Chernoff lower tail of a chi-square variable: for
ProvedGaussianMatrix.chi_square_lower_tailLet be a standard Gaussian vector in , , so that has the chi-square distribution with degrees of freedom. For every with ,
This is the elementary Chernoff (moment-generating-function) estimate for the lower tail of a chi-square variable (Dasgupta–Gupta, Lemma 2.2); it is the probabilistic input for small-ball estimates of distances of a Gaussian vector to a fixed subspace, and hence for the small-ball bound on the smallest singular value of a Gaussian matrix.
Formalization Note. The chi-square variable is realized as under the product of standard normal measures on ; the exponent is a real power. For the right side is and the left side is the probability of the null event .
import Definitions.Def_GaussianMatrix_basic open MeasureTheory ProbabilityTheory open scoped Matrix
namespace GaussianMatrix
theorem chi_square_lower_tail {d : ℕ} (hd : 1 ≤ d) (u : ℝ) (hu : 0 ≤ u) (hud : u ≤ d) :
(Measure.pi fun _ : Fin d => gaussianReal 0 1) {g | ∑ i, g i ^ 2 ≤ u}
≤ ENNReal.ofReal ((Real.exp 1 * u / d) ^ ((d : ℝ) / 2)) := by sorry
end GaussianMatrixRead-back
What the Lean code literally says, in plain math · claude-fable-5-1
Let be a natural number with , and let be a real number with . Let be the product measure on (functions ) under which the coordinates are independent standard normals ; this is a probability measure and is written directly as a product of one-dimensional Gaussians, not via the matrix-law construction. The statement asserts
where , the division is real division (no division by zero since ), the exponent is the real number , and the power is the real power (for base and exponent this is the usual nonnegative power; in particular ). The left side is the outer measure of the set (no measurability assumed); the right side is in and the base is nonnegative so does not truncate. Edge cases: when the right-hand side is , so the claim is that ; when the right-hand side is and the inequality is automatic for a probability measure.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.