AFTD Auto-Formalizing Theoretical Domains

map_const_mul_gaussian

theoremverifiedimported Lean

Scaling a centered real Gaussian. Let $X\colon \Omega \to \bbr$ be a measurable random variable on a measure space $(\Omega, \mu)$ whose law is the centered Gaussian $N(0, \sigma_0)$ with variance $\sigma_0 \ge 0$. Then for every scalar $c \in \bbr$, the scaled variable $c\,X$ has law \[ N\!\bigl(0,\; c^2 \sigma_0\bigr), \] the centered Gaussian with variance $c^2 \sigma_0$.

Statement

set_option relaxedAutoImplicit false in
set_option autoImplicit false in
open MeasureTheory ProbabilityTheory Real NNReal Matrix Finset in
variable {d k : ℕ} in
lemma map_const_mul_gaussian
    {Ω : Type*} [MeasurableSpace Ω] {μ : Measure Ω}
    {X : Ω → ℝ} {σ₀ : NNReal}
    (hX_meas : Measurable X)
    (hX : Measure.map X μ = gaussianReal 0 σ₀)
    (c : ℝ) :
    Measure.map (fun ω => c * X ω) μ =
      gaussianReal 0 (⟨c ^ 2, sq_nonneg _⟩ * σ₀)
Source
TCSlib, `map_const_mul_gaussian`
Lean proof
By Ganesh Sankar, from github.com/Shilun-Allan-Li/tcslib at c8a02591b1e5 (Copyright (c) 2026 Ganesh Sankar. All rights reserved, Apache-2.0). Compiled here against leanprover/lean4:v4.33.1, Mathlib v4.33.1; written for leanprover/lean4:v4.25.0.
What was changed in map_const_mul_gaussian
--- TCSlib/LearningTheory/JohnsonLindenstrauss/RowDistribution.lean:96 (map_const_mul_gaussian)
+++ as compiled here
@@ -1,6 +1,5 @@
-/-- If a measurable real random variable `X` has law `N(0, σ₀)` (Mathlib's `gaussianReal`
-with variance parameter `σ₀`), then for any real `c` the scalar multiple `c·X` has law
-`N(0, c²·σ₀)`. This is the scaling step of [DG03, §2] (a linear image of a centered
-Gaussian is a centered Gaussian); it is an auxiliary restatement of Mathlib's
-`gaussianReal_map_const_mul` that is easier to apply. -/
+set_option relaxedAutoImplicit false in
+set_option autoImplicit false in
+open MeasureTheory ProbabilityTheory Real NNReal Matrix Finset in
+variable {d k : ℕ} in
 lemma map_const_mul_gaussian
@@ -16,2 +15,3 @@
     gaussianReal_map_const_mul]
-  simp
+  simp only [mul_zero]
+  rfl
Verified
16 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Used by 2
  • sum_scaled_iid_gaussian_mapWeighted sum of independent Gaussians is Gaussian. Let $(\Omega,\mu)$ be a probability space, and let $(Y_j)_{j}$ be a family of…
  • jl_concentration_single_via_chi_squaredJohnson–Lindenstrauss concentration for a single vector. Let $k \ge 1$, and let $A$ be a random $k \times d$ real matrix on a probability…

Lean source view module on GitHub

set_option relaxedAutoImplicit false in
set_option autoImplicit false in
open MeasureTheory ProbabilityTheory Real NNReal Matrix Finset in
variable {d k : ℕ} in
/-- If a measurable real random variable `X` has law `N(0, σ₀)` (Mathlib's `gaussianReal` with variance parameter `σ₀`), then for any real `c` the scalar multiple `c·X` has law `N(0, c²·σ₀)`. This is the scaling step of [DG03, §2] (a linear image of a centered Gaussian is a centered Gaussian); it is an auxiliary restatement of Mathlib's `gaussianReal_map_const_mul` that is easier to apply. -/
lemma map_const_mul_gaussian
    {Ω : Type*} [MeasurableSpace Ω] {μ : Measure Ω}
    {X : Ω → ℝ} {σ₀ : NNReal}
    (hX_meas : Measurable X)
    (hX : Measure.map X μ = gaussianReal 0 σ₀)
    (c : ℝ) :
    Measure.map (fun ω => c * X ω) μ =
      gaussianReal 0 (⟨c ^ 2, sq_nonneg _⟩ * σ₀) := by
  have : (fun ω => c * X ω) = (fun r : ℝ => c * r) ∘ X := rfl
  rw [this, ← Measure.map_map (by fun_prop) hX_meas, hX,
    gaussianReal_map_const_mul]
  simp only [mul_zero]
  rfl
See it in the graphDiscuss this resultChallenge it