Empirical risk computation: nominal risk under the empirical distribution is the sample average
ProvedWassersteinDRO.Regularization.empirical_risk_computationboreldecompositiondroleafleaf-1wasserstein
The nominal risk of a measurable loss under the empirical distribution of N samples equals the sample average of the loss.
Preamble
import Mathlib import Definitions.Def_WassersteinDRO_Regularization_empiricalDistribution import Definitions.Def_WassersteinDRO_Regularization_nominalRisk open MeasureTheory
Formal statement
namespace WassersteinDRO.Regularization
theorem empirical_risk_computation {E : Type*} [MeasurableSpace E] [MeasurableSingletonClass E]
{N : ℕ} (hN : 0 < N) (ξhat : Fin N → E) (ℓ : E → ℝ) (hm : Measurable ℓ) :
nominalRisk (empiricalDistribution ξhat) ℓ =
(1 / (N : ℝ)) * Finset.sum Finset.univ (fun i => ℓ (ξhat i)) := by sorry
end WassersteinDRO.Regularization
Source
Kuhn, Mohajerin Esfahani, Nguyen & Shafieezadeh-Abadeh, INFORMS TutORials 2019 (Theorem 10, Lipschitz regularization of Wasserstein DRO); decomposition leaves of WassersteinDRO.Regularization.convex_loss_p1_borel (d58d7d19-a202-486d-8c9a-5855000c0837).