Documentation

MIPRE.Background.Repetition.CommutingRepetition.VN.TensorStep

The product trace #

The product trace τ(a ⊗ b) = τ₁(a) τ₂(b) on the algebraic tensor product.

Equations
Instances For
    @[simp]
    theorem CommutingRepetition.StdTracialAlgebra.stepτ_tmul (M₁ M₂ : StdTracialAlgebra) (a : M₁.A) (b : M₂.A) :
    (M₁.stepτ M₂) (a ⊗ₜ[ℂ] b) = M₁.τ a * M₂.τ b
    theorem CommutingRepetition.StdTracialAlgebra.stepτ_mul_comm (M₁ M₂ : StdTracialAlgebra) (x y : TensorProduct ℂ M₁.A M₂.A) :
    (M₁.stepτ M₂) (x * y) = (M₁.stepτ M₂) (y * x)
    theorem CommutingRepetition.StdTracialAlgebra.stepτ_star (M₁ M₂ : StdTracialAlgebra) (x : TensorProduct ℂ M₁.A M₂.A) :
    (M₁.stepτ M₂) (star x) = star ((M₁.stepτ M₂) x)

    The GNS space #

    @[reducible, inline]

    The GNS space of the product trace: the completion of the inner-product tensor product of the two GNS spaces.

    Equations
    Instances For

      The GNS embedding: ι₁ ⊗ ι₂ into the completion.

      Equations
      Instances For
        @[simp]
        theorem CommutingRepetition.StdTracialAlgebra.stepι_tmul (M₁ M₂ : StdTracialAlgebra) (a : M₁.A) (b : M₂.A) :
        (M₁.stepι M₂) (a ⊗ₜ[ℂ] b) = ↑(M₁.ι a ⊗ₜ[ℂ] M₂.ι b)
        theorem CommutingRepetition.StdTracialAlgebra.stepι_inner (M₁ M₂ : StdTracialAlgebra) (x y : TensorProduct ℂ M₁.A M₂.A) :
        inner ℂ ((M₁.stepι M₂) x) ((M₁.stepι M₂) y) = (M₁.stepτ M₂) (star x * y)

        Density of the algebraic image inside the tensor product (before completion): the range of ι₁ ⊗ ι₂ is dense in H₁ ⊗ H₂. Two-slot closed-preimage argument: for fixed a, the set of y with ι₁ a ⊗ y in the closure is closed and contains the dense range of ι₂; then the set of x with x ⊗ y in the closure is closed and contains the dense range of ι₁; conclude on pure tensors and extend by the submodule structure of the closure.

        The GNS embedding has dense range (density in the algebraic tensor product composed with density of the completion coercion).

        Extending operators to the completion #

        The completion coercion as a continuous linear map.

        Equations
        Instances For
          @[simp]
          theorem CommutingRepetition.StdTracialAlgebra.toStepH_apply (M₁ M₂ : StdTracialAlgebra) (w : TensorProduct ℂ M₁.H M₂.H) :
          (M₁.toStepH M₂) w = ↑w
          noncomputable def CommutingRepetition.StdTracialAlgebra.extendStep (M₁ M₂ : StdTracialAlgebra) (T : TensorProduct ℂ M₁.H M₂.H →L[ℂ] TensorProduct ℂ M₁.H M₂.H) :
          M₁.StepH M₂ →L[ℂ] M₁.StepH M₂

          The unique bounded extension of an operator on the algebraic tensor product to the completed GNS space.

          Equations
          Instances For
            theorem CommutingRepetition.StdTracialAlgebra.extendStep_coe (M₁ M₂ : StdTracialAlgebra) (T : TensorProduct ℂ M₁.H M₂.H →L[ℂ] TensorProduct ℂ M₁.H M₂.H) (w : TensorProduct ℂ M₁.H M₂.H) :
            (M₁.extendStep M₂ T) ↑w = ↑(T w)
            theorem CommutingRepetition.StdTracialAlgebra.ext_of_coe (M₁ M₂ : StdTracialAlgebra) {P Q : M₁.StepH M₂ →L[ℂ] M₁.StepH M₂} (h : ∀ (w : TensorProduct ℂ M₁.H M₂.H), P ↑w = Q ↑w) :
            P = Q

            Two operators on the completion agreeing on coercions are equal.

