Quantities of quantum relative entropy, namely the (standard) quantum relative entropy, and generalizations such as sandwiched Rényi relative entropy.
The operator norm of a matrix, with respect to the Euclidean norm (l2 norm) on the domain and codomain.
Equations
Instances For
The matrix representation of the map $v \mapsto v \otimes \sum_k |k\rangle|k\rangle$.
The output index is (d1 \times d2) \times d2.
Equations
Instances For
The isometry V_rho from the paper.
Equations
- V_rho ρAB = Matrix.kroneckerMap (fun (x1 x2 : ℂ) => x1 * x2) (↑ρAB.sqrt) 1 * map_to_tensor_MES dA dB * ↑ρAB.traceRight⁻¹.sqrt
Instances For
The isometry V_sigma from the paper.
Equations
- V_sigma σBC = (Matrix.reindex ((Equiv.prodComm (dC × dB) dB).trans ((Equiv.refl dB).prodCongr (Equiv.prodComm dC dB))) (Equiv.refl dC)) (V_rho (σBC.reindex (Equiv.prodComm dB dC)))
Instances For
V_rho^H * V_rho simplifies to sandwiching the traceRight by the inverse square root.
The partial trace (left) of a positive definite matrix is positive definite.
V_rho is an isometry.
V_sigma is an isometry.
The operator W from the paper, defined as a matrix product.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The operator norm of a matrix product is at most the product of the operator norms.
Define U_rho as the Kronecker product of V_rho and the identity.
Instances For
Define U_sigma as the Kronecker product of the identity and V_sigma.
Equations
- U_sigma σBC = Matrix.kronecker 1 (V_sigma σBC)
Instances For
The operator norm of the conjugate transpose is equal to the operator norm.
Helper lemmas for operator_ineq_SSA #
Core inequality: W†W ≤ I. This is the key step, following from the isometry argument: V_rho ⊗ I_C and I_A ⊗ V_sigma are isometries, their cross product has norm ≤ 1, and the result can be related to W_mat through the MES computation (Eq. 6 in Lin-Kim-Hsieh).
The intermediate operator inequality: ρ_AB ⊗ σ_C⁻¹ ≤ (ρ_A ⊗ σ_BC⁻¹).reindex(assoc⁻¹). This is derived from W_mat_sq_le_one by algebraic manipulation (conjugation and simplification).
Operator extension of SSA (Main result of Lin-Kim-Hsieh).
For positive definite ρ_AB and σ_BC:
ρ_A⁻¹ ⊗ σ_BC ≤ ρ_AB⁻¹ ⊗ σ_C
where ρ_A = Tr_B(ρ_AB) and σ_C = Tr_B(σ_BC), and the RHS is reindexed
via the associator (dA × dB) × dC ≃ dA × (dB × dC).
Weak monotonicity and SSA proof infrastructure #
Weak monotonicity, version with partial traces.
Strong subadditivity on a tripartite system
"Ordinary" subadditivity of von Neumann entropy
Triangle inequality for pure tripartite states: S(A) ≤ S(B) + S(C).
One direction of the Araki-Lieb triangle inequality: S(A) ≤ S(B) + S(AB).
Araki-Lieb triangle inequality on von Neumann entropy
Weak monotonicity of quantum conditional entropy: S(A|B) + S(A|C) ≥ 0.
Strong subadditivity, stated in terms of conditional entropies. Also called the data processing inequality. H(A|BC) ≤ H(A|B).
Strong subadditivity, stated in terms of quantum mutual information. I(A;BC) ≥ I(A;B).
The quantum conditional mutual information QCMI is nonnegative.
The quantum conditional mutual information QCMI ρABC is at most 2 log dA.
The quantum conditional mutual information QCMI ρABC is at most 2 log dC.