. Equality of measures means equality on every measurable set of the product sigma-algebra. Caveats on the pushforward: Mathlib's image measure is defined to be the zero measure if the map is not almost-everywhere measurable; here the map is linear on a finite-dimensional space, hence measurable, so this convention is not triggered. Edge cases: if p=0p = 0p=0 or m=0m = 0m=0 the array space is a single point, both sides are the unit point mass, and UUU or VVV is an empty matrix for which the orthogonality hypothesis holds automatically.

Human review
  • Endorsed by Shuze Chen · Oct 9, 2026

    Confirmed by the moderator at approval.

  • Endorsed by tc · Oct 9, 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