            theorem CommutingRepetition.StdTracialAlgebra.extendStep_add (M₁ M₂ : StdTracialAlgebra) (S T : TensorProduct ℂ M₁.H M₂.H →L[ℂ] TensorProduct ℂ M₁.H M₂.H) :
            M₁.extendStep M₂ (S + T) = M₁.extendStep M₂ S + M₁.extendStep M₂ T
            theorem CommutingRepetition.StdTracialAlgebra.extendStep_smul (M₁ M₂ : StdTracialAlgebra) (c : ℂ) (T : TensorProduct ℂ M₁.H M₂.H →L[ℂ] TensorProduct ℂ M₁.H M₂.H) :
            M₁.extendStep M₂ (c • T) = c • M₁.extendStep M₂ T
            theorem CommutingRepetition.StdTracialAlgebra.inner_mapL_star (M₁ M₂ : StdTracialAlgebra) (S : M₁.H →L[ℂ] M₁.H) (T : M₂.H →L[ℂ] M₂.H) (w w' : TensorProduct ℂ M₁.H M₂.H) :

            Adjoint duality for factorwise operators on the algebraic tensor product: ⟪(S* ⊗ T*) w, w'⟫ = ⟪w, (S ⊗ T) w'⟫.

            theorem CommutingRepetition.StdTracialAlgebra.extendStep_star_of_inner (M₁ M₂ : StdTracialAlgebra) {P Q : TensorProduct ℂ M₁.H M₂.H →L[ℂ] TensorProduct ℂ M₁.H M₂.H} (hPQ : ∀ (w w' : TensorProduct ℂ M₁.H M₂.H), inner ℂ (P w) w' = inner ℂ w (Q w')) :
            M₁.extendStep M₂ P = star (M₁.extendStep M₂ Q)

            Adjoint compatibility of the extension, from inner-product duality of the unextended operators: if ⟪P w, w'⟫ = ⟪w, Q w'⟫ on the algebraic tensor product then the extension of P is the adjoint of the extension of Q.

            The factorwise representation #

            noncomputable def CommutingRepetition.StdTracialAlgebra.tensorRep0 (M₁ M₂ : StdTracialAlgebra) {A₁ : Type u_1} [Ring A₁] [StarRing A₁] [Algebra ℂ A₁] {A₂ : Type u_2} [Ring A₂] [StarRing A₂] [Algebra ℂ A₂] (φ₁ : A₁ →⋆ₐ[ℂ] M₁.H →L[ℂ] M₁.H) (φ₂ : A₂ →⋆ₐ[ℂ] M₂.H →L[ℂ] M₂.H) :

            The factorwise action on the algebraic tensor product: a ⊗ b ↦ φ₁(a) ⊗ φ₂(b) as a bounded operator.

