plackett_luce_surprisal_gaussian_limit_const_one
For the constant weight profile w(x) = 1 on [0,1], the limiting Gaussian distribution gaussianReal with variance Real.toNNReal (plackett_luce_surprisal_variance_profile (fun _ => 1)) degenerates to the Dirac delta measure at 0.
Statement
theorem plackett_luce_surprisal_gaussian_limit_const_one : ProbabilityTheory.gaussianReal 0 (Real.toNNReal (plackett_luce_surprisal_variance_profile (fun _ => 1))) = MeasureTheory.Measure.dirac 0
- Source
- arXiv:2009.09172 (Preference Elicitation as Average-Case Sorting, AAAI 2021)
- Verified
- 15 Nov 2026
- Axioms
Classical.choiceQuot.soundpropext- Built on 2
- plackett_luce_surprisal_variance_profileThe asymptotic surprisal variance V for a weight profile w is the integral over x in [0,1] of the variance of log(L(X_x)) + w(x) * X_x,…
- plackett_luce_surprisal_variance_profile_const_oneFor the constant weight profile w(x) = 1 on [0,1], the asymptotic surprisal variance plackett_luce_surprisal_variance_profile equals 0.
Read back from the Lean
The Gaussian measure on ℝ with mean 0 and variance parameter given by the nonnegative part (Real.toNNReal) of the Plackett-Luce surprisal variance profile of the constant function w(x) = 1 is equal to the Dirac measure at 0 (MeasureTheory.Measure.dirac 0).
Written by a model that saw only the Lean, never the English above. If the two disagree, that is worth a challenge.
The proof, explained
The limiting Gaussian distribution has mean 0 and variance given by the surprisal variance profile of the constant weight profile w(x) = 1. Since this variance profile is zero, the distribution reduces to a real Gaussian with zero variance. A Gaussian distribution with zero variance degenerates to the Dirac delta measure concentrated at its mean, which is 0.
The variance parameter evaluates to 0, reducing the goal to showing that the Gaussian distribution with mean 0 and variance 0 equals the Dirac delta measure at 0.
For the constant weight profile w(x) = 1, the surprisal variance profile vanishes by plackett_luce_surprisal_variance_profile_const_one. Converting the resulting real number 0 to the non-negative reals via Real.toNNReal_zero simplifies the variance argument of the Gaussian distribution to 0.
uses
plackett_luce_surprisal_variance_profile_const_one,Real.toNNReal_zerorw [plackett_luce_surprisal_variance_profile_const_one, Real.toNNReal_zero]The Gaussian distribution with mean 0 and variance 0 is the Dirac delta measure at 0.
By ProbabilityTheory.gaussianReal_zero_var, any real Gaussian measure with variance 0 degenerates to the Dirac delta measure concentrated at its mean, which in this case is 0.
uses
ProbabilityTheory.gaussianReal_zero_varexact ProbabilityTheory.gaussianReal_zero_var 0
plackett_luce_surprisal_variance_profile_const_one- The surprisal variance profile evaluated at the constant function 1 equals 0.
Real.toNNReal_zero- The projection of the real number 0 to the non-negative real numbers is 0.
ProbabilityTheory.gaussianReal_zero_var- A real Gaussian distribution with mean μ and variance 0 is the Dirac delta measure at μ.
Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.
Lean source view module on GitHub
Choose a tactic — a marked line — to see the goals Lean had before and after it, as in Lean’s infoview. Step with ← → or j k.
/-- The limiting Gaussian distribution for the constant weight profile w = 1 is Dirac delta at 0. -/theorem plackett_luce_surprisal_gaussian_limit_const_one : ProbabilityTheory.gaussianReal 0 (Real.toNNReal (plackett_luce_surprisal_variance_profile (fun _ => 1))) = MeasureTheory.Measure.dirac 0 := by rw [plackett_luce_surprisal_variance_profile_const_one, Real.toNNReal_zero] exact ProbabilityTheory.gaussianReal_zero_var 0