The Setting
For a generalized amplitude-damping channel with damping strength and stationary excited-state population , ordinary amplitude damping is the boundary .
The computational basis of qubits consists of bit strings , where .✦ An -qubit operator is a complex matrix indexed by these bit strings.✦ A density operator is a positive semidefinite operator with .✦
Write , with , for the single-qubit Kraus operators. For a label and computational basis bit strings and , the tensor product Kraus operator has the matrix entries in (1): (1)✦ For an -qubit operator , the memoryless channel acts as in (2); it is the tensor power of one fixed single-qubit channel. (2)✦ In (2), denotes the adjoint.
For a finite complex square matrix , the entropy is the real part of the trace of the spectral function applied to , with .✦ For density operators it is the von Neumann entropy in (3), which also specifies the binary entropy , both in nats. (3) The qubit entropy as a function of determinant is given by (4): (4)✦ For input excitation population , the phase-pair output determinant and Holevo objective are the functions in (5): (5) An optimal excitation population is a point of satisfying (6): (6)✦
A finite quantum ensemble on qubits consists of density operators and nonnegative weights summing to one, indexed by .✦ The ensemble Holevo information of a finite quantum ensemble through is the quantity in (7), measured in bits: (7)✦ The -use Holevo information of the channel is the supremum of ensemble Holevo information over finite quantum ensembles on qubits, as in (8): (8)✦
A classical communication code for channel uses consists of encoding density operators and a collective decoding positive operator-valued measure , with and .✦ The average decoding error of a classical communication code for equally likely messages is the quantity in (9): (9)✦ A rate in bits per channel use is achievable if there is a sequence of codes with such that, for every , their message counts satisfy for all sufficiently large .✦ The unassisted classical capacity of the channel is the supremum of its achievable rates, as in (10): (10)✦
The Claim
Theorem 1. For real and , the single-qubit Kraus operators satisfy (11), (12), (13) and (14): (11)✦ (12)✦ (13)✦ (14)✦ Let and let be an optimal excitation population. For every nonnegative integer , the Holevo information satisfies (15), and the operational capacity satisfies (16): (15)✦ (16)✦
In the paper, the result reads: for , let have Kraus matrices and ; for every integer , its Holevo information and capacity satisfy (17): (17)
What Is Checked
The Kraus identities are stated for real parameters, while the information identities concern tensor powers of one generalized amplitude-damping channel with , an optimal population , finite quantum ensembles, and classical codes with vanishing average decoding error.