DPI (Data Processing Inequality) #
The Data Processing Inequality (DPI) for the sandwiched Rényi relative entropy, and as a consequence, the quantum relative entropy.
noncomputable def
weighted_norm
{d : Type u_1}
[Fintype d]
[DecidableEq d]
(p : ℝ)
(σ : MState d)
(X : Matrix d d ℂ)
:
The weighted norm |X|_{p, σ} defined in the paper.
Equations
Instances For
The spectral norm (operator norm) of a matrix.
Equations
- spectralNorm_mat A = if h : Finset.univ.Nonempty then let A_dag_A := ⟨A.conjTranspose * A, ⋯⟩; √((Finset.image ⋯.eigenvalues Finset.univ).max' ⋯) else 0
Instances For
The weighted norm for p = \infty.
Equations
- weighted_norm_infty x✝ X = spectralNorm_mat X
Instances For
noncomputable def
induced_norm
{d : Type u_1}
{d₂ : Type u_3}
[Fintype d]
[Fintype d₂]
[DecidableEq d]
[DecidableEq d₂]
(p : ℝ)
(σ : MState d)
(Φ : CPTPMap d d₂ ℂ)
(Ψ : Matrix d d ℂ → Matrix d₂ d₂ ℂ)
:
The induced norm of a map Ψ: M_d -> M_d2 with respect to weighted norms.
Equations
- induced_norm p σ Φ Ψ = sSup {x : ℝ | ∃ (X : Matrix d d ℂ) (_ : weighted_norm p σ X ≠ 0), weighted_norm p (Φ σ) (Ψ X) / weighted_norm p σ X = x}
Instances For
noncomputable def
induced_norm_infty_map
{d : Type u_1}
{d₂ : Type u_3}
[Fintype d]
[Fintype d₂]
[DecidableEq d]
[DecidableEq d₂]
(σ : MState d)
(Φ : CPTPMap d d₂ ℂ)
(Ψ : Matrix d d ℂ → Matrix d₂ d₂ ℂ)
:
Equations
- induced_norm_infty_map σ Φ Ψ = sSup {x : ℝ | ∃ (X : Matrix d d ℂ) (_ : weighted_norm_infty σ X ≠ 0), weighted_norm_infty (Φ σ) (Ψ X) / weighted_norm_infty σ X = x}
Instances For
Equations
- Gamma_inv_map σ = MatrixMap.conj ↑((↑σ).cfc fun (x : ℝ) => x ^ (-1 / 2))
Instances For
theorem
Gamma_inv_map_eq
{d : Type u_1}
[Fintype d]
[DecidableEq d]
(σ : MState d)
(X : Matrix d d ℂ)
:
Equations
- sigma_inv_sqrt σ = ↑((↑σ).cfc fun (x : ℝ) => x ^ (-1 / 2))
Instances For
theorem
T_is_CP
{d : Type u_1}
{d₂ : Type u_3}
[Fintype d]
[Fintype d₂]
[DecidableEq d]
[DecidableEq d₂]
(σ : MState d)
(Φ : CPTPMap d d₂ ℂ)
:
(T_map σ Φ).IsCompletelyPositive
theorem
T_is_positive
{d : Type u_1}
{d₂ : Type u_3}
[Fintype d]
[Fintype d₂]
[DecidableEq d]
[DecidableEq d₂]
(σ : MState d)
(Φ : CPTPMap d d₂ ℂ)
:
(T_map σ Φ).IsPositive
theorem
weighted_norm_one_eq_trace_norm_Gamma
{d : Type u_1}
[Fintype d]
[DecidableEq d]
(σ : MState d)
(X : Matrix d d ℂ)
:
noncomputable def
general_induced_norm
{d : Type u_1}
{d₂ : Type u_3}
[Fintype d]
[Fintype d₂]
[DecidableEq d]
[DecidableEq d₂]
(p q : ℝ)
(σ : MState d)
(σ' : MState d₂)
(Ψ : MatrixMap d d₂ ℂ)
:
Equations
- general_induced_norm p q σ σ' Ψ = sSup {x : ℝ | ∃ (X : Matrix d d ℂ) (_ : weighted_norm p σ X ≠ 0), weighted_norm q σ' (Ψ X) / weighted_norm p σ X = x}
Instances For
theorem
Gamma_inv_self
{d : Type u_1}
[Fintype d]
[DecidableEq d]
(σ : MState d)
(hσ : σ.m.PosDef)
:
theorem
T_map_unital
{d : Type u_1}
{d₂ : Type u_3}
[Fintype d]
[Fintype d₂]
[DecidableEq d]
[DecidableEq d₂]
(σ : MState d)
(Φ : CPTPMap d d₂ ℂ)
(hΦσ : (Φ σ).m.PosDef)
:
theorem
T_map_is_CP_proof
{d : Type u_1}
{d₂ : Type u_3}
[Fintype d]
[Fintype d₂]
[DecidableEq d]
[DecidableEq d₂]
(σ : MState d)
(Φ : CPTPMap d d₂ ℂ)
:
(T_map σ Φ).IsCompletelyPositive
theorem
HermitianMat.le_smul_one_imp_eigenvalues_le
{d : Type u_1}
[Fintype d]
[DecidableEq d]
(A : HermitianMat d ℂ)
(M : ℝ)
(h : A ≤ M • 1)
(i : d)
:
theorem
HermitianMat.eigenvalues_le_imp_le_smul_one
{d : Type u_1}
[Fintype d]
[DecidableEq d]
(A : HermitianMat d ℂ)
(M : ℝ)
(h : ∀ (i : d), ⋯.eigenvalues i ≤ M)
:
theorem
block_matrix_posSemidef
{m : Type u_11}
{n : Type u_12}
{k : Type u_13}
[Fintype m]
[Fintype n]
[Fintype k]
(X : Matrix k n ℂ)
(Y : Matrix k m ℂ)
:
(Matrix.fromBlocks (Y.conjTranspose * Y) (Y.conjTranspose * X) (X.conjTranspose * Y) (X.conjTranspose * X)).PosSemidef
The block matrix [[1, X], [X†, X†X]] is positive semidefinite.
theorem
block_matrix_one_posSemidef
{m : Type u_11}
{n : Type u_12}
[Fintype m]
[Fintype n]
[DecidableEq m]
(X : Matrix m n ℂ)
:
(Matrix.fromBlocks 1 X X.conjTranspose (X.conjTranspose * X)).PosSemidef
theorem
sandwichedRenyiEntropy_DPI
{d : Type u_1}
{d₂ : Type u_3}
[Fintype d]
[Fintype d₂]
[DecidableEq d]
[DecidableEq d₂]
{α : ℝ}
(hα : 1 ≤ α)
(ρ σ : MState d)
(Φ : CPTPMap d d₂ ℂ)
:
The Data Processing Inequality for the Sandwiched Renyi relative entropy.
Proved in https://arxiv.org/pdf/1306.5920. Seems kind of involved.