OAI/Analysis/KadisonKastler/CanonicalAmplificationDescent.lean

GitHub source file

9 declarations, 0 contain sorry, 0 declared as axioms

The coordinate slice and descent from the canonical amplification

Compression along the zeroth coordinate is an actual unital completely positive retraction of the diagonal embedding. Matrix-entry membership shows that it maps the amplified target algebra into the original one.

  1. def OAI.KadisonKastler.operatorCompressionLinear line 33
  2. theorem OAI.KadisonKastler.operatorCompressionLinear_apply line 39
  3. def OAI.KadisonKastler.operatorCompressionCP line 44

    Source documentation

    Compression by any bounded operator is completely positive.

    def operatorCompressionCP (S : H →L[ℂ] K) : (K →L[ℂ] K) →CP (H →L[ℂ] H) where
    …
  4. def OAI.KadisonKastler.canonicalSlice line 79

    Source documentation

    The actual completely positive coordinate-zero slice.

    def canonicalSlice :
        (CountableHilbertSum H →L[ℂ] CountableHilbertSum H) →CP (H →L[ℂ] H) :=
    …

    Mentions: OAI.KadisonKastler.CountableHilbertSum

  5. theorem OAI.KadisonKastler.canonicalSlice_apply line 83
  6. theorem OAI.KadisonKastler.canonicalSlice_one line 90
  7. theorem OAI.KadisonKastler.canonicalSlice_l2Diagonal line 94
  8. theorem OAI.KadisonKastler.canonicalSlice_mem line 99
  9. theorem OAI.KadisonKastler.concrete_unitary_of_canonical_amplification line 117

    Source documentation

    A near-identity conjugacy of the actual canonical amplifications descends through the coordinate slice to a conjugacy on the original space.

    theorem concrete_unitary_of_canonical_amplification {M N : VonNeumannAlgebra H}
        (hCorrection : PublishedMultiplicationCorrection M)
        (hCommutant : PublishedBlockCommutantApproximation (Algebra M) H)
        (z : unitary (CountableHilbertSum H →L[ℂ] CountableHilbertSum H))
        (hz : Conjugates z (canonicalAmplifiedAlgebra H M) (canonicalAmplifiedAlgebra H N))
        {g : ℝ} (hgap : kkDistance M N < g)
        (hsmall : 2 * ‖(z : CountableHilbertSum H →L[ℂ] CountableHilbertSum H) - 1‖ < 1 / 1000)
        (hge : g + 2 * ‖(z : CountableHilbertSum H →L[ℂ] CountableHilbertSum H) - 1‖ < 1) :
        ∃ u : unitary (H →L[ℂ] H), Conjugates u M N ∧
          ‖(u : H →L[ℂ] H) - 1‖ ≤
            256 * ‖(z : CountableHilbertSum H →L[ℂ] CountableHilbertSum H) - 1‖ := by
    …

    Mentions: OAI.KadisonKastler.PublishedMultiplicationCorrection, OAI.KadisonKastler.PublishedBlockCommutantApproximation, OAI.KadisonKastler.Algebra, OAI.KadisonKastler.CountableHilbertSum, OAI.KadisonKastler.Conjugates, OAI.KadisonKastler.canonicalAmplifiedAlgebra, OAI.KadisonKastler.kkDistance