OAI/Analysis/NuclearUltrapower/TensorGap.lean

GitHub source file

38 declarations, 0 contain sorry, 0 declared as axioms

  1. def OAI.NuclearUltrapower.operatorMatrixAction line 26

    Source documentation

    The actual left action of an operator-valued matrix on a Hilbert direct sum.

    def operatorMatrixAction (M : CStarMatrix n n (H →L[ℂ] H)) :
        PiLp 2 (fun _ : n => H) →L[ℂ] PiLp 2 (fun _ : n => H) where
    …
  2. theorem OAI.NuclearUltrapower.operatorMatrixAction_apply line 38
  3. def OAI.NuclearUltrapower.operatorMatrixActionHom line 42

    Source documentation

    Matrix amplification is a genuine unital star homomorphism.

    def operatorMatrixActionHom : CStarMatrix n n (H →L[ℂ] H) →⋆ₐ[ℂ]
        (PiLp 2 (fun _ : n => H) →L[ℂ] PiLp 2 (fun _ : n => H)) where
    …
  4. theorem OAI.NuclearUltrapower.operatorMatrixActionHom_apply line 65
  5. theorem OAI.NuclearUltrapower.operatorMatrixActionHom_injective line 69

    Source documentation

    One-column vectors recover every coefficient; amplification is faithful.

    theorem operatorMatrixActionHom_injective : Function.Injective (operatorMatrixActionHom (n := n) (H := H)) := by
    …

    Mentions: OAI.NuclearUltrapower.operatorMatrixActionHom

  6. theorem OAI.NuclearUltrapower.operatorMatrixActionHom_isometry line 78
  7. def OAI.NuclearUltrapower.tensorPiLinearEquiv line 90

    Source documentation

    The fixed-factor tensor/direct-sum identification, with its actual Hilbert norm.

    def tensorPiLinearEquiv (b : OrthonormalBasis n ℂ F) :
        E ⊗[ℂ] F ≃ₗ[ℂ] PiLp 2 (fun _ : n => E) :=
    …
  8. theorem OAI.NuclearUltrapower.tensorPiLinearEquiv_tmul line 96
  9. theorem OAI.NuclearUltrapower.tensorPiLinearEquiv_inner line 100
  10. def OAI.NuclearUltrapower.tensorPiIsometry line 121
  11. theorem OAI.NuclearUltrapower.tensorPiIsometry_tmul line 125
  12. def OAI.NuclearUltrapower.tensorOperatorMatrix line 131

    Source documentation

    This matrix has scalar coefficients in the fixed factor; the first factor may be infinite-dimensional.

    def tensorOperatorMatrix (b : OrthonormalBasis n ℂ F)
        (U : E →L[ℂ] E) (V : F →L[ℂ] F) : CStarMatrix n n (E →L[ℂ] E) :=
    …
  13. theorem OAI.NuclearUltrapower.tensorOperatorMatrix_apply line 136
  14. theorem OAI.NuclearUltrapower.tensorOperatorMatrix_action line 141

    Source documentation

    Entry-by-entry compatibility of the operator tensor with matrix amplification.

    theorem tensorOperatorMatrix_action [CompleteSpace E] (b : OrthonormalBasis n ℂ F)
        (U : E →L[ℂ] E) (V : F →L[ℂ] F) (z : E ⊗[ℂ] F) :
        operatorMatrixActionHom (n := n) (H := E) (tensorOperatorMatrix b U V) (tensorPiIsometry b z) =
    …

    Mentions: OAI.NuclearUltrapower.operatorMatrixActionHom, OAI.NuclearUltrapower.tensorOperatorMatrix, OAI.NuclearUltrapower.tensorPiIsometry

  15. def OAI.NuclearUltrapower.tensorPiUnitary line 167
  16. theorem OAI.NuclearUltrapower.tensorPiUnitary_apply_tensorPi line 172
  17. theorem OAI.NuclearUltrapower.tensorOperatorMatrix_conjugacy line 178
  18. theorem OAI.NuclearUltrapower.tensorMatrix_complement_bound line 193