            Equations
            Instances For
              @[simp]
              theorem CommutingRepetition.StdTracialAlgebra.tensorRep0_tmul (M₁ M₂ : StdTracialAlgebra) {A₁ : Type u_1} [Ring A₁] [StarRing A₁] [Algebra ℂ A₁] [StarModule ℂ A₁] {A₂ : Type u_2} [Ring A₂] [StarRing A₂] [Algebra ℂ A₂] [StarModule ℂ A₂] (φ₁ : A₁ →⋆ₐ[ℂ] M₁.H →L[ℂ] M₁.H) (φ₂ : A₂ →⋆ₐ[ℂ] M₂.H →L[ℂ] M₂.H) (a : A₁) (b : A₂) :
              (M₁.tensorRep0 M₂ φ₁ φ₂) (a ⊗ₜ[ℂ] b) = TensorProduct.mapL (φ₁ a) (φ₂ b)
              theorem CommutingRepetition.StdTracialAlgebra.tensorRep0_one (M₁ M₂ : StdTracialAlgebra) {A₁ : Type u_1} [Ring A₁] [StarRing A₁] [Algebra ℂ A₁] [StarModule ℂ A₁] {A₂ : Type u_2} [Ring A₂] [StarRing A₂] [Algebra ℂ A₂] [StarModule ℂ A₂] (φ₁ : A₁ →⋆ₐ[ℂ] M₁.H →L[ℂ] M₁.H) (φ₂ : A₂ →⋆ₐ[ℂ] M₂.H →L[ℂ] M₂.H) :
              (M₁.tensorRep0 M₂ φ₁ φ₂) 1 = 1
              theorem CommutingRepetition.StdTracialAlgebra.tensorRep0_mul (M₁ M₂ : StdTracialAlgebra) {A₁ : Type u_1} [Ring A₁] [StarRing A₁] [Algebra ℂ A₁] [StarModule ℂ A₁] {A₂ : Type u_2} [Ring A₂] [StarRing A₂] [Algebra ℂ A₂] [StarModule ℂ A₂] (φ₁ : A₁ →⋆ₐ[ℂ] M₁.H →L[ℂ] M₁.H) (φ₂ : A₂ →⋆ₐ[ℂ] M₂.H →L[ℂ] M₂.H) (x y : TensorProduct ℂ A₁ A₂) :
              (M₁.tensorRep0 M₂ φ₁ φ₂) (x * y) = (M₁.tensorRep0 M₂ φ₁ φ₂) x * (M₁.tensorRep0 M₂ φ₁ φ₂) y
              theorem CommutingRepetition.StdTracialAlgebra.tensorRep0_inner_star (M₁ M₂ : StdTracialAlgebra) {A₁ : Type u_1} [Ring A₁] [StarRing A₁] [Algebra ℂ A₁] [StarModule ℂ A₁] {A₂ : Type u_2} [Ring A₂] [StarRing A₂] [Algebra ℂ A₂] [StarModule ℂ A₂] (φ₁ : A₁ →⋆ₐ[ℂ] M₁.H →L[ℂ] M₁.H) (φ₂ : A₂ →⋆ₐ[ℂ] M₂.H →L[ℂ] M₂.H) (x : TensorProduct ℂ A₁ A₂) (w w' : TensorProduct ℂ M₁.H M₂.H) :
              inner ℂ (((M₁.tensorRep0 M₂ φ₁ φ₂) (star x)) w) w' = inner ℂ w (((M₁.tensorRep0 M₂ φ₁ φ₂) x) w')
              noncomputable def CommutingRepetition.StdTracialAlgebra.tensorRep (M₁ M₂ : StdTracialAlgebra) {A₁ : Type u_1} [Ring A₁] [StarRing A₁] [Algebra ℂ A₁] [StarModule ℂ A₁] {A₂ : Type u_2} [Ring A₂] [StarRing A₂] [Algebra ℂ A₂] [StarModule ℂ A₂] (φ₁ : A₁ →⋆ₐ[ℂ] M₁.H →L[ℂ] M₁.H) (φ₂ : A₂ →⋆ₐ[ℂ] M₂.H →L[ℂ] M₂.H) :
              TensorProduct ℂ A₁ A₂ →⋆ₐ[ℂ] M₁.StepH M₂ →L[ℂ] M₁.StepH M₂

              The factorwise representation on the completed GNS space, as a ∗-algebra homomorphism: a ⊗ b acts by the extension of φ₁(a) ⊗ φ₂(b).

              Equations
              • One or more equations did not get rendered due to their size.
              Instances For
                theorem CommutingRepetition.StdTracialAlgebra.tensorRep_apply (M₁ M₂ : StdTracialAlgebra) {A₁ : Type u_1} [Ring A₁] [StarRing A₁] [Algebra ℂ A₁] [StarModule ℂ A₁] {A₂ : Type u_2} [Ring A₂] [StarRing A₂] [Algebra ℂ A₂] [StarModule ℂ A₂] (φ₁ : A₁ →⋆ₐ[ℂ] M₁.H →L[ℂ] M₁.H) (φ₂ : A₂ →⋆ₐ[ℂ] M₂.H →L[ℂ] M₂.H) (x : TensorProduct ℂ A₁ A₂) :
                (M₁.tensorRep M₂ φ₁ φ₂) x = M₁.extendStep M₂ ((M₁.tensorRep0 M₂ φ₁ φ₂) x)
                theorem CommutingRepetition.StdTracialAlgebra.tensorRep_coe (M₁ M₂ : StdTracialAlgebra) {A₁ : Type u_1} [Ring A₁] [StarRing A₁] [Algebra ℂ A₁] [StarModule ℂ A₁] {A₂ : Type u_2} [Ring A₂] [StarRing A₂] [Algebra ℂ A₂] [StarModule ℂ A₂] (φ₁ : A₁ →⋆ₐ[ℂ] M₁.H →L[ℂ] M₁.H) (φ₂ : A₂ →⋆ₐ[ℂ] M₂.H →L[ℂ] M₂.H) (x : TensorProduct ℂ A₁ A₂) (w : TensorProduct ℂ M₁.H M₂.H) :
                ((M₁.tensorRep M₂ φ₁ φ₂) x) ↑w = ↑(((M₁.tensorRep0 M₂ φ₁ φ₂) x) w)

                The left and right representations and the assembly #

                The opposite-algebra distribution (A₁ ⊗ A₂)ᵐᵒᵖ → A₁ᵐᵒᵖ ⊗ A₂ᵐᵒᵖ as a ∗-algebra homomorphism.

