OAI/NumberTheory/DukePrimeDegree/Entropy/EntropyInverse.lean

GitHub source file

5 declarations, 0 contain sorry, 0 declared as axioms

Time reversal of entropy

Reversing a finite orbit name and shifting its starting point does not change its Shannon entropy. This proves equality of Kolmogorov--Sinai entropy for mutually inverse measure-preserving maps, as used in EKL (2006), equation (3.2).

  1. def OAI.DukePrimeDegree.reverseWordEquiv line 22

    Source documentation

    Reversal is a permutation of the finite word alphabet.

    def reverseWordEquiv (α : Type*) (N : ℕ) : (Fin N → α) ≃ (Fin N → α) where
    …
  2. lemma OAI.DukePrimeDegree.inverse_orbitName line 28
  3. theorem OAI.DukePrimeDegree.orbitEntropy_inverse line 42

    Source documentation

    Finite block entropies agree exactly under time reversal.

    theorem orbitEntropy_inverse {Y α : Type*} [MeasurableSpace Y] [Fintype α]
        (μ : Measure Y) (T S : Y → Y)
        (hT : MeasurePreserving T μ μ) (hS : Measurable S)
        (hST : Function.LeftInverse S T)
        (P : Y → α) (hP : ∀ a, MeasurableSet {y | P y = a}) (N : ℕ) :
        observationEntropy μ (orbitName S P N) =
          observationEntropy μ (orbitName T P N) := by
    …
  4. theorem OAI.DukePrimeDegree.observationEntropyRate_inverse line 61

    Source documentation

    Equality holds for every finite observation, before taking the supremum.

    theorem observationEntropyRate_inverse {Y α : Type*}
        [MeasurableSpace Y] [Fintype α]
        (μ : Measure Y) (T S : Y → Y)
        (hT : MeasurePreserving T μ μ) (hS : Measurable S)
        (hST : Function.LeftInverse S T)
        (P : Y → α) (hP : ∀ a, MeasurableSet {y | P y = a}) :
        observationEntropyRate μ S P = observationEntropyRate μ T P := by
    …
  5. theorem OAI.DukePrimeDegree.measureEntropy_inverse line 74

    Source documentation

    Entropy is invariant under reversing an invertible measure-preserving map.

    theorem measureEntropy_inverse {Y : Type*} [MeasurableSpace Y]
        (μ : Measure Y) (T S : Y → Y)
        (hT : MeasurePreserving T μ μ) (hS : Measurable S)
        (hST : Function.LeftInverse S T) :
        measureEntropy μ S = measureEntropy μ T := by
    …