OAI/Analysis/KadisonSimilarity/Representation/RankOneFactorCriterion.lean

GitHub source file

4 declarations, 0 contain sorry, 0 declared as axioms

Rank-one tests for scalarity. A countable dense family suffices, including on a variable projection range. These are the tests used to prove that the central-decomposition fibers are factors.

  1. theorem OAI.KadisonSimilarity.scalar_of_commutes_rankOne line 16

    Source documentation

    Commuting with all rank-one operators forces an operator to be scalar.

    theorem scalar_of_commutes_rankOne (T : H →L[ℂ] H)
        (hT : ∀ v w : H, T * InnerProductSpace.rankOne ℂ v w =
          InnerProductSpace.rankOne ℂ v w * T) :
        ∃ c : ℂ, T = c • (1 : H →L[ℂ] H) := by
    …
  2. theorem OAI.KadisonSimilarity.scalar_of_commutes_dense_rankOne line 40

    Source documentation

    It suffices to test rank-one operators on a dense family of vectors.

    theorem scalar_of_commutes_dense_rankOne {ι : Type*} (d : ι → H)
        (hd : DenseRange d) (T : H →L[ℂ] H)
        (hT : ∀ i j, T * InnerProductSpace.rankOne ℂ (d i) (d j) =
          InnerProductSpace.rankOne ℂ (d i) (d j) * T) :
        ∃ c : ℂ, T = c • (1 : H →L[ℂ] H) := by
    …
  3. theorem OAI.KadisonSimilarity.projectionFiber_denseRange line 58

    Source documentation

    Projecting a dense family gives a dense family in the actual range.

    theorem projectionFiber_denseRange {ι : Type*} (d : ι → H) (hd : DenseRange d)
        (p : H →L[ℂ] H) (hp : IsStarProjection p) :
        DenseRange (fun i => projectionFiberRetraction p hp (d i)) := by
    …

    Mentions: OAI.KadisonSimilarity.projectionFiberRetraction

  4. theorem OAI.KadisonSimilarity.scalar_of_commutes_projection_rankOne line 70

    Source documentation

    Countably many compressed rank-one commutation identities suffice for scalarity on the projection fiber.

    theorem scalar_of_commutes_projection_rankOne {ι : Type*} (d : ι → H)
        (hd : DenseRange d) (p : H →L[ℂ] H) (hp : IsStarProjection p)
        (T : projectionFiber p hp →L[ℂ] projectionFiber p hp)
        (hT : ∀ i j,
          T * InnerProductSpace.rankOne ℂ
            (projectionFiberRetraction p hp (d i)) (projectionFiberRetraction p hp (d j)) =
          InnerProductSpace.rankOne ℂ
            (projectionFiberRetraction p hp (d i)) (projectionFiberRetraction p hp (d j)) * T) :
        ∃ c : ℂ, T = c • (1 : projectionFiber p hp →L[ℂ] projectionFiber p hp) :=
    …

    Mentions: OAI.KadisonSimilarity.projectionFiber, OAI.KadisonSimilarity.projectionFiberRetraction