OAI/Analysis/KadisonKastler/ControlledPowerDrift.lean

GitHub source file

3 declarations, 0 contain sorry, 0 declared as axioms

Adding the controlled lift to the positive word error

  1. theorem OAI.KadisonKastler.controlled_word_error_identity line 19

    Source documentation

    Cancelling the test unitary separates the positive word error into its word commutator and the controlled lift commutator.

    theorem controlled_word_error_identity
        (T : K →L[ℂ] K) (Q s U : unitary (K →L[ℂ] K)) (ξ : K) :
        T (((Q : K →L[ℂ] K) * s * U) ξ) -
          ((Q : K →L[ℂ] K) * s * U) ((star (U : K →L[ℂ] K)) (T ((U : K →L[ℂ] K) ξ))) =
        (T * (Q : K →L[ℂ] K) - Q * T) ((s : K →L[ℂ] K) ((U : K →L[ℂ] K) ξ)) +
          (Q : K →L[ℂ] K) ((T * (s : K →L[ℂ] K) - s * T) ((U : K →L[ℂ] K) ξ)) := by
    …
  2. theorem OAI.KadisonKastler.controlled_word_error_rms_le line 36

    Source documentation

    The controlled lift adds at most 2δ to the word's root mean square.

    theorem controlled_word_error_rms_le
        (T : K →L[ℂ] K) (Q : Ω → unitary (K →L[ℂ] K))
        (s U : unitary (K →L[ℂ] K)) (ξ : K) {δ : ℝ} (hδ : 0 ≤ δ)
        (hs : ‖T * (s : K →L[ℂ] K) - s * T‖ ≤ 2 * δ) :
        finiteRMS (fun θ => T (((Q θ : K →L[ℂ] K) * s * U) ξ) -
          ((Q θ : K →L[ℂ] K) * s * U)
            ((star (U : K →L[ℂ] K)) (T ((U : K →L[ℂ] K) ξ)))) ≤
        finiteRMS (fun θ => (T * (Q θ : K →L[ℂ] K) - Q θ * T)
          ((s : K →L[ℂ] K) ((U : K →L[ℂ] K) ξ))) + 2 * δ * ‖ξ‖ := by
    …

    Mentions: OAI.KadisonKastler.finiteRMS

  3. theorem OAI.KadisonKastler.eventually_controlled_word_error line 59

    Source documentation

    The 68δ positive-word limit and the controlled lift give 70δ.

    theorem eventually_controlled_word_error
        (l : Filter J) (T : K →L[ℂ] K)
        (Q : J → Ω → unitary (K →L[ℂ] K))
        (s : J → unitary (K →L[ℂ] K)) (U : unitary (K →L[ℂ] K))
        (ξ : K) {δ : ℝ} (hδ : 0 ≤ δ)
        (hs : ∀ j, ‖T * (s j : K →L[ℂ] K) - s j * T‖ ≤ 2 * δ)
        (hword : ∀ ε > 0, ∀ᶠ j in l,
          finiteRMS (fun θ => (T * (Q j θ : K →L[ℂ] K) - Q j θ * T)
            ((s j : K →L[ℂ] K) ((U : K →L[ℂ] K) ξ))) ≤ 68 * δ * ‖ξ‖ + ε) :
        ∀ ε > 0, ∀ᶠ j in l,
          finiteRMS (fun θ => T (((Q j θ : K →L[ℂ] K) * s j * U) ξ) -
            ((Q j θ : K →L[ℂ] K) * s j * U)
    …

    Mentions: OAI.KadisonKastler.finiteRMS