bernoulli_expectation_monotone
ProvedMonotonicity of the Bernoulli expectation in the integrand.
Fix dimensions and an inclusion probability . Let with weights , . If for every observation set , then
The hypothesis is exactly what makes every weight , so the inequality survives summation. This is the Bernoulli-model analogue of rademacher_expectation_monotone, used to push per- conditional bounds through the outer expectation over the sampling set.
Preamble
import Definitions.Def_matrix_completion_bernoulli open MatrixCompletion
Formal statement
theorem bernoulli_expectation_monotone
{n1 n2 : ℕ} (p : ℝ)
(F G : Finset (Fin n1 × Fin n2) → ℝ) :
0 ≤ p → p ≤ 1 → (∀ Ω, F Ω ≤ G Ω) →
bernoulliExpectation p F ≤ bernoulliExpectation p G := by sorrySource
Candès–Recht, Exact Matrix Completion via Convex Optimization (Bernoulli model)