                Equations
                Instances For

                  The left representation of the tensor step.

                  Equations
                  Instances For

                    The right representation of the tensor step.

                    Equations
                    Instances For
                      theorem CommutingRepetition.StdTracialAlgebra.tensorRep0_L_map_ι (M₁ M₂ : StdTracialAlgebra) (x y : TensorProduct ℂ M₁.A M₂.A) :
                      ((M₁.tensorRep0 M₂ M₁.L M₂.L) x) ((TensorProduct.map M₁.ι M₂.ι) y) = (TensorProduct.map M₁.ι M₂.ι) (x * y)

                      The left evaluation identity on the algebraic image.

                      theorem CommutingRepetition.StdTracialAlgebra.stepL_apply (M₁ M₂ : StdTracialAlgebra) (x y : TensorProduct ℂ M₁.A M₂.A) :
                      ((M₁.stepL M₂) x) ((M₁.stepι M₂) y) = (M₁.stepι M₂) (x * y)
                      theorem CommutingRepetition.StdTracialAlgebra.tensorRep0_R_map_ι (M₁ M₂ : StdTracialAlgebra) (x y : TensorProduct ℂ M₁.A M₂.A) :
                      ((M₁.tensorRep0 M₂ M₁.R M₂.R) ((M₁.opDistrib M₂) (MulOpposite.op x))) ((TensorProduct.map M₁.ι M₂.ι) y) = (TensorProduct.map M₁.ι M₂.ι) (y * x)

                      The right evaluation identity on the algebraic image.

                      theorem CommutingRepetition.StdTracialAlgebra.stepR_apply (M₁ M₂ : StdTracialAlgebra) (x y : TensorProduct ℂ M₁.A M₂.A) :
                      ((M₁.stepR M₂) (MulOpposite.op x)) ((M₁.stepι M₂) y) = (M₁.stepι M₂) (y * x)
                      theorem CommutingRepetition.StdTracialAlgebra.tensorRep0_commute (M₁ M₂ : StdTracialAlgebra) {A₁ : Type u_1} [Ring A₁] [StarRing A₁] [Algebra ℂ A₁] [StarModule ℂ A₁] {A₂ : Type u_2} [Ring A₂] [StarRing A₂] [Algebra ℂ A₂] [StarModule ℂ A₂] {B₁ : Type u_3} [Ring B₁] [StarRing B₁] [Algebra ℂ B₁] [StarModule ℂ B₁] {B₂ : Type u_4} [Ring B₂] [StarRing B₂] [Algebra ℂ B₂] [StarModule ℂ B₂] (φ₁ : A₁ →⋆ₐ[ℂ] M₁.H →L[ℂ] M₁.H) (φ₂ : A₂ →⋆ₐ[ℂ] M₂.H →L[ℂ] M₂.H) (ψ₁ : B₁ →⋆ₐ[ℂ] M₁.H →L[ℂ] M₁.H) (ψ₂ : B₂ →⋆ₐ[ℂ] M₂.H →L[ℂ] M₂.H) (h₁ : ∀ (a : A₁) (c : B₁), Commute (φ₁ a) (ψ₁ c)) (h₂ : ∀ (b : A₂) (d : B₂), Commute (φ₂ b) (ψ₂ d)) (x : TensorProduct ℂ A₁ A₂) (z : TensorProduct ℂ B₁ B₂) :
                      Commute ((M₁.tensorRep0 M₂ φ₁ φ₂) x) ((M₁.tensorRep0 M₂ ψ₁ ψ₂) z)

                      Factorwise commutation on the algebraic tensor product, from pointwise commutation of the factor actions.

                      theorem CommutingRepetition.StdTracialAlgebra.stepLR_commute (M₁ M₂ : StdTracialAlgebra) (x y : TensorProduct ℂ M₁.A M₂.A) :
                      Commute ((M₁.stepL M₂) x) ((M₁.stepR M₂) (MulOpposite.op y))

                      The binary tensor step: the algebraic tensor product with the product trace, in standard form on the completed inner-product tensor of the GNS spaces (Stage B, WP-B3b).

                      Equations
                      • One or more equations did not get rendered due to their size.
                      Instances For
                        @[simp]
                        theorem CommutingRepetition.StdTracialAlgebra.tensorStep_τ (M₁ M₂ : StdTracialAlgebra) (x : TensorProduct ℂ M₁.A M₂.A) :
                        (M₁.tensorStep M₂).τ x = (M₁.stepτ M₂) x