Gaussian generating function (eq. 1.41)
ProvedDeepLearningTheory.gaussian_generating_functionLet be a symmetric positive-definite real matrix, with inverse , and let be a source. Then the generating function (1.38) has the closed form
Its Taylor coefficients at are the Gaussian moments, which is how the book derives Wick's theorem.
import Mathlib import Definitions.Def_DLT_GaussianIntegrals open MeasureTheory
namespace DeepLearningTheory
theorem gaussian_generating_function {N : ℕ} (K : Matrix (Fin N) (Fin N) ℝ)
(hK : K.PosDef) (J : Fin N → ℝ) :
∫ z : Fin N → ℝ, Real.exp (-(1 / 2) * gaussQuadForm K z + ∑ μ : Fin N, J μ * z μ)
= Real.sqrt ((2 * Real.pi) ^ N * K.det) *
Real.exp ((1 / 2) * ∑ μ : Fin N, ∑ ν : Fin N, J μ * K μ ν * J ν) := by sorry
end DeepLearningTheoryRead-back
What the Lean code literally says, in plain math · Aristotle (Harmonic) - non-blind, same agent as the drafter
Non-blind read-back. This read-back was written by the same agent that drafted the Lean statement, with full knowledge of the source text and of the intended meaning. It is not independent testimony and must not be mistaken for a blind audit; reviewers should compare the Lean code against the source themselves.
For every , every positive-definite (hence symmetric) real matrix and every , the Lebesgue integral over of
equals . The integral is a Bochner integral. No restriction is placed on .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.