    Source documentation

    Fixed-factor matrix norms see precisely the tensor Kazhdan estimate.

    theorem tensorMatrix_complement_bound {Γ : Type*} (S : Finset Γ)
        (b : OrthonormalBasis n ℂ F) (U : Γ → (E ≃ₗᵢ[ℂ] E)) (V : Γ → (F ≃ₗᵢ[ℂ] F))
        {κ : ℝ} (hκ : κ ≤ 4 * (S.card : ℝ))
        (hgap : ∀ z : E ⊗[ℂ] F, κ * ‖z‖ ^ 2 ≤
          ∑ s ∈ S, ‖TensorProduct.congrIsometry (U s) (V s) z - z‖ ^ 2) :
        ‖(4 * (S.card : ℝ)) • (1 : CStarMatrix n n (E →L[ℂ] E)) -
          laplacian S (fun s => tensorOperatorMatrix b (U s) (V s))‖ ≤
            4 * (S.card : ℝ) - κ := by
    …

    Mentions: OAI.NuclearUltrapower.laplacian, OAI.NuclearUltrapower.tensorOperatorMatrix

  19. def OAI.NuclearUltrapower.coefficientMatrix line 242

    Source documentation

    An elementary tensor in coordinates of the fixed second factor.

    def coefficientMatrix (c : n → n → ℂ) (a : A) : CStarMatrix n n A :=
    …
  20. theorem OAI.NuclearUltrapower.coefficientMatrix_apply line 246
  21. theorem OAI.NuclearUltrapower.matrixMap_coefficientMatrix line 249
  22. theorem OAI.NuclearUltrapower.map_laplacian line 254
  23. theorem OAI.NuclearUltrapower.map_laplacian_complement line 258
  24. def OAI.NuclearUltrapower.VariableUltra.eval line 271

    Source documentation

    Evaluation is an actual unital star homomorphism.

    def eval (i : ι) : Seq A →⋆ₐ[ℂ] A i where
    …

    Mentions: OAI.NuclearUltrapower.VariableUltra.Seq

  25. theorem OAI.NuclearUltrapower.VariableUltra.eval_apply line 280
  26. def OAI.NuclearUltrapower.VariableUltra.unitarySequence line 283

    Source documentation

    Arbitrary coordinate unitaries are bounded, without a dimension bound.

    def unitarySequence (u : ∀ i, unitary (A i)) : unitary (Seq A) :=
    …

    Mentions: OAI.NuclearUltrapower.VariableUltra.Seq

  27. theorem OAI.NuclearUltrapower.VariableUltra.unitarySequence_apply line 291
  28. theorem OAI.NuclearUltrapower.VariableUltra.matrixFamily_eval line 303
  29. theorem OAI.NuclearUltrapower.VariableUltra.coefficient_complement_quotient_bound line 307

    Source documentation

    Fixed-size polynomial transfer, with only an EVENTUAL coordinate bound.

    theorem coefficient_complement_quotient_bound {Γ : Type*} (S : Finset Γ)
        (u : Γ → Seq A) (c : Γ → n → n → ℂ) (R C : ℝ)
        (h : ∀ᶠ i in (ω : Filter ι),
          ‖R • (1 : CStarMatrix n n (A i)) -
            laplacian S (fun s => coefficientMatrix (c s) (u s i))‖ ≤ C) :
        ‖R • (1 : CStarMatrix n n (Algebra A ω)) -
          laplacian S (fun s => coefficientMatrix (c s) (mk A ω (u s)))‖ ≤ C := by
    …

    Mentions: OAI.NuclearUltrapower.VariableUltra.Seq, OAI.NuclearUltrapower.laplacian, OAI.NuclearUltrapower.coefficientMatrix, OAI.NuclearUltrapower.VariableUltra.Algebra, OAI.NuclearUltrapower.VariableUltra.mk

  30. theorem OAI.NuclearUltrapower.no_invariant_of_complement_norm line 349

    Source documentation

    A strict complemented-norm gap excludes every common fixed vector.

    theorem no_invariant_of_complement_norm (S : Finset Γ) (a : Γ → (H →L[ℂ] H))
        {κ : ℝ} (hκ : 0 < κ)
        (h : ‖(4 * (S.card : ℝ)) • (1 : H →L[ℂ] H) - laplacian S a‖ ≤
          4 * (S.card : ℝ) - κ) (x : H) (hx : ∀ s ∈ S, a s x = x) : x = 0 := by
    …

    Mentions: OAI.NuclearUltrapower.laplacian

  31. abbrev OAI.NuclearUltrapower.VariableUltra.operatorSeq line 371
  32. abbrev OAI.NuclearUltrapower.VariableUltra.operatorQuotient line 372
  33. def OAI.NuclearUltrapower.VariableUltra.isoSequence line 374
  34. def OAI.NuclearUltrapower.VariableUltra.starAlgToStarMonoid line 377
  35. def OAI.NuclearUltrapower.VariableUltra.quotientIsoUnitary line 384
  36. def OAI.NuclearUltrapower.VariableUltra.representedIso line 391
  37. theorem OAI.NuclearUltrapower.VariableUltra.quotient_tensor_complement_bound line 402

    Source documentation

    The strict coordinate tensor gap survives in EVERY unital representation of the varying matrix quotient. No faithfulness is needed here.

    theorem quotient_tensor_complement_bound {Γ : Type*} (S : Finset Γ)
        (b : OrthonormalBasis n ℂ F) (U : ∀ i, Γ → (E i ≃ₗᵢ[ℂ] E i))
        (V : Γ → (F ≃ₗᵢ[ℂ] F)) {κ : ℝ} (hκ : κ ≤ 4 * (S.card : ℝ))
        (hgap : ∀ᶠ i in (ω : Filter ι), ∀ z : E i ⊗[ℂ] F,
          κ * ‖z‖ ^ 2 ≤ ∑ s ∈ S, ‖TensorProduct.congrIsometry (U i s) (V s) z - z‖ ^ 2)
        (φ : operatorQuotient E ω →⋆ₐ[ℂ] (K →L[ℂ] K)) :
        ‖(4 * (S.card : ℝ)) • (1 : CStarMatrix n n (K →L[ℂ] K)) -
          laplacian S (fun s => tensorOperatorMatrix b
            (representedIso E ω φ (fun i => U i s)) (V s))‖ ≤ 4 * (S.card : ℝ) - κ := by
    …

    Mentions: OAI.NuclearUltrapower.VariableUltra.operatorQuotient, OAI.NuclearUltrapower.laplacian, OAI.NuclearUltrapower.tensorOperatorMatrix, OAI.NuclearUltrapower.VariableUltra.representedIso

  38. theorem OAI.NuclearUltrapower.VariableUltra.quotient_tensor_no_invariants line 438

    Source documentation

    In particular a tensor intertwiner in a quotient representation must vanish.

    theorem quotient_tensor_no_invariants {Γ : Type*} (S : Finset Γ)
        (b : OrthonormalBasis n ℂ F) (U : ∀ i, Γ → (E i ≃ₗᵢ[ℂ] E i))
        (V : Γ → (F ≃ₗᵢ[ℂ] F)) {κ : ℝ} (hκ0 : 0 < κ) (hκ : κ ≤ 4 * (S.card : ℝ))
        (hgap : ∀ᶠ i in (ω : Filter ι), ∀ z : E i ⊗[ℂ] F,
          κ * ‖z‖ ^ 2 ≤ ∑ s ∈ S, ‖TensorProduct.congrIsometry (U i s) (V s) z - z‖ ^ 2)
        (φ : operatorQuotient E ω →⋆ₐ[ℂ] (K →L[ℂ] K))
        (z : K ⊗[ℂ] F)
        (hz : ∀ s ∈ S, TensorProduct.congrIsometry
          (representedIso E ω φ (fun i => U i s)) (V s) z = z) : z = 0 := by
    …

    Mentions: OAI.NuclearUltrapower.VariableUltra.operatorQuotient, OAI.NuclearUltrapower.VariableUltra.representedIso