Randomness and Solovay degrees #
Page 3 says:
We sometimes identify a real α with its binary expansion X ∈ 2^ω
and the corresponding set X = {n ∈ ℕ : X (n) = 1 }
Here we prove the well known fact that this representation is not unique.
We also show that consequently, if we use Mathlib's current definition of
the pullback measure as of June 23, 2024 as a basis for a measure on Cantor space,
the Cantor space gets measure 0.
However, if we use Etienne Marion's KolmogorovExtension4 library things work out well.
Instances For
Equations
- fairCoin'' = PMF.bernoulli (1 / 2) fairCoin''._proof_2
Instances For
Instances For
instance
instIsProbabilityMeasureBoolValMeasureβ
(n : ℕ)
:
MeasureTheory.IsProbabilityMeasure ((fun (x : ℕ) => ↑β) n)
noncomputable def
μ_bernoulli
(p : NNReal)
(hp : p ≤ 1)
:
MeasureTheory.Measure ((i : ℕ) → (fun (x : ℕ) => Bool) i)
The Bernoulli measure with parameter p.
Equations
- μ_bernoulli p hp = MeasureTheory.Measure.infinitePi fun (x : ℕ) => ↑(β'measure p